Get File PDF

International Journal of Food Science & Technology
Volume 43 ,Issue 9 , Pages.1507 - 1727 (September 2008)
Editorial
Application of sensory evaluation in food research (p 1507-1511)
Sarah E. Kemp
Published Online: Aug 8 2008 7:08AM
DOI: 10.1111/j.1365-2621.2008.01780.x
Original articles
Sensory aroma from Maillard reaction of individual and combinations of amino acids with glucose in acidic
conditions (p 1512-1519)
Kam Huey Wong, Suraini Abdul Aziz, Suhaila Mohamed
Published Online: May 15 2008 12:00AM
DOI: 10.1111/j.1365-2621.2006.01445.x
Physicochemical and sensory optimisation of a low glycemic index ice cream formulation (p 1520-1527)
Anthony P. Whelan, Cesar Vega, Joseph P. Kerry, H. Douglas Goff
Published Online: Jul 30 2007 12:00AM
DOI: 10.1111/j.1365-2621.2007.01502.x
Effect of glutamate and inosinate on sensory and instrumental texture of extruded products (p 1528-1533)
Renata Cassar, Ftima A. Sardinha, Jos A. G. Aras
Published Online: Oct 31 2007 12:00AM
DOI: 10.1111/j.1365-2621.2007.01548.x
Physicochemical and sensory characteristics of cookies containing residue from king palm (Archontophoenix
alexandrae) processing (p 1534-1540)
Manoela A. Vieira, Karina C. Tramonte, Rossana Podest, Sandra R. P. Avancini, Renata D. de M. C. Amboni, Edna R.
Amante
Published Online: Dec 14 2007 12:00AM
DOI: 10.1111/j.1365-2621.2007.01568.x
Impact of ripening stages of banana flour on the quality of extruded products (p 1541-1548)
Shirani Gamlath
DOI: 10.1111/j.1365-2621.2007.01574.x
Quality assessment of whole and gutted sardines (Sardina pilchardus) stored in ice (p 1549-1559)
Nuray Erkan, zkan zden
Published Online: Nov 21 2007 12:00AM
DOI: 10.1111/j.1365-2621.2007.01579.x
Probiotic potential and sensory properties of coconut flan supplemented with Lactobacillus paracasei and
Bifidobacterium lactis (p 1560-1568)
Sabrina B. M. Corra, Inar A. Castro, Susana M. I. Saad
DOI: 10.1111/j.1365-2621.2007.01585.x
Aromatic profiles of spray-dried encapsulated orange flavours: influence of matrix composition on the aroma
retention evaluated by sensory analysis and electronic nose techniques (p 1569-1576)
M. V Galmarini, M. C. Zamora, R. Baby, J. Chirife, V. Mesina
DOI: 10.1111/j.1365-2621.2007.01592.x
Probiotic foods: consumer perception and attitudes (p 1577-1580)

Julia V. Viana, Adriano G. da Cruz, Sidney S. Zoellner, Ramon Silva, Aline L. D. Batista
DOI: 10.1111/j.1365-2621.2007.01596.x
Comparison of full-fat and low-fat cheese analogues with or without pectin gel through microstructure, texture,
rheology, thermal and sensory analysis (p 1581-1592)
He Liu, Xue Ming Xu, Shi Dong Guo
Published Online: Feb 15 2008 12:00AM
DOI: 10.1111/j.1365-2621.2007.01616.x
Consumer attitude and behaviour towards tomatoes after 10 years of Flandria quality labelling (p 1593-1601)
Wim Verbeke, Liesbeth Van de Velde, Koen Mondelaers, Bianka Khne, Guido Van Huylenbroeck
DOI: 10.1111/j.1365-2621.2007.01621.x
Colour improvement of common carp (Cyprinus carpio) fillets by hydrogen peroxide for surimi production (p 16021609)
Ali Jafarpour, Frank Sherkat, Brian Leonard, Elisabeth M. Gorczyca
DOI: 10.1111/j.1365-2621.2007.01622.x
Use of a toasted durum whole meal in the production of a traditional Italian pasta: chemical, mechanical, sensory
and image analyses (p 1610-1618)
Antonietta Baiano, Clara Fares, Giorgio Peri, Roberto Romaniello, Antonella M. Taurino, Pietro Siciliano, Giuseppe
Gambacorta, Carmela Lamacchia, Sandra Pati, Ennio La Notte
Published Online: Apr 16 2008 12:00AM
DOI: 10.1111/j.1365-2621.2007.01632.x
The role of volatile compounds on aroma and flavour perception in coloured raw carrot genotypes (p 1619-1627)
Stine Kreutzmann, Anette K. Thybo, Merete Edelenbos, Lars P. Christensen
Published Online: Apr 16 2008 12:00AM
DOI: 10.1111/j.1365-2621.2007.01662.x
Effect of film and temperature on the sensory, microbiological and nutritional quality of minimally processed
cauliflower (p 1628-1636)
Ana Simn, Elena Gonzlez-Fandos, Domingo Rodrguez
Published Online: Jun 28 2008 6:13AM
DOI: 10.1111/j.1365-2621.2007.01672.x
Effects of gelatinisation level, gum and transglutaminase on the quality characteristics of rice noodle (p 1637-1644)
Seda Yalcin, Arzu Basman
DOI: 10.1111/j.1365-2621.2007.01674.x
Quality factors and grades for the classification and standardisation of complex ready pasta meals (p 1645-1656)
Maddalena Kindt, Palmira Mazzaracchio, Giancarlo Barbiroli
Published Online: Mar 25 2008 12:00AM
DOI: 10.1111/j.1365-2621.2007.01693.x
Effect of gaseous ozone on microbial inactivation and sensory of flaked red peppers (p 1657-1662)
Meltem Y. Akbas, Murat Ozdemir
Published Online: Jun 4 2008 12:00AM
DOI: 10.1111/j.1365-2621.2008.01722.x
The effect of pectin concentration on viscoelastic and sensory properties of processed cheese (p 1663-1670)
Ivana Mack, Franti ek Buka, Vladimr Pavlnek, Petra Lecinov, Jan Hrab
DOI: 10.1111/j.1365-2621.2008.01734.x
Sensory shelf life of butterhead lettuce leaves in active and passive modified atmosphere packages (p 1671-1677)
Gastn Ares, Claudia Lareo, Patricia Lema
DOI: 10.1111/j.1365-2621.2008.01736.x
Effect of soluble CO2 stabilisation and vacuum packaging in the shelf life of farmed sea bream and sea bass
fillets (p 1678-1687)
Rogrio Mendes, Amparo Gonalves
DOI: 10.1111/j.1365-2621.2008.01737.x
The effect of cutting and fish-orientation systems on the deheading yield of carp (p 1688-1692)
Andrzej Dowgiallo
DOI: 10.1111/j.1365-2621.2008.01750.x
Effect of fortification of defatted soy flour on sensory and rheological properties of wheat bread (p 1693-1698)
Morteza Mashayekh, Mohammad Reza Mahmoodi, Mohammad Hasan Entezari
DOI: 10.1111/j.1365-2621.2008.01755.x
Effects of gelatine type and concentration on the shelf-life stability and quality of marshmallows (p 1699-1704)
Johanna M. Tan, Miang H. Lim
DOI: 10.1111/j.1365-2621.2008.01756.x
Differences in chemical, microbial and sensory quality parameters of the marinated ascidia Microcosmus sabatieri
Roule, 1885 during storage at 6 C under vacuum conditions (p 1705-1713)
Nikolaos Stamatis, John Arkoudelos, Dimitris Vafidis
DOI: 10.1111/j.1365-2621.2008.01761.x
Effect of freezethaw cycles and additives on rheological and sensory properties of ready to bake frozen
chapaties (p 1714-1720)
Deep N. Yadav, Prakash E. Patki, Mohammad A. Khan, Gopal K. Sharma, Amrindar S. Bawa
DOI: 10.1111/j.1365-2621.2008.01763.x
Short communication
Enzymatic coagulation of milk: animal rennets and microbial coagulants differ in their gelation behaviour as
affected by pH and temperature (p 1721-1727)
Doris Jaros, Katrin Seitler, Harald Rohm
DOI: 10.1111/j.1365-2621.2008.01749.x
International Journal of Food Science and Technology 2008, 43, 15071511
Editorial
Application of sensory evaluation in food research
Eating and drinking should be pleasurable. The sensory

experiences evoked by foods and beverages are key to
the delivery of pleasure and crucial to commercial
success. Advertising and branding motivate consumers
to try products, but if they dont like a products
appearance, aroma, avour or texture, they will not buy
it again. Specialised research techniques are available to
measure, understand and optimise consumers sensory
experiences, so that products can be designed and
marketed to meet consumers sensory needs, thus
reducing the risk of product failure. This scientic eld
is known as sensory evaluation.
At a time when the application of sensory evaluation
is growing, it is appropriate that the IJFST has
dedicated an issue to papers in this eld. The Professional Food Sensory Group (PFSG) of The Institute of
Food Science and Technology (IFST) are pleased to be
invited to provide this editorial (with no formal
involvement in selecting and reviewing papers) and,
thereby, contribute to greater awareness and understanding of the eld.
The PFSG is the body through which the IFST
promotes sensory evaluation. It aims to:
Establish, maintain and enhance the professional
status of food sensory scientists.
Facilitate better communication of food sensory
matters.
Provide objective scientic comment on food sensory
issues.
Provide guidance on food sensory practices, such as
ethics, laboratory practices and relevant legislation.
Assess, monitor and review courses designed to enable
professional recognition of food sensory scientists.
Provide career advice for food sensory scientists.
Encourage young food sensory scientists into the
profession and assist the professional development of
food sensory scientists through membership of IFST
and PFSG.
The PFSG is one of the rst organisations in the
world to develop an accreditation scheme for sensory
evaluation training to introduce consistent standards
and proper guidance on the quality and content of basic
sensory training courses. It is also one of the rst
organisations to set ethical and professional practices
for the sensory analysis of foods. It regularly hosts
conferences, symposia and workshops on sensory evaluation and sits on the British Standards Institute
sensory committee. (Further information on the PFSG
and how to join, can be found at http://www.IFST.

org.uk.)
Sensory evaluation is traditionally dened as a scientic method used to evoke, measure, analyse and interpret those responses to products as perceived through the
senses of sight, smell, touch, taste and hearing (Stone and
Sidel, 1993). It can be divided into two areas: objective
(analytic) and subjective (hedonic). In objective testing,
the sensory attributes of a product are evaluated by a
selected and trained panel. In subjective testing, the
reactions of consumers to the sensory properties of
products are measured. The power of sensory evaluation
is realised when these two elements are combined to
reveal insights into the way in which sensory properties
drive consumer acceptance. Linking sensory properties to
physical, chemical, formulation and or process variables
then enables the product to be designed to deliver
optimum or appropriate consumer quality, functional
and emotional benets.
The science of sensory evaluation emerged in the
1940s from its origins in quality assessments made by
expert tasters employed in industries such as tea, coee
and cheese. It is now integrated into the entire product
life cycle with applications in:
Product development, including determining preferences, identifying sensory drivers of liking, targeting
sensory-based consumer segments, competitor analysis, new concept development, product design and
optimisation, scale-up and cost reduction.
QA QC, including raw material quality, sensory
specications to ensure acceptability, taint testing,
shelf-life and quality auditing through supply chain.
Marketing, including claim substantiation, delivering
emotional benets, and synergy with brand communication and advertising.
Research to improve fundamental understanding of
consumer behaviour and perception.
The papers in this special edition of IJFST illustrate
some of the broad range of areas in which sensory
evaluation can be applied in research in the food
industry, including ingredients and formulation, quality,
storage and shelf-life, and health and nutrition.
In the area of ingredients and formulation:
An investigation of the effect of pectin concentration
on processed cheese showed that pectin can be
increased without damaging appearance or avour
(Macku et al.).
doi:10.1111/j.1365-2621.2008.01780.x
2008 The Author. Journal compilation 2008 Institute of Food Science and Technology
1507
1508
Editorial
Gamlath investigated a novel use for over-ripe

bananas in our to produce acceptable extruded snack
products, measured using on a 19 Liking Scale with
thirty consumers, that showed potential for further
investigation.
Wong et al. investigated sensory aroma, using descriptive analysis with eleven trained panellists, from
Maillard reaction of amino acids with glucose in
acidic conditions to predict avour characteristics and
found that the Maillard reaction can be used as a basis
for production of specic avours.
Galmarini et al. found that matrix composition determines the aromatic prole, measured using descriptive
analysis with twelve trained panellists, of spray-dried
encapsulated orange avours.
In the area of quality:

Treatment with hydrogen peroxide was found to
increase whitening (measured using instrumental
methods) of common carp llets to be used in surimi
production (Jafarpour et al.).
The role of volatile compounds on aroma and avour
of coloured raw carrot genotypes was investigated
using descriptive analysis with ten screened, trained
panellists and instrumental techniques. A sensory map
was developed for coloured carrots related to genotype (Kreutzmann et al.).
Consumer attitude and behaviour towards tomatoes
after 10 years of using the Belgian Flandria quality
labelling system on fruit and vegetables was investigated using a survey with 373 consumers and a change
in consumer attitudes was found (Verbeke et al.).
In the areas of shelf-life and storage:

An investigation of the effect of packaging lms and
storage temperature on minimally processed cauliower showed that cauliower maintained acceptable
appearance in all the lms studied (Simon et al.).
Sensory assessment of shelf-life of butterhead lettuce
leaves in active- and passive-modied atmosphere
packages found that low storage temperatures
and active atmosphere increased acceptable shelf-life,
as determined by consumer purchase intent (Ares
et al.).
Soluble CO2 stabilisation prior to vacuum packaging
of sea bream and sea bass was found to give longer
shelf-life (Mendes & Goncalves).
A quality and shelf-life assessment of whole and
gutted sardines stored in ice was carried out (Erkan &
Ozden).
Exposure to ozone reduced microbes on aked red
pepper, with only a small change in acceptability
(Akbas & Ozdemir).
International Journal of Food Science and Technology 2008
Freeze-thaw cycles were found to have a detrimental

effect on ready-to-bake frozen chapaties, which could
be improved by adding glycerol (Yadav et al.).
In the area of nutrition and health:

The addition of gums to noodles to make them gluten
free, and thereby suitable for consumers with celiac
disease, showed promise for further investigation
(Yalcin & Basman).
Fortication of wheat bread with defatted soy our
was found to give as good a loaf as a 100% wheat
bread with higher nutritional quality and good consumer acceptability, measured on a 19 Acceptabilty
Scale with 145 consumers (Mashayekh et al.).
A characterisation of higher-bre our, dough and
pasta produced using toasted durum whole meal to
partially replace semolina was carried out (Baiano
et al.).
Flours with a higher bre content from inclusion of
king-palm residue produced acceptable cookies, with
consumer response measured using a 19 Liking Scale
and 15 Purchase Intent Scale with 100 consumers
(Vieira et al.).
Acceptable snack products were produced from highly
nutritious novel raw materials amaranth and chickpea
mixed with bovine lung, using added monosodium
glutamate and disodium inosinate to improve texture
(Cassar et al.).
Sensory and physico-chemical optimisation of ice
cream formulated using new sweeteners to produce a
low GI Index product resulted in an acceptable
candidate formulation (Whelan et al.).
An physical and sensory investigation of the addition
of pectin to improve texture of low fat cheese yielded
promising results for further investigation (Liu et al.).
The addition of probiotics to coconut an, a traditional Brazilian dessert, showed no signicant difference in acceptability over 21 days storage (Correa
et al.).
Consumer perceptions and attitudes to probiotic foods
in Rio de Janeiro, Brazil, measured using a survey
with 420 consumers, revealed that the majority of
consumers could not correctly dene probiotic, indicating the need for education to embed probiotic
concepts (Viana et al.).
It is not possible to review all the papers in detail here,

but there are a few that warrant special attention for
highlighting noteworthy points in sensory evaluation.
Two papers demonstrate that, despite advances in
instrumental analysis, sensory evaluation remains the
most sensitive and reliable measurement technique in
some situations. Galmarini et al. found that sensory
analysis was more sensitive than the e-nose for analysing
Editorial
orange avours and provided more information, in that

it can be used to assess pungency and provides names
for the sensations measured. Mendes & Goncalves
found that sensory evaluation was the most reliable
indicator of quality in sea bream and sea bass, whereas
chemical indicators, except for pH, were found to be
poor indicators of change in quality.
Consumer attitudes and behaviours change over time
because of many factors, such as availability of new
information, new product introductions, changes in lifestyle, economic trends, etc. Verbeke et al. provide an
example of such a change. They assessed consumer
attitudes and behaviours to tomatoes after 10 years of
using the Belgian Flandria quality labelling system. Over
time, the Flandria label has become the norm, appearing
on many fruits and vegetables, and so is now associated
with a normal, standard tomato, giving little perceived
differentiation in quality, apart from origin and production. Results from consumer research, such as liking,
preference, purchase intent, attitudes, beliefs, habits,
etc., need to be reassessed regularly at appropriate time
intervals or when a change occurs that impacts on the
commercial environment. It also means that systems
based on consumer data, such as quality and grading
schemes, must be checked regularly to ensure they
remain consumer-relevant.
The full potential of sensory evaluation is realised
when sensory, consumer and and or instrumental analyses are combined. Many of the papers in this edition
use instrumental, sensory and or consumer data as
independent pieces of evidence. A few papers stand out
as examples of statistically combining data from dierent sources to give more information.
Ares et al. investigated the shelf-life of butterhead
lettuce leaves in active- and passive-modied atmosphere packages. As part of the study, a sensory panel
identied the defects in lettuce leaf appearance that best
differentiated fresh and stored samples. They were
trained on these defects and then carried out an
objective quantitative descriptive assessment on lettuce
leaves stored in different packaging conditions at each
storage time. A separate group of forty consumers of
lettuce were presented with the lettuce leave samples as
above and asked a purchase intent question: Imagine
you are in a supermarket, you want to buy a package of
minimally-processed lettuce, and you nd a package of
lettuce with leaves like this, would you normally buy it?
to which they must answer yes or no. The consumer
purchase intent data were linked to the objective sensory
panel attributes (defects) using regression analysis. This
was used to estimate the level of sensory defects with the
most rapid onset, at which 25% consumer rejection
would occur. The estimated levels were considerably
more conservative than the arbitrary gure typically
used in shelf-life studies on fruit and vegetables. In the
words of the authors: Results showed the importance of
performing consumer studies in order to establish

proper criteria to estimate the shelf life of fresh fruit
and vegetables.
Kreutzmann et al. combined sensory and chemical
data using principal component analysis (PCA) to
understand how volatile compounds relate to aroma
and avour of coloured raw carrot genotypes. The PCA
was represented as a bi-plot, which served as a visual
aid to help interpretation. Correlations enabled
compounds and sensory attributes to be linked. The
presence of terpenes was found to be related to harsh
avour attributes. Interactions between sugars and
volatiles affected sensory perception, so that the
increasing sugar content resulted in masking of harsh
avour qualities.
Reliable and meaningful sensory data can only be
produced if the appropriate procedures are used. The
paper by Galmarini et al. illustrates many of these
procedures for objective sensory quantitative description, such as using an appropriate number of panellists,
training the panel on attributes and scale use, using
physical references to illustrate attributes, presenting
samples in an unbiased fashion, assessing samples in at
least duplicate (Kreutzmann et al. assessed in triplicate)
and checking sensory data to ensure results are due to
product differences rather than unreliable panel performance. In addition, it is good practice to screen for and
select panellists with appropriate sensory abilities prior
to using them in a study, as applied by Kreutzmann
et al.
Sensory evaluation has evolved into a complex, multidisciplinary eld that requires a high degree of scientic
knowledge and skill to carry out successfully. A key
challenge is to use a highly variable human measuring
instrument to gather robust, unbiased data. Even a
seemingly simple triangle test is fraught with potential
pitfalls. Sensory analysis is rarely routine and often
needs to be tailor made to match the product, situation
or assessor. There remains a tendency for companies to
hire untrained sta to run sensory programmes, sometimes driven by a shortage of experienced sensory
scientists, and for scientists from other elds to attempt
to carry out sensory testing without consulting a skilled
sensory professional and or statistician trained in sensory techniques (sensometrician).
IJFST often nds it necessary to reject papers with a
sensory element because fundamental errors have been
made. The following are some of the basic elements
necessary to carry out sound sensory research:
The study must be conducted safely and ethically. For
guidance see Ethical and Professional Practices for
the Sensory Analysis of Foods published by the IFST
PFSG. Considerations include:
s Safe and ethical treatment of assessors (panellists,
consumers and subjects):
1509
1510
Editorial
Short- and long-term health effects on

assessor groups (e.g. existing medical
conditions, allergies, effects of long-term
product usage, etc.)
Ethical treatment (e.g. informed consent,
special considerations for children, etc.)
s Safe test materials:
Safety of ingredients, (e.g. approved for
use in country of test, origin, recommended intakes, allergenic effects, identication and safety assessment on taint
compounds prior to tasting, etc.)
Effect of processing (e.g. creation of sideproducts, change in nature, etc.)
Safety of nished product (e.g. microbiologically safe, no leeching of harmful
packaging material into product, etc.)
s Safe experimental design and techniques:
Safety of test procedure (e.g. differences
from normal life conditions, safe and
hygienic test sample preparation, etc.)
Test objectives must be pre-dened and an appropriate experimental design should be used to meet these
objectives. A typical error is not using a formal
experimental design to systematically assess effects of
treatments.
The correct technique(s) to meet objectives should be
used, including:
s Appropriate
test type. Typical errors include
applying one technique, often descriptive analysis,
in all situations, and using sample labelling that
will bias response such as 1, 2, 3 or A, B, C.
s Appropriate experimental design for serving order.
A typical error is presenting samples in the same
order for all assessors.
Sufcient numbers of assessors must be used to meet
the statistical power requirements of the sensory test.
The correct type of assessors should be used. A classic
error is using a trained sensory panel or quality panel
to give consumer liking, pleasantness, preference or
acceptability ratings. (Prediction of consumer response using a trained panel is possible, but only if a
relationship has been statistically established previously.) Another error is failing to test and or target
the appropriate consumer group.
Appropriate instruction on how to perform the assessment should be given to assessors. Typical errors include
presenting scales with no information on what the scale
is measuring or how the scale should be used, and
assuming that allowing familiarisation with a product
or test procedure is the same as providing training.
Replication should be used where appropriate, such

as for descriptive analysis. As in other scientic
disciplines, replication reduces the chance of a random
or indiosyncratic result, increases reliability and
increases the statistical power of a test. It enables
additional statistical analysis to be carried out, such as
checking panel performance.
Appropriate statistical analysis techniques must be
applied to the data. Typical errors include using means
when modes and medians are appropriate for the data,
such as data generated from category scales; using
multiple comparison tests without rst carrying out an
analysis of variance (anova) or when the anova did
not show a signicant dierence; inappropriate application of Fishers LSD test; carrying out multidimensional scaling with an insucient number of samples.
The correct conclusions must be drawn from the data.
Typical errors are to conclude that products are are
not signicantly different from the results of a liking
or preference test (rather than a sensory rating or
difference test), and assuming that a sensory difference
has been caused by a particular factor when no
evidence is present.
Sufcient information on how the sensory test was
carried out must be reported. As in other scientic
elds, this must be enough information to enable the
study to be replicated. For example, how samples were
prepared (cooking, serving size, etc.) and presented
(serving vessel, labelling, temperature, lighting, etc.),
experimental design (serving order, blocking,
etc.), type of assessors (gender, age, level of training,
etc.), etc.
Sensory evaluation is a growing, dynamic eld. It

continues to broaden its applications from its roots in
food and beverages to include categories as diverse as
personal care products, household products, cars,
mobile phones and environments, to name but a few.
The role that sensory evaluation plays in organisations
continues to grow. Sensory departments were typically a
part of the QA function supplying data as a service.
They are increasingly becoming fully integrated in all
stages of a products life cycle from product conception
to post-launch monitoring and work as partners with
R&D and marketing to provide insights to help guide
development and commercial strategy. As many brands
and products are now global, sensory projects are
increasingly required to have a global perspective.
Indeed, the structure of the professional bodies in the
sensory evaluation eld is changing to become more
global. Sensory research is delivering increasingly
sophisticated techniques and better understanding of
consumer perception and behaviour using a multidisciplinary approach by linking with elds such as
Editorial
physico-chemistry, psychophysics, psychology, physiology, neuroscience and genomics.

Some challenges remain. Education continues to be a
need as a result of skills shortage in the eld. Some
traditional techniques are perceived as slow compared
with the increasing speed of product launches, and
quicker methods are under development. Testing with
human subjects requires increasing attention to health,
safety and ethical considerations, coupled with an
increased risk of litigation. Changes in employment
law in the EU are making it increasingly complex to
employ sensory assessors.
The future for sensory evaluation is bright, however,
as it transitions from an option to a necessity to meet
consumers sensory needs and deliver competitive
advantage in the quest for successful products.
Sarah E. Kemp
On behalf of the IFST PFSG committee.
References
Akbas, M.Y. & Ozdemir, M. (2008). Eect of gaseous ozone on
microbial inactivation and sensory of aked red peppers. International Journal of Food Science and Technology, 43, 16571662.
Baiano, A., Fares, C., Peri, G., Romaniello, R., Taurino, A.M.,
Siciliano, P., Gambacorta, G., Lamacchia, C., Pati, S. & La Notte,
E. (2008). Use of a toasted durum whole meal in the production of a
traditional Italian pasta: chemical, mechanical, sensory and image
analyses. International Journal of Food Science and Technology, 43,
16101618.
Cassar, R., Sardinha, F.A. & Areas, J.A.G. (2008). Eect of glutamate
and inosinate on sensory and instrumental texture of extruded
products. International Journal of Food Science and Technology, 43,
15281533.
Correa, S.B.M., Castro, I.A. & Saad, M.I. (2008). Probiotic potential
and sensory properties of coconut an supplemented with Lactobacillus paracasei and Bidobacterium lactis. International Journal of
Food Science and Technology, 43, 15601568.
Erkan, N. & Ozden, O (2008). Quality assessment of whole and gutted
sardines (Sardina pilchardus) stored in ice. International Journal of
Galmarini, M.V., Zamora, M.C., Baby, R., Chirife, J. & Mesina, V.

(2008). Aromatic proles of spray-dried encapsulated orange
avours: inuence of matrix composition on the aroma retention
evaluated by sensory analysis and electronic nose techniques.
International Journal of Food Science and Technology, 43, 1569
1576.
Gamlath, S. (2008). Impact of ripening stages of banana our on the
quality of extruded products . International Journal of Food Science
and Technology, 43, 15411548.
Jafarpour, A., Sherkat, F., Leonard, B. & Gorczyca, E.M. (2008).
Colour improvement of common carp (Cyprinus carpio) llets by
hydrogen peroxide for surimi production. International Journal of
Kreutzmann, S., Thybo, A.K., Edelenbos, M. & Christensen, L.P.
(2008). The role of volatile compounds on aroma and avor
perception in coloured raw carrot geotypes. International Journal of
Macku, I., Bunnka, F., Pavl nek, V., Lecianova, P. & Harbe, J. (2008).
The eect of pectin concentration on visoelastic and sensory
properties of processed cheese. International Journal of Food Science
Mashayekh, M., Mahmoodi, R. & Entezari, M.H. (2008). Eect of
fortication of defatted soy our on sensory and rheological
properties of wheat bread. International Journal of Food Science
Mendes, R & Goncalves, A. (2008). Eect of soluble CO2 stabilisation
and vacuum packaging in the shelf life of farmed sea bream and sea
bass llets. International Journal of Food Science and Technology, 43,
16781687.
Stone, H. & Sidel, J.L. (1993) Sensory evaluation practices, 2nd Edn.
San Diego, CA: Academic Press.
Verbeke, W., Van de Velde, L., Mondelaers, K., Kuhne, B. & Van
Huylenbroeck, G. (2008). Consumer attitude and behaviour towards
tomatoes after 10 years of Flandria quality labelling. International
Journal of Food Science and Technology, 43, 15931601.
Viana, J.V., da Cruz, A.G., Zoellner, S.S., Silva, R. & Batista, A.L.D.
(2008). Probiotic foods: consumer perception and attitudes. International Journal of Food Science and Technology, 43, 15771580.
Yadav, D.N., Patki, P.E., Khan, M.A., Sharma, G.K. & Bawa, A.S.
(2008). Eect of freeze-thaw cycles and additives on rheological and
sensory properties of ready to bake frozen chapaties. International
Yalcin, S & Basman, A. (2008). Eects of gelatinisation level, gum and
transglutaminase on the quality characteristics of rice noodle.
International Journal of Food Science and Technology, 43, 1637
1644.
1511
1512
Original article
Sensory aroma from Maillard reaction of individual and
combinations of amino acids with glucose in acidic conditions
Kam Huey Wong, Suraini Abdul Aziz & Suhaila Mohamed*
Faculty of Food Science and Technology, Universiti Putra Malaysia, 43400 Serdang, Malaysia
(Received 20 December 2005; Accepted in revised form 24 August 2006)
Summary
The aroma produced in glucoseamino acids (individual and in combination) Maillard reaction, under acidic
conditions at 100 C were determined and compared by trained panellist. Proline produced pleasant, owery
and fragrant aroma. Phenylalanine and tyrosine produced dried roses aroma. Alanine produced fruity and
owery odour, while aspartic acid and serine both produced pleasant, fruity aroma. Arginine, produced a
pleasant, fruity and sour aroma at pH 5.2, but not at its natural pH. Glycine, lysine, threonine and valine
produced a pleasant caramel-like odour. Isoleucine and leucine gave o a burnt caramel aroma. Methionine
developed a fried potato odour. Cysteine and methionine produced savoury, meaty and soy sauce-like
avours. A combination of these amino acids produced dierent types of aroma, with the stronger note
dominating the odour of the mixture. This study will help the prediction of avour characteristics of
hydrolysates from dierent protein sources.
Keywords:
Amino acids, aroma, Maillard reaction, glucose.
Introduction
The parameters that inuence the overall avours and

aroma in a Maillard reaction, are the type of amino acid
and sugar, pH, temperature, time, moisture content
(Lane & Nursten, 1983; Schieberle & Hofmann, 1997)
water activity, oxygen, the reaction medium, sulphur
dioxide and phosphates (Hurrell & Carpenter, 1977;
Namiki, 1988; Shenoy, 1993). Aroma compounds generated from Maillard reaction were mostly studied using
simple model systems with amino acids (Baltes et al.,
1989; Tressl et al., 1989). Many studies have been done
to clarify the mechanisms in organic solvent, rather than
in aqueous solution, and using amines not related to food
ingredients instead of amino acids (Hofmann, 1998a).
The taste of traditional Japanese foods like miso and
soy sauce is due to the release of amino acids from
naturally occurring proteins during fermentation. The
proper type and level of free amino acids can signicantly improve the taste of food products in naturally
occurring or intentionally added avour potentiators.
Although people have been heating sugars and amino
acids at dierent pHs and temperatures, there is still
no comprehensive study that clearly distinguishes the
specic avour formation and browning development,
especially under very acidic conditions in a Maillard
reaction. A reaction between one amino acid and one
*Correspondent: E-mail: mohamed.suhaila@gmail.com
sugar will yield hundreds of volatile compounds

(Farmer et al., 1989), which include a range of heterocyclic compounds having a ring structure containing an
atom of N, O or S, depending on the heteroatoms
present in the amino acid, producing more than one
odour in the sensory evaluation.
This study attempts to determine and compare the
various aroma resulting from the Maillard reaction of
individual and combinations of amino acids to relate to
the production of avours formed in acid-hydrolysed
protein. The sensory results of these amino acids on
avour notes may assist in the prediction of possible
avour properties of dierent hydrolysates by referring
to the types of amino acids present through amino acids
analysis.
Materials and methods
Materials
l-alanine, l-arginine hydrochloride, l-aspartic acid,

l-cysteine, l-glutamic acid, glycine, l-histidine monohydrochloride monohydrate, l-leucine, l-isoleucine,
l-lysine monohydrochloride, l-methionine, l-phenylalanine, l-proline, l-serine, l-threonine, l-tyrosine and
l-valine were purchased from Sigma Chemical Co.
(Kuala Lumpur, Malaysia). Glucose monohydrate was
obtained from Hamburg Chemical Co. (Hamburg,
Germany). Concentrated HCl (37%) and sodium
doi:10.1111/j.1365-2621.2006.01445.x
2008 Institute of Food Science and Technology
Glucose-amino acids Maillard reaction aromas K. H. Wong et al.
hydroxide pellets of analytical grade were obtained from

Merck (Kuala Lumpur, Malaysia). Puried water was
used except otherwise stated.
Maillard reaction of amino acids and d-glucose
d-glucose monohydrate (10.0 mmol in 20 mL distilled

water) were reacted with amino acids (3.3 mmol),
individually and in combinations in sealed tubes (Hofmann & Schieberle, 1995) in triplicates, under ve
dierent conditions, i.e.
Sample A: At the original pH of the amino acids and
heated at 100 1 C in conventional temperaturecontrolled thermostatic oven for 24 h;
Sample B: At the original pH of the amino acids and
heated at 100 1 C for 14 h;
Sample C: At pH 5.2 0.1 and heated at 100 1 C
for 24 h;
Sample D: In 6 m HCl as the medium and heated at
100 1 C for 24 h;
Sample E: In 6 m HCl as the medium and heated at
100 1 C for 1 h.
After the reaction, the tubes were removed and
immediately cooled in ice. Samples D and E were then
neutralized to pH 5.2 0.1. The amounts and types of
amino acids used for the amino acid combination
reaction mixtures are shown in Table 1, with and
without sulphur amino acids (cysteine and methionine).
These combinations were based on the amino acid

prole of a seed protein acid hydrolysate.
Sensory evaluation was carried independently by
eleven trained assessors from the faculty (ASTM,
1968, 1981). The descriptors for sensory analysis were
initially discussed, and a set of reference solutions was
made. The panel was introduced to these reference
solutions in two training sessions.
A set of reference solutions was prepared based on the
odour descriptor set, which consisted of bouillon,
meaty, soy sauce, smoky, malty, caramel, beany,
chicken, fruity, owery, pandan, vanilla, prawn, seafood, chocolate, coee, buttery and medicinal. The
reference solutions used in order to obtain the most
characteristic avour were:
1 Bouillon: one bouillon cube (beef avour, from Knorr,
CPC/AJI, Kuala Lumpur, Malaysia) dissolved in boiling
water.
2 Meaty: commercial acid hydrolysate (Ajieki, Ajinomoto
(Malaysia) Bhd., Kuala Lumpur, Malaysia) and concentrated yeast extract (Vegemite, Kraft Foods Limited,
Victoria, Australia).
3 Soy sauce: soy sauce (Po-Po, Hung Chun Sdn. Bhd.,
Ipoh, Malaysia).
4 Smoky: hickory smoke barbecue sauce (Knorr, CPC/
AJI).
5 Malty: soymilk with malt avour (Soyfresh, Lam Soon
Singapore Pte Ltd).
Table 1 The types and amounts of amino acids involved in the Maillard reaction of fteen and seventeen types of amino acids combined
Without sulphur amino acids
With cysteine and methionine
Amino acid
Molecular
weight
Concentration of
amino acid used
without sulphur
amino acids (mmol)
Amount of amino acids

used without sulphur
amino acids (g)
Concentration of
amino acid used
with sulphur
amino acids (mmol)
Amount of amino acids

used with sulphur
amino acids (g)
Asp
Glu
Ser
Gly
His
Arg
Thr
Ala
Pro
Tyr
Val
Met
Cys
Ile
Leu
Phe
Lys
133.11
147.13
105.09
75.07
155.18
174.20
119.12
89.09
115.13
181.19
117.15
149.20
240.30
131.18
131.18
165.19
146.19
0.4599
0.6596
0.1385
0.1970
0.0000
0.3116
0.1041
0.1371
0.1011
0.0560
0.1656
0
0
0.1337
0.3895
0.3002
0.1460
0.0122
0.0194
0.0029
0.0030
0.0000
0.0109
0.0025
0.0024
0.0023
0.0020
0.0039
0
0
0.0035
0.0102
0.0099
0.0043
0.3986
0.5717
0.1200
0.1708
0.0000
0.2700
0.0902
0.1188
0.0876
0.0486
0.1435
0.2200
0.2200
0.1159
0.3376
0.2602
0.1265
0.0106
0.0168
0.0025
0.0026
0.0000
0.0094
0.0021
0.0021
0.0020
0.0018
0.0034
0.0066
0.0106
0.0030
0.0089
0.0086
0.0037
Total
3.3000
3.3000
1513
1514
6 Caramel: burned sugar.

7 Beany: soymilk (Drinho, Lam Soon Singapore Pte Ltd).
8 Chicken: concentrated chicken stock (Maggi, Nestle
Products Sdn. Bhd., Kuala Lumpur, Malaysia).
9 Fruity: raisin (Big Sister, Madina Setia Sdn. Bhd., Kuala
Lumpur, Malaysia).
10 Flowery: chrysanthemum tea (Yeos, Yeo Hiap Seng
(Malaysia) Bhd), Kuala Lumpur, Malaysia.
11 Pandan: pandan avour essence (Star, Bush Boake Allen
Sdn. Bhd., Kuala Lumpur, Malaysia).
12 Vanilla: vanilla avour essence (Star, Bush Boake Allen
Sdn. Bhd.).
13 Prawn: prawn avour
4219 (Matrix Flavours &
Fragrances Sdn. Bhd., Kuala Lumpur, Malaysia).
14 Seafood: seafood avour powder 1879 (Matrix Flavours
& Fragrances Sdn. Bhd.).
15 Chocolate: chocolate milk (Dutch Lady, Dutch Lady
Milk Ind. Bhd., Kuala Lumpur, Malaysia).
16 Coffee: instant coffee powder (Nescafe Classic, Nestle
Products Sdn. Bhd., Kuala Lumpur, Malaysia) was
diluted in hot water.
17 Buttery: butter (Fern, New Zealand Dairy Board,
Wellington, New Zealand).
Colour references were prepared for the samples of

Maillard reaction and assessed visually by the panellists
as follows:
1 Dark brown: soy sauce (Po-Po, Hung Chun Sdn. Bhd.).
2 Brown: 20% soy sauce in water.
3 Light brown: 5% soy sauce in water.
4 Very light brown: 2% soy sauce in water.
5 Light yellow: 0.5% soy sauce in water.
The references for taste were 1% of the following
ingredients in distilled water:
1 Sweet: sucrose (Hamburg).
2 Sour: citric acid (Merck).
3 Salty: sodium chloride (Merck).
4 Bitter: caffeine (Merck).
5 Umami: Monosodium glutamate [Ajinomoto (Malaysia)
Bhd.].
The 20-mL samples of the Maillard reaction were
served at room temperature and at 60 C. All the
samples were coded with 3-digit numbers and served in
a randomised order. Six samples were served per
session and the experiments were replicated. The
detection and evaluation of the avour notes of amino
acids used only simple descriptive test (ISO, 1985). The
results were collated to produce a list of descriptive
terms applicable to the samples, based on the frequency
of usage of each descriptive word. There were no
restrictions as to how many and what terms to use. The
odour descriptions included in the results were only
those used by four or more panellists (v2 > 5.18;
signicant at P < 0.05).
Results and discussion
Flavour notes of individual amino acids
Temperature, substrates ratio, pH and time may aect

the avour notes formed during Maillard reaction.
Temperature and substrates ratio were therefore xed,
while pH and time were varied accordingly. Some
samples, such as aspartic acid and glutamic acid had low
natural pH values (pH 3.0 and 3.2 respectively), while
the others were within the range of pH 5.35.8. A variety
of aromas were detected from arginine, aspartic acid
and cysteine, implying the suitability of pH 5.2 towards
the avour formation. Serine, proline and phenylalanine
gave signicant fruity and owery aroma at their
original pH. Alanine produced a owery odour, while
aspartic acid produced a pleasant odour (Table 2). The
aroma detected from isoleucine, leucine, lysine and
threonine were caramel-like and became more distinct
on extended heating. No odour was detected from
arginine, glutamic acid and histidine (Table 2), even
after a colour change. This may be due to the very
minute amounts of volatile odorous compounds released that cannot be detected by the human olfactory
organ. The detection of odour is dependent on the
concentration and odour threshold of the key odour
impact compounds, and also on the sensitivity of the
human nose to that particular compound. Although no
odour was detected from arginine when Maillard was
performed at its natural pH values, a pleasant, fruity
and sour odour were produced at pH 5.2. Leucine,
threonine, tyrosine and valine also released slightly
stronger aroma at pH 5.2. No aroma were detected from
glutamic acid and histidine (Tables 2 and 3), probably
because pH 5.2 were not suitable for the production of
odorous substances from these two amino acids. Under
controlled pH 5.2 conditions, alanine, serine and aspartic acid produced a distinctly pleasant, fruity odour.
Cysteine formed a distinct sulphury odour. Glycine,
lysine, threonine and valine gave o a pleasant caramellike odour; while isoleucine and leucine produced a
distinct burnt caramel-like aroma. Methionine developed a distinct fried potato odour, while proline
produced a distinct owery, pleasant and pandan-like
aroma. Phenylalanine and tyrosine gave o a distinct
owery odour almost similar to dried roses. Under
uncontrolled pH, with the initial pH 5.2, most aroma
formed were almost the same as under pH 5.2 0.1
controlled conditions except alanine had additional
persimmon-like odour, arginine had no odour, glycine
and lysine had additional pleasant caramel-like odour,
serine had pleasant, slightly ciku, fresh dates-like
odour, proline had additional slightly persimmon-like
odour, phenylalanine had additional dried roses and
almond-like odour, and tyrosine had a slight dried roses
odour.
Table 2 Colour and odour of Maillard

products of amino acids and glucose formed
after heating for 14 and/or 24 h at
original pH
Amino acid
14 and/or 24 hours
Original
measured pH Colour
Odour
Alanine
5.6
Arginine
Aspartic acid
Cysteine
Glutamic acid
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Threonine
Serine
5.3
3.0
4.5
3.2
5.7
4.0
5.8
5.7
5.4
5.6
5.6
5.5
Proline
5.7
Phenylalanine
Tyrosine
Valine
None (glucose),
control
5.4
5.6
5.6
5.4
Brown
Fruity** (Persimmon), pleasant/sweet**,

flowery**,
Light yellow
None, bitter taste*
Very light brown Fruity** (ciku, fresh dates), pleasant/sweet**
Light yellow
Sulphury**
No change
None, sour taste**
Brown
Caramel-like**, pleasant/sweet*
Light yellow
None, slight sweet-sour taste
Brown
Burnt*, caramel-like*
Brown
Burnt**, caramel-like*
Brown
Pleasant/sweet*, caramel-like**, bitter taste
Brown
Fried potatoes, prawn cracker*
Light brown
Pleasant/sweet**, astringent sweet taste
Brown
Fruity* (slightly ciku, fresh dates),
pleasant/sweet*
Light brown
Flowery**, pleasant/sweet**, pandan**,
slightly persimmon, bitter taste
Brown
Flowery** (dried roses), Almond, bitter taste
Brown
Flowery** (slightly dried roses), sweet taste
Brown
Caramel-like**, bitter taste
No change
None
Ciku Manilkara achras mill.; pandan Pandanus odorus, Ridl.

*Results from the chi-square distribution showed the flavour, significantly differed from the blank
(P < 0.05, v2 < 4.8), i.e. identified by at least four or five of the eleven panellists.
**Results from the chi-square distribution showed the flavour very significantly differed from the
blank (P < 0.01, v2 < 8.25), i.e. identified by at least six or more of the eleven panellists.

after heating for 24 h at starting pH 5.2
and controlled at pH 5.2 0.1
conditions
Amino acid
Colour
Odour
Alanine
Arginine
Aspartic acid
Cysteine
Glutamic acid
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Threonine
Serine
Proline
Phenylalanine
Tyrosine
Valine
None (glucose)
Brown
Brown
Brown
Very light brown
Brown
Brown
Light yellow
Brown
Brown
Brown
Brown
Light brown
Brown
Brown
Brown (cloudy)
Brown
Brown
No change
Fruity** (ciku, fresh dates), pleasant/sweet**, flowery**

Pleasant/sweet**, fruity*, sour*
Fruity (ciku, fresh dates), pleasant/sweet**, caramel-like**
Sulphury**, slightly meaty, boiled chickpeas
None, sour umami taste,
Pleasant/sweet**, flowery*
None, sour taste
Burnt**, caramel-like**
Burnt**, caramel-like*, biscuit-like
Pleasant/sweet**, pandan*,bitter taste
Fried potatoes, prawn crackers*
Pleasant/sweet**, fruity**
Pleasant/sweet**
Flowery*, pleasant/sweet*, pandan**
Flowery** (roses), almond, Mimusops elengi flower**
Flowery*, fruity*, pleasant/sweet, tea-like
Caramel-like**, biscuit-like, malty, chocolate, bitter taste
None
Ciku Manilkara achras mill.; pandan Pandanus odorus, Ridl.

**Results from the chi-square distribution showed the flavour very significantly differed from the
1515
1516
Amino acid
Colour
Odour
Alanine
Brown
Arginine
Aspartic Acid
Cysteine
Glutamic Acid
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Threonine
Brown
Brown
Dark brown
Brown
Brown
Brown
Brown
Brown
Brown
Brown
Brown
Serine
Proline
Phenylalanine
Tyrosine
Valine
Brown
Brown
Brown
Brown
Brown
Pleasant/sweet**, fruity* (ciku, fresh dates), caramel-like**,

slightly burnt**
Caramel-like*, burnt**, slight sweet/pleasant*
Pleasant/sweet**, fruity (ciku, fresh dates), caramel-like**
Sulphury**, meaty*
Pleasant/sweet**, caramel-like**, burnt**
Pleasant/sweet**, caramel-like**, burnt**
Caramel-like**, burnt*
Burnt**, coffee-like*, prune
Burnt*, coffee-like*, caramel-like*, malty*
Pleasant/sweet**, cardboard, Herbal tea*
Meaty**, sulphury**, fried potatoes*
Pleasant/sweet**, fruity** (ciku, fresh dates), flowery*,
caramel-like**
Slightly burnt**, pleasant/sweet*
Flowery**, pleasant/sweet**, pandan*, slightly alkaline
Flowery** (roses)
Pleasant/sweet*, caramel-like*, burnt**
Pleasant/sweet**, caramel-like**, burnt*, malty

after heating for 1 h in the presence of 6 m
HCl (pH < 0.1)
Ciku Manilkara achras mill.; pandan Pandanus odorus, Ridl. All products contained black
sediment. After 24 h most of the samples had burnt, smoky aroma, except phenylalanine and
proline samples retained their pleasant, flowery aroma.
**Results from the chi-square distribution showed the flavour, very significantly differed from the
During the Strecker degradation, sulphur-containing

amino acids will lose one carbon as CO2 to form highly
odorous thioaldehydes. For example, methionine will
form methional, which produces a signicant potato
avour (Self, 1967) as similarly found for methionine in
this study (Tables 24). Compounds of particular
importance for the development of meaty avours,
reportedly have a furan or thiophene ring with a thiol
group in the 3-position, while similar compounds with a
thiol in the 2-position tend to be burnt and sulphurous (Farmer, 1994). In this study, cysteine was found to
emit a sulphury odour at its natural pH, and both
sulphury and meaty aroma in acidic pH.
Maillard-type reaction between the amino acid proline and reducing carbohydrates is well known for
generating popcorn-like, roasty aroma upon thermal
treatment (Hunter et al., 1969). Nine key odourants are
generated in thermally treated proline/glucose reaction
mixtures, of which one gave buttery odour, another gave
caramel-like odour, six gave popcorn-like odour and the
last unknown odourant gave a roasty odour (Hofmann
& Schieberle, 1998a). The distillate obtained by boiling
and simultaneously extracting an aqueous mixture of
proline and d-glucose, elicited an intense roasty, popcorn-like odour (Hofmann & Schieberle, 1998a). The
dierence in the avour developed from proline-glucose
interaction between this study and those reported
previously is probably due to the acid pH of the
reaction medium, and that the panellists were more

familiar with pandan than popcorn avour.
The factors inuencing the generation of aroma from
cysteine/carbohydrate reaction mixtures are still not
fully understood, particularly because the contribution
of a single odorant to the key avour notes has not been
established. A sensory evaluation of cysteine/glucose
mixture revealed meat-like and pungent odour notes as
the predominant odour qualities, while a cysteine/
rhamnose mixture revealed roasty, meat-like and
sulphury odour notes with the highest intensities and
cysteine/ribose mixture produced an additional caramellike and a seasoning-like odour note (Salter et al., 1988;
Hofmann & Schieberle, 1998b). Similar results were
found in this study.
Effect of extreme low pH
At 6 m HCl (which is the concentration of HCl used in

producing acid protein hydrolysates), most samples
turned brown after 1 h at 100 C (Table 4). Cysteine
was dark brown in colour after the 1-h heating, and
gradually turned black after 24 h. Tyrosine also exhibited colour change from light to a darker colour with
prolonged heating. Results clearly show that 1-h heating
at an extremely low pH is sucient, and some samples
produced a burnt odour. Other samples such as aspartic acid, cysteine, methionine, threonine, proline and
phenylalanine emitted a pleasant odour. When the

heating period was prolonged to 24 h, more samples
with the exception of methionine, proline and phenylalanine emitted burnt odour.
Flavour notes of amino acids combinations
Amino acids combinations heated at their natural pH

values at 100 C for 24 h produced an odour similar to
soy sauce, hydrolysed yeast extract (Marmite) and
slightly of dried salted sh (Table 6). When they are
heated in the presence of 6 m HCl as for the production
of acid-hydrolysed proteins, a strong chemical-like
disinfectant and chlorine-like aroma were produced.
This may be due to the usage of a strong acid in the
reaction. Heating at pH 5.2, produced a burnt, smoked
and slightly soy sauce-like aroma. Cysteine and methionine may be involved in the production of soy sauce
odour, which may have been destroyed when heated in
the presence of 6 m HCl. A good odour was produced
under a very strong acidic condition, after just 1 h of
heating at 100 C. Heating the amino acids at
Serving temperature
Table 5 shows that while serving temperature had no

eect on the colour, apparent dierences in odour
were detected. Glutamic acid and histidine had no
detectable odour when served at room temperature,
but when served at 60 C, odour was easily detected.
Glycine, lysine and serine also emitted additional
aroma. This might be due to the minute amounts of
volatiles probably hydrogen bonded to the aqueous
medium which made detection at room temperature
dicult.
Table 5 Comparison of colour and odour of Maillard products of amino acids and glucose heated and maintained at pH 5.2 0.1 served at
room temperature and 60 C
Served at room temperature
Served at 60 C
Amino acid
Colour
Odour
Colour
Odour
Arginine
Glutamic acid
Glycine
Histidine
Lysine
Threonine
Serine
Tyrosine
Brown
Brown
Brown
Light yellow
Brown
Very light brown
Brown
Brown
Brown
Brown
Brown
Light yellow
Brown
Very light brown
Brown
Brown
Pleasant/sweet**, caramel-like**, fruity, sour

Pleasant/sweet**, caramel-like*, biscuit-like*
Pleasant/sweet**, flowery*, caramel-like**
Pleasant/sweet**, caramel-like*
Pleasant/sweet**, pandan*, flowery*
Pleasant/sweet**, fruity*
Pleasant/sweet**, caramel-like**
Flowery** (roses), fruity, Pleasant/sweet, tea-like
Valine
Brown
Pleasant/sweet**, fruity*, sour*

None**, fruity sour umami taste
Pleasant/sweet**, flowery*
None**, sour taste
Pleasant/sweet**, pandan*
Pleasant/sweet**, fruity**
Pleasant/sweet**
Flowery*, fruity*, Pleasant/sweet*,
tea-like
Caramel-like**, biscuit-like, malty,
chocolate
Brown
Caramel-like**, biscuit-like, malty,

chocolate, burnt**
*Results from the chi-square distribution showed the flavour, significantly differed from the blank (P < 0.05, v2 < 4.8), i.e. identified by at least four or
five of the eleven panellists.
**Results from the chi-square distribution showed the flavour, very significantly differed from the blank (P < 0.01, v2 < 8.25), i.e. identified by at least six
or more of the eleven panellists.
Table 6 Mixture of aroma detected from the Maillard reaction of a combination of amino acids (as in Table 1 in the presence of sulphur
amino acids) with glucose under ve dierent reaction conditions
Samples
Description of odour
A: At the original pH of the amino acids and heated at 100 1 C in

conventional temperature-controlled thermostatic oven for 24 h
B: At the original pH of the amino acids and heated at 100 1 C for 14 h
C: At pH 5.2 0.1 and heated at 100 1 C for 24 h
D: In 6 M HCl as the medium and heated at 100 1 C for 24 h.
Neutralised to pH 5.2 0.1 after immediate cooling in ice
E: In 6 M HCl as the medium and heated at 100 1 C for 1 h.
Neutralised to pH 5.2 0.1 after immediate cooling in ice
Soy sauce**, marmite*, salted fish-like*

Chemical*, strong disinfectant*, chlorine-like*
Burnt*, smoky*, slightly soy sauce-like*
Flowery*, caramel-like*
Chemical*, strong disinfectant*, chlorine-like*
*Results from the chi-square distribution showed the flavour, significantly differed from the blank (P < 0.05, v2 < 4.8), i.e. identified by at least four or
more of the eleven panellists.
**Results from the chi-square distribution showed the flavour, very significantly differed from the blank (P < 0.01, v2 < 8.25), i.e. identified by at least six
or more of the eleven panellists.
1517
1518
100 1 C for 14 h, under the original pH, released a

owery and caramel-like aroma. No soy sauce-like
avours were released when cysteine and methionine
were absent. The sensation of odour is produced by the
volatile chemical substances, which stimulate the receptors in the nasal epithelium.
A combination of these amino acids produced dierent types of aroma. The types of odour produced were a
combination of aroma of each individual amino acid
with the stronger note dominating and becoming the
main odour of the samples (Table 6). For example,
sample D produced a owery odour, most probably
caused by phenylalanine and tyrosine. When cysteine
and methionine were added to the reaction composition,
the main odour was dierent. It produced soy sauce,
hydrolysed yeast extract (Marmite) and slightly salted
sh-like aroma as in sample A of Table 6. On the other
hand, the lack of cysteine and methionine in samples D
and E caused the production of a non-soy sauce odour.
This clearly shows that these two sulphur-containing
amino acids (cysteine and methionine) are the main
amino acids that are responsible for the production of a
meaty avour. The determined avour notes of the
amino acids Maillard reactions are closely related to the
avour produced by vegetable protein hydrolysates.
Colour
Glutamic acid did not exhibit any colour changes at its

original pH even after heating for 24 h (Table 2). The
Maillard reaction may have already taken place long
before any colour changes were observed. Under controlled pH, the mixture turned brown in colour at the
end of the reaction. Variation of odour was more
prominent than the colour changes under both reaction
conditions. Most samples had similar colours when
reacted at dierent pH values for 24 h, as this long
period of time is believed to allow maximum interaction
of the substrates. Arginine, aspartic acid, cysteine,
glutamic acid and proline exhibited prominent colour
changes especially at pH 5.2.
The complexity of the nonenzymatic browning reactions is known to be at least partly due to the sugar
caramelisation processes (Greenshields, 1973). In this
study, glucose solution without amino acid was also
treated under similar conditions to examine the appearance of the brown colour. Sensory results show no
changes in colour when glucose was present alone in the
reaction medium at approximately pH 5.4, as previously
reported (Ajandouz & Puigserver, 1999), and that the
intensity of browning increased with pH values, and the
presence of almost all of the essential amino acids. None
of these amino acids, with the exception of tryptophan,
gave rise to any browning in the absence of glucose.
Maillard reaction is temperature dependent and the
reaction rate may increase two to three times for each
10 C rise in temperature of model systems (Birch,

1977), and even more than this in the natural systems.
Up to 60 C, browning is normally a zero order
reaction, but at higher temperatures, changes may
follow a rst order reaction (Labuza & Saltmarch,
1981). The eect of temperature on Maillard reaction, is
also related to other variables such as acidity and water
activity. Activation energy decreases with increasing
water activity, resulting in increasing browning rate.
Reaction at 37 C occurred over a period of days, while
at above 100 C, reaction time occurred within minutes.
In a mixture of albumin and glucose, 76% of the eamino lysine were unavailable after 30 days at 37 C,
compared with 85% in 15 min at 121 C (Armstrong,
1994; Mustapha, 1997). pH value has a major inuence
on many important pathways of the Maillard reaction
and the products (Ames, 1988). The browning rate of
glucose/glycine peptides below pH 6 is greater than
glucose/glycine at the same molar concentration, but the
reverse was observed at a higher pH (Labuza & Schmidl,
1986). Increasing the pH values of the reaction medium
will generally enhance the reaction (Labuza et al., 1983;
Ashoor & Zent, 1984; Petriella et al., 1985). The type
and the amount of nal products may dier if the
experimental conditions are not exactly identical (Ames
et al., 1997), and if the pH is not controlled during the
reaction (Ames et al., 1993). The present results agree
with the previous researchers observations, and new
aromas are reported here due to the slight dierences in
reaction conditions.
The likelihood of mutagenic or carcinogenic compounds formed under the conditions of the reaction is
beyond the scope of this study. The formation and
occurrence of carcinogenic heterocyclic amines in glucoseamino acids model systems have been reported
(Johansson et al., 1995) and reviewed (Skog et al.,
1998).
Conclusions
In this study, most amino acids produced a pleasant or

acceptable avour with only a few producing unpleasant
burnt and sulphury aroma. Consequently, the Maillard
reaction can be used as a basis for the production of
specic avouring products by carefully selecting the
sugars and amino acids, and controlling the processing
conditions within narrow specications. Both pH and
time play an important role in controlling the type of
odorous compounds that are produced and the use of
extreme pH, such as 6 m HCl would produce a burnt
odour even with only an hour of heating. Prolonged
heating time under this condition would cause the
unpleasant odour to become even more intense. Combination of amino acids produced dierent results.
Slightly acidic amino acid mixtures heated for 24 h
(Table 6, sample C) produced a better odour than amino
acid mixtures heated at their natural pH for (sample B)

14 h. The serving temperature also plays an important
role in the detection of odour. At 60 C, higher
concentrations of volatile compounds are being released.
References
Ajandouz, E.H. & Puigserver, A. (1999). Nonenzymatic browning
reaction of essential amino acids: eect of pH on caramelization and
Maillard reaction kinetics. Journal of Agricultural and Food
Chemistry, 47, 17861793.
Ames, J.M. (1988). The Maillard browning reaction an update.
Chemistry and Industry, 5, 558561.
Ames, J.M., Apriyantono, A. & Arnoldi, A. (1993). A low molecular
weight coloured compounds formed in xylose-lysine model system.
Food Chemistry, 46, 121127.
Ames, J.M., Arnoldi, A., Bates, L. & Negroni, M. (1997). Analysis of
the methanol-extractable of a model extrusion-cooked cereal
product. Journal of Agricultural and Food Chemistry, 45, 12561263.
Armstrong, H.J. (1994). Enhanced protein gelation using the Maillard
reaction and elevated temperatures. PhD Thesis, University of
Nottingham, UK.
Ashoor, S.H. & Zent, J.B. (1984). Maillard browning in common
amino acids and sugars. Journal of Food Science, 49, 12061207.
ASTM (1968). ASTM Committee E-18 on Sensory Evaluation of
Materials and Products. ASTM STP 434. In: J.B. Wheeler, H.M.
Hoersch, H.P. Mahy & A.S. Kleinberg (eds), Manual on Sensory
Testing Methods. Pp. 78. Philadelphia, PA: American Society for
Testing and Materials.
ASTM (1981). ASTM Committee E-18 on Sensory Evaluation of
Materials and Products. ASTM STP 758. In: Guidelines for the
Selection and Training of Sensory Panel Members (edited by J.B.
Wheeler, H.M. Hoersch, H.P. Mahy & A.S. Kleinberg). Pp. 528.
Philadelphia, PA: American Society for Testing and Materials.
Baltes, W., Kunert-Kirchho, J. & Reese, G. (1989). Model reactions
on generation of thermal aroma compounds. In: Thermal Generation
of Aromas (edited by T.H. Parliment, R.J. McGorrin & C.T. Ho).
Pp. 141155. ACS Symposium Series 409. Washington, DC:
American Chemical Society.
Birch, G.G. (1977). Chemical, physical and biological changes in
carbohydrates induced by thermal processing. In: Physical, Chemical
and Biological Changes in Food Caused by Thermal Processing
(edited by T. Hoyem & O. Kvale). Pp. 152167. London: Applied
Science Publishers Limited.
Farmer, L.J. (1994). The role of nutrients in meat avour. Proceedings
of the Nutrition Society, 53, 327333.
Farmer, L.J., Mottram, D.S. & Whiteld, F.B. (1989). Volatile
compounds produced in Maillard reactions involving cysteine,
ribose and phospholipid. Journal of the Science of Food and
Agriculture, 49, 347368.
Greenshields, R.N. (1973). Caramel-part 2. Manufacture, composition
and properties. Process Biochemistry, 4, 1720.
Hofmann, T. (1998a). Characterization of the chemical structure of
novel colored Maillard reaction products from furan-2-aldehyde
and amino acids. Journal of Agricultural and Food Chemistry, 46,
932940.
Hofmann, T. & Schieberle, P. (1995). Evaluation of the key odorants
in a thermally treated solution of ribose and cysteine by aroma
extraction dilution techniques. Journal of Agriculture and Food
Chemistry, 43, 21872194.
Hofmann, T. & Schieberle, P. (1998a). Flavor contribution and
formation of the intense roast-smelling odorants 2-propionyl-1pyrroline and 2-propionyltetrahydropyridine in Maillard-type reactions. Journal of Agriculture and Food Chemistry, 46, 27212726.
Hofmann, T. & Schieberle, P. (1998b). Identication of key aroma
compounds generated from cysteine and carbohydrates under
roasting conditions. Zeitschrift fur Lebensmittel Untersuchung und

-Forschung A, 207, 229236.
Hunter, I.R., Walden, M.K., Scherer, J.R. & Lundin, R.E. (1969).
Preparation and properties of 1,4,5,6-tetrahydro-2-acetopyridine, a
cracker-odor constituent of bread aroma. Cereal Chemistry, 46, 189
195.
Hurrell, R.F. & Carpenter, K.J. (1977). Maillard reactions in foods. In:
Physical, Chemical and Biological Changes in Food Caused by
Thermal Processing (edited by T. Hoyem & O. Kvale). Pp. 168184.
London: Applied Science Publishers Limited.
ISO (1985). Sensory Analysis Methodology General Guidance, 1st
edn. International Standard 6658, Ref. No. ISO 66581985 (E). Pp.
89. Switzerland, Gene`ve: International Organization for Standardization.
Johansson M.A.E., Fay L.B., Gross G.A., Olsson K. & Jagerstad M.
(1995). Inuence of amino acids on the formation of mutagenic/
carcinogenic heterocyclic amines in a model system. Carcinogenesis,
16, 25532560.
Labuza, T.P. & Saltmarch, M. (1981). The nonenzymatic browning
reaction as aected by water in foods. In: Water Activity: Inuences
on Food Quality (edited by L.B. Rockland & G.F. Stewart). Pp. 605
650. New York: Academic Press.
Labuza, T.P. & Schmidl, M.K. (1986). Advances in the control of
browning reactions in foods. In: Role of Chemicals in the Quality of
Processed Food (edited by O.R. Fennema, W.H. Chang & C.Y. Lii).
Pp. 6595. Westport, CT: Food and Nutrition Press.
Labuza, T.P., Loncin, M. & Weiss, M. (1983). Investigations into the
inuence of water activity, pH and heat treatment on the velocity of
the Maillard reaction in foods. Journal of Food Quality, 6, 103118.
Lane, M.C. & Nursten, H. E. (1983). The variety of odors produced in
Maillard model systems and how they are inuenced by the reaction
conditions. In: G.R. Waller & M.S. Feather (eds), The Maillard
Reaction in Food and Nutrition. ACS Symposium Series 215. Pp 141
158. Washington, DC: American Chemical Society.
Mustapha, W.A.W. (1997). Inuence of the physico-chemical environment as product formation in the Maillard reaction. PhD Thesis,
University of Nottingham, UK.
Namiki, M. (1988). Chemistry of Maillard reactions: recent studies on
the browning reaction mechanism and the development of antioxidants and mutagens. Advances in Food Research, 32, 115143.
Petriella, C., Resnik, S.L., Lozano, R.D. & Chirife, J. (1985). Kinetics
of deteriorative reactions in model food systems of high water
activity: color changes due to non enzymatic browning. Journal of
Food Science, 50, 622625.
Salter, L.J., Mottram, D.S. & Whiteld, F.B. (1988). Volatile
compounds produced in Maillard reactions involving glycine, ribose
and phospholipid. Journal of the Science of Food and Agriculture, 46,
227242.
Schieberle, P. & Hofmann, T. (1997). Identication of the character
impact odorants in fresh strawberry juice by quantitative measurements and sensory studies on model mixtures. Journal of Agriculture
and Food Chemistry, 45, 227232.
Self, R. (1967). Potato avor. In: Chemistry and Physiology of Flavors
(edited by H.W. Schultz, E.A. Day & L.M. Libbey). Pp 362389.
Westport, CT: AVI Publishing Company, Inc.
Shenoy, V.R. (1993). Maillard reactions and its applications. Journal
of Scientic and Industrial Research, 52, 684689.
Skog K.I., Johansson M.A.E. & Jagerstad M.I. (1998). Carcinogenic
heterocyclic amines in model systems and cooked foods: a review on
formation, occurrence and intake. Food and Chemical Toxicology,
36, 879896.
Tressl, R., Helak, B., Martin, N. & Kersten, E. (1989). Formation of
amino acid specic Maillard products and their contribution in
thermally generated aromas. In: Thermal Generation of Aromas
(edited by T.H. Parliment, R.J. McGorrin & C.T. Ho). ACS
Symposium Series 409. Pp. 156171. Washington, DC: American
Chemical Society.
1519
1520
Original article
Physicochemical and sensory optimisation of a low glycemic
index ice cream formulation
Anthony P. Whelan,1,2 Cesar Vega,1 Joseph P. Kerry1 & H. Douglas Goff3*
1 Department of Food and Nutritional Sciences, University College Cork, Ireland
2 Macs Ice Cream Ltd, Killarney, County Kerry, Ireland
3 Department of Food Science, University of Guelph, Guelph, ON, N1G 2W1, Canada
(Received 21 June 2006; Accepted in revised form 25 October 2006)
Summary
The development of a low glycemic index ice cream with as close as possible physicochemical properties and
sensory quality compared with a sucrose-sweetened ice cream was investigated. Three relatively new novel
commercial sweeteners tagatose, erythritol and trehalose were studied, along with maltitol and
polydextrose. Once the freezing curves were matched, other physicochemical properties also were found to
match. Sweetness and sweet taste could then be adjusted for sensory optimisation with a combination
of these sugars and supplementation with sucralose to boost the sweetness as necessary.
Keywords
Glycemic index, ice cream, no sugar added, sweetener.
Introduction
The glycemic index (GI) is a measure of the ability of

food (specically, the carbohydrate in food) to raise
blood sugar (glucose) levels after consumption compared with an equivalent dose of glucose. Low-GI foods
release glucose slowly into the blood, producing a
gradual and relatively low rise in blood glucose and
insulin levels. Foods with a low GI play an important
role in the dietary management of diabetes, weight
reduction, peak sports performance and the reduction of
risks associated with heart disease and hypertension
(Jenkins et al., 1981).
Although some studies on diabetic ice creams and
low-calorie desserts have been reported, very little has
been published on formulating ice creams suitable
specically for low-GI diets. Abril et al. (1982), using
a combination of xylitol, fructose and corn syrup solids
(CSS), produced an ice cream of acceptable quality that
was reported to have physicochemical attributes similar
to sucrose-sweetened ice cream. An ice cream sweetened
with combinations of xylitol (315%) and sucrose (15
3%) also has been reported to have characteristics
similar to sucrose-sweetened ice cream (Marco &
Pearson, 1982). Go & Pearson (1983) and Go &
Jordan (1984) reported that an acceptable low-calorie
frozen dessert could be formulated with polydextrose,
CSS, aspartame and microcrystalline cellulose. Specter
*Correspondent: E-mail: dgo@uoguelph.ca
& Setser (1994) also used polydextrose and aspartame in

combinations with milk fat substitutes (tapioca dextrin
or potato malto-dextrin) to formulate low-calorie frozen
desserts and found no signicant dierences in the
textural attributes between such formulations and a
sucrose-sweetened control. Ozdemir et al. (2003) produced diabetic ice cream using maltitol, sorbitol and
high-fructose corn syrup as the sweetening agents and
compared them with a sucrose-sweetened control. Sensory analysis showed that maltitol-based ice cream was
more preferred than that containing sorbitol. Bordi
et al. (2004) compared a regular 12% fat vanilla ice
cream and a maltitol-sweetened ice cream using a large
taste panel and showed an overall consumer preference
for the maltitol-sweetened ice cream.
Table 1 shows some of the properties of commercially
available sweeteners. Sucrose is the most widely used
sweetener, although some other sweeteners, such as CSS
or combinations of functional ingredients and highintensity sweeteners, have reduced its use in industrial
food applications. Tagatose a new low-calorie, low GI
sweetener derived from lactose exhibits a prebiotic
eect and promotes the growth of healthy bacteria in the
gut. The laxative eects of tagatose limit its use to
2030 g day)1 (Levin et al., 1995). Erythritol is a relatively new 4-carbon polyol that cannot be metabolised by
the human enzymatic system; thus, it is excreted 90%
unchanged (Portmann & Kilcast 1996). Trehalose has
been reported to have a clean taste prole with no
aftertaste and a sweetness prole that is characterised by
doi:10.1111/j.1365-2621.2007.01502.x
Low glycemic index ice cream formulation A. P. Whelan et al.
Table 1 Properties of commercially available sweeteners
Sucrose*
Lactose*
Fructose*
Erythritol
FOSd (Inulin)
Isomalt
Lactitol
Mannitol
Maltitol
Sorbitol
Xylitol
Tagatose
Trehalose
Polydextrose
10 DE Mdxf*
Laxative threshold
(g per day) intestinal
discomfort
Glycemic
index (GI)
Molecular
weight
None
Casesg
None
>100 low
2030 high
35 high
High
20 high
100 low
70 medium
50 medium
30 high
Casesg
90 low
None
59 10
46
19 7
0
0
17 15
22
0
55 16
72
71
31
Unknown
62
8090
342
342
180.16
122
>464
344
344
182
344
182
152
180
378
1825000
FPDa
SEb
Caloric
valuec
(kcal g)1)
1.00
1.00
1.90
2.80
0.74
0.99
0.99
1.88
0.99
1.88
2.25
1.90
0.90
0.60
0.21
100
14
180190
5370
4060
3560
3040
5060
8590
60
87100
92
45
40
5
4.0
4.0
3.7
0.2
2.0
2.0
2.0
1.6
3.0
2.6
2.4
1.5
3.62
1.0
4.0
Solubility
ww%
(25 C)
67 high
22 low
High
36 med
Medium
Medium
High
Low
60 high
72 high
66 high
High
>45 med
High
High
Relative
cost
1
<1
>1
56
34
>1
<1
Molecular formula
C12H22O11
C12H22O11
C12H12O6
C4H10O4
Mixture of FOSd
C24H48O22
C12H24O11.2.H2O
C6H14O6
C12H24O11
C6H14O6
C5H12O5
C6H12O6
C12H22O11.2.H2O
Long-chain PSe
Long-chain dextrins
Freezing point depression.

Sucrose equivalents, based on mol. wt.
c
Reported values for the USA; the European Union has a standard value of 2.4 kCal g)1 for all polyols.
d
Fructo-oligosaccharides.
e
Polysaccharides.
f
Maltodextrins.
g
In cases of lactase or trehalase deficiencies.
*Caries causing.
Source: Carabin & Flamm (1999); DuBois (2000); Foster-Powell et al. (2002); Jenkins et al. (1981); Marie (1991); Marshall et al. (2003); Nabors (2001);
Newsome (1986); Noda et al. (1994); Patil et al. (1987); Portmann & Birch (1995); Portmann and Kilcast (1996).
b
a rapid onset of sweetness with comparable persistence

to sucrose. Unlike sucrose, however, the sweetness of
trehalose increases in a nonlinear fashion with increasing
concentration (Portmann & Birch, 1995).
Three factors have to be considered when developing a
sweetening system in ice cream: desired sweetness and
sweet taste; freezing point depression (FPD); and contribution to total solids (Rothwell, 1985; Stampanoni
Koeferli et al., 1995; Guven & Karaca, 2002). Sweeteners
are classied according to their sucrose equivalence (SE),
where sucrose has an arbitrary value of 100 (Nabors,
2001). The desired sweetness in ice cream ranges between
13% and 16% SE (Guinard et al., 1996). Most sweeteners contribute to the FPD of ice cream mix (Smith et al.,
1984). Ice cream must be easily scooped at intended
serving temperature and provide an optimum melting
sensation when consumed (Marshall et al., 2003). In
order to make a suitable low-GI ice cream, the carbohydrate content (sucrose and lactose) must be reduced
and replaced with alternative carbohydrate sources.
The objectives of this study were to investigate the
suitability of alternative sweeteners for the manufacture
of low-GI ice cream, to optimise formulations to match
their freezing curves and other physicochemical properties and to enhance the sensory attributes of the most
suitable formulations.
Mix formulation
The ingredient sources were as follows: whole milk

(CMP Dairies, Cork, Ireland); skim milk powder (SMP;
Golden Vale Plc, Cork, Ireland); anhydrous milk fat
(AMF; Dairygold, Kerry, Ireland); sugar (Irish sugar,
Carlow, Ireland); milk protein concentrate (MPC 80;
Alaplex 4850, NZMP, New Zealand); maltitol
(C*Maltidex CH 16385; Cerestar, Milan, Italy); erythritol (C*Eridex 16952; Cerestar, Milan, Italy); tagatose
(Arla Foods Ingredients, Skanderborgvej, Denmark);
polydextrose (litesse P) and fructose (Fructon C)
(Danisco Sweeteners Ltd, Surrey, UK); locust bean
gum, j-carrageenan (from Eucheuma cottonii), and
emulsier (a distilled monoglyceride made from edible,
rened rapeseed oil and blended with fully hardened
rapeseed oil with an iodine value of 30) (Danisco,
Brabrand, Denmark); and sucralose (Tate & Lyle,
McIntosh, Alabama).
All formulations consisted of 6% milk fat, 3.6% milk
protein, 0.2% mono- and di-glycerides, 0.14% locust
bean gum and 0.014% carrageenan. To minimise the
lactose content, AMF and MPC were used to supply the
fat and protein. All mixes were formulated to obtain
1521
1522
Table 2 Ice cream mix formulations for physical property

characterisation
Ingredients
Control
Full fat milk

Water
AMF
SMP
MPC 80
Sugar
Fructose
Erythritol
Tagatose
Maltitol
Polydextrose
Trehalose
Total solids %
Lactose %
Sucrose equivalents %
77.2
3.5
3.9
Table 3 Formulations developed for optimisation of sensory

attributes
Mix 1
Mix 2
Mix 3
Mix 4
Mix 5
67.6
6.0
69.7
6.0
68.8
6.0
65.7
6.0
64.3
6.0
4.5
4.5
4.5
4.5
4.5
15.0
1.6
5.5
6.0
3.0
3.4
15.0
31.9
6.0
16.0
14.0
14.0
14.0
31.9
0.04
14.2
29.8
0.04
9.4
30.8
0.04
10.8
8.4
33.0
0.04
18.3
24.9
32.9
0.04
10.1
AMF, anhydrous milk fat; SMP, skim milk powder; MPC, milk protein
concentrate.
the same theoretical FPD ()2.15 0.05 C), the calculation of which is described as follows. A target of
32 4% was set for the total solids. Overrun was xed
at 100 10%. The carbohydrate content of the various
formulations is shown in Table 2.
Mix 1 contained the maximum amount of tagatose
(6%) recommended by the manufacturer and 1.6%
fructose to make up total solids and desired FPD. Mix 2
contained as much erythritol as possible (5.5%) to
balance the FPD of the control. Mix 3 contained a
combination of erythritol (3.0%) and tagatose (3.4%).
Mix 4 contained 15% maltitol and 8.45% trehalose. Mix
5 contained only trehalose. Sucralose (600 times the
sweetness of sucrose) was added to boost the sweetness
as necessary.
A second set of formulations, aimed to improve the
dairy character of the ice creams (compared with those
detailed earlier), included the use of milk and cream.
The carbohydrate content of these mixes is shown in
Table 3. Mix A was similar to mix 4 earlier. Mix B
contained the maximum usage of tagatose (6%) with
6% polydextrose and 3% maltitol used to make up total
solids. Mix C contained 15% maltitol and 2.5%
trehalose. Mix D contained as much erythritol (4.2%)
as possible with 6% polydextrose and 3% maltitol
making up total solids.
Ice cream manufacture
When polydextrose was used, it was pre-homogenised

with the water fraction of the mix at 70 C to improve
hydration. Powdered ingredients were pre-blended and
dispersed into the liquid fraction at 40 C. The fat
source was added at 50 C. Ice cream mixes, 10 L, were
Ingredients
Control
Full fat milk

Water
Cream
AMF
SMP
MPC 80
Sugar
Erythritol
Tagatose
Maltitol
Polydextrose
Trehalose
Total solids %
Lactose %
Sucrose equivalents %
Calorie reduction %
71.8
74.2
71.4
76.3
8.9
9.2
8.8
4.5
1.5
1.6
1.4
15.0
6.0
3.0
6.0
65.7
9.1
6.0
3.8
15.0
4.2
31.7
5.9
16.0
0
8.4
33.0
0.04
18.3
16.5
32.0
4.0
11.5
30.5
15.0
2.5
32.0
3.9
16.6
18.2
3.0
6.0
27.9
4.1
9.0
35.5
AMF, anhydrous milk fat; SMP, skim milk powder; MPC, milk protein
concentrate.
batch pasteurised (85 C, 10 s), homogenised at

21 3 MPa (2-stage, Model APV 1000; Albertslund,
Denmark; for larger batches, 80 L, a Niro homogeniser
was used), cooled to 4 C and aged overnight. Before
freezing, 3 ml L)1 of vanilla avour was added to the
mix. Ice cream was frozen using a scraped surface
continuous ice cream freezer (Armeld FT25A; Armeld Ltd., Ringwood, UK) with a draw temperature of
)5.0 C 0.5 C, dasher speed of 200 r.p.m. and
product back pressure of 2 Bar. A continuous Cherry
Burrell Vogt VA-80 ice cream freezer (capacity
300 L h)1) was used to freeze ice cream for larger
batches, also at a draw temperature of )5 0.5 C. Ice
cream was lled into 500-mL or 2-L plastic containers
and transferred to a blast freezer at )35 C for 3 h
before being transferred into a freezer at )25 C until
subsequent analysis.
Analyses
Freezing characteristics
The theoretical model most frequently used by the ice

cream industry to calculate the freezing curves of ice
cream mixes is that of Leighton (1927) as modied by
Bradley & Smith (1983), Bradley (1984) and Marshall
et al. (2003). Table 4 shows an easier adaptation that,
when fed into an Excel spreadsheet, will give the predicted
FPDTotal (freezing point) and theoretical freezing curve,
plotted as % water frozen in the mix (rst column)
graphed against the FPDTotal (last column).
Modulated temperaturedierential scanning calorimetry (MT-DSC) (Q1000 TA Instruments, New Castle,
DE, USA) was used to perform thermal evaluations.
Table 4 Theoretical initial freezing point and freezing curve

WFd (%)
g Suc 100 g water
WNF (g)
FPDh,bSweetener
FPDSalt
FPDTotal
090 in
(% W *(1 ) (WF
=(()9*0.00001)*
FPDSweetener + FPDSalt
((SE*100) WNF (g))
(% MSNF *2.37 ) (g)
increments
(g Suc 100 g water)2 ) (0.0612)*
water not frozen
(%) 100)))
of 5
(g Suc 100 g water))
Example: % water frozen in mix = 0; (g) water not frozen = 62.85; % SE = 21.86; % MSNFf = 12
% SEg = the sum of all carbohydrate (carbohydrate*FPDh potential compared to sucrose) sources in the mix; e.g. Sucrose (15*1) + Lactose
(6.86*1) = 21.86
(12*2.37) 62.85 = )0.45
2.14 + 0.45 = )2.68
0
(62.85*(1 ) (0 100)
(21.86*100) 62.85
=(()9*0.00001)*(34.78)2 )
(0.0612)*(34.78)) = )2.23
= 62.85
= 34.78
5
(62.85*(1 ) (5 100)
(21.86*100) 59.71
=(()9*0.00001)*(36.62)2 )
(12*2.37) 59.71 = )0.48
2.29 + 0.48 = )2.84
= 59.71
= 36.62
(0.0612)*(36.62)) = )2.36
10, etc.
c
The constant 2.37 is based on the average molecular weight and concentrations of salts present in milk.
Polynominal equation with intercept through zero derived from regression model where g sucrose 100 g water is graphed against FPD C. Data were
extrapolated from Leighton (1927).
c
W, water in the mix.
d
WF, water frozen in the mix.
e
WNF, water not frozen.
f
MSNF, milk solids not fat.
g
SE, sucrose equivalent.
h
FPD, freezing point depression.
b
The instrument was calibrated using sapphire, gallium

(mp 29.8 C; DH = 80 J g)1) and indium (mp =
156.6 C; DH = 28.45 J g)1). The sample size was
15 mg. The temperature protocol used to measure
the melting of ice was: equilibration at 25 C, cooled to
)25 C at a rate of 2 C min)1 and held for 30 min at
)25 C, cooled to )60 C at a rate of 2 C min)1,
warmed to )36 C and held for 30 min, modulation on
(amplitude 0.318 C every 60 s), held for 5 min and
warmed to 15 C at a rate of 2 C min)1 (Go et al.,
2003). Freezing points were calculated from the DSC
freezing curves as the temperature at which the steepest
slope was observed on the melting endotherm. Using the
method of De Cindio et al. (1995), as applied by Livney
et al. (2003), the amount of ice frozen at any given
temperature was calculated, generating freezing curves
for the dierent solutions studied.
Mix viscosity
Viscosity was determined at 4 C after aging using an

R S Soft Solids Tester Rheometer (Brookeld Viscometers Ltd, Essex, England) with concentric cylinder and
cone spindle CC-45. The Rheometer was programmed
to rotate for 5 min at increasing shear rate from 0 to
100 s)1, with apparent viscosity taken as the value at
30 s)1.
Fat droplet particle size
The particle size distribution of the mix and melted ice

cream were measured by integrated laser light scattering
using a Mastersizer model S equipped with a 300 RF
(reverse Fourier) lens and a He-Ne (k = 633 nm) laser
(Manufacturers presentation code 0303, specic for the
refractive index of milk fat, Malvern Instrument Ltd.,
Worcestershire, UK). The mean particle size diameter
D[4,3] (the volume-surface-weighted diameter) was used
to characterise the droplet size in the mix. The cumulative percentage of the particles 2.3 lm and the
volume-weighted-diameter D(v,0.9) were used to represent partially coalesced fat in the ice cream (Bolliger
et al., 2000).
Fat destabilisation index
The mix and ice cream samples (1.00 g) were diluted

1:500 in two steps with deionised water, and the
absorbance was measured by a spectrophotometer
(Beckman DU640; Ramsey, MN, USA) at 540 nm
against deionised water as a blank. Fat destabilisation
index was calculated as follows (Bolliger et al., 2000):
% fat destabilisation Absorbance in mix
Absorbance in ice cream
=Absorbance in mix 100
Overrun
Overrun was measured by comparing the weight of mix

and ice cream in a xed volume container. Overrun was
calculated as follows: % Overrun = [(weight of mix )
weight of ice cream) 100] (weight of ice cream).
Melting test
Ice cream samples (500 mL) were placed on a 6-mesh

grid at room temperature (20 C). The weight of the
ice cream at time 0 and of the dripped portion passing
1523
through the screen was recorded every 10 min for

120 min. Tests were done in triplicate with two repetitions within each replicate. The time (min) was plotted
against the dripped volume (as % mass lost) and the
maximum meltdown rate corresponded to the highest
gradient (slope) in the ascending meltdown curve.
Hardness
The texture analysis was conducted using a TA.XT2

Texture Analyzer (Stable Micro Systems, Reading, UK)
tted with a 2.5-cm diameter acrylic conical probe (SMS
P 60C), set to record the force needed to penetrate the
sample (500 mL) to a depth of 20 mm at a speed of
2 mm s)1. The hardness was measured as the peak
compression force (N) during the penetration of the
sample. Means were calculated from four readings from
separate containers from each of the three separate
batches (n = 12).
Sensory analysis
A ve-member expert taste panel assessed all of the ice

creams described in Table 2, while a larger taste panel
(n = 70) assessed the ice creams described in Table 3.
Ice creams were compared with the control for intensity
(sweetness intensity, vanilla avor and dairy avour)
and liking attributes (sweetness liking and overall
liking). The samples were served in the 100-mL plastic
container straight from a 2-L-tempered container to
ensure that all ice creams were of the same consistency.
The cups were labelled with random three-digit codes.
The order of presentation of the samples was randomised across subjects. Subjects indicated their degree of
liking (DOL) on a 140-mm line scale. Ice creams were
ranked as either less or more intense or liked compared
with the control, which was given a xed value of 0
anchored in the centre.
Statistical analysis
Statistical analysis of the data was carried out using

Microsoft Excel 2000 (Microsoft Corporation, USA),
anova single-factor and two-factor test. When signicant eects were evidenced (P < 0.05) between sample
treatments, T test (LSD) was used to compare the means

of each parameter. All analyses were performed in duplicate from three separate replications unless otherwise
stated.
Physical properties
Our rst goal was to utilise the alternative carbohydrates in mixes that contained the same freezing curve
as the control and then measure their other physical
parameters. The calculated and measured freezing
points (Table 5) and freezing curves (Fig. 1) were
similar regardless of the ice cream formulation. This
was in agreement with previous ndings (Livney et al.
2003). Comparison of the measured percentage of total
water frozen with the theoretical value showed good
accuracy up to 5055% of total water frozen (Fig. 1),
after which the theoretical method overestimated the
amount of water frozen in the system also concurrent
with previous reports (Livney et al. 2003). The use of a
10
20
% total water frozen

30
40
50
60
70
80
90
0
5
10
C
1524
15
20
25
30
35
Figure 1 Experimental freezing curves for ice cream mixes as described

in Table 2, obtained from dierential scanning calorimetry (DSC)
measurements using de De Cindio et al. (1995) method. Control ()),
mix 1 (n), mix 2 (s), mix 3 (*), mix 4 (), mix 5 ()) and theoretical
(h). Means from duplicate measurements from three replications
(n = 6).
Table 5 Physicochemical parameters of mix and ice cream from formulations described in Table 2
Control
Mix 1
Mix 2
Mix 3
Mix 4
Mix 5
Measured freezing
ptc (C)
Viscosity at
30 s)1 (mPa s)
Fat destabilization
(%)
Slope meltdown
% Min
Fat in drip (%)
Hardness (N)
)2.4a
)2.2a
)2.2a
)2.1a
)2.1a
)2.2a
54a
51a
46b
50ab
55a
53a
40.8a
52.8a
49.2a
52.6a
50.1a
51.5a
5.5a
5.3a
5.6a
5.3a
5.5a
5.3a
3.2a
2.7ab
2.7ab
2.0b
2.7ab
2.5ab
57.3ab
59.1a
46.2b
52.6ab
50.5ab
45.8ab
0.28
0.12
0.24
0.38
0.14
0.10
3
2
2
2
3
1
10.51
4.62
6.38
8.86
4.10
5.28
0.35
0.34
0.29
0.56
0.72
0.51
0.27
0.94
0.95
0.81
0.29
0.53
12.58
8.43
2.96
6.05
8.73
14.50
a,b
Means with different letters in the same column are significantly different. Means from duplicate measurements from three replications (n = 6).
Theoretical freezing point was )2.15 C for all mixes (standard deviation).
dierent DSC protocol with increasing annealing time

(60 min) showed that the experimental curves matched
the theoretical freezing curves more closely. This
suggests that maximum ice formation is a prerequisite
for more precise estimation of the freezing curve.
All mixes from Table 2, except mix 2, had the same
apparent viscosity as the control at 4 C after aging
(Table 5). The lower viscosity of mix 2 was attributed to
its lower total solid content (29.8%). Alvarez et al.
(2005) reported increased viscosity in mixes containing
MPC (56% and 85% protein) and polydextrose substituted for SMP (at levels of 20% and 50%) with the same
protein content and total solids. Viscosity increased with
increasing substitution (20% and 50%) but not with
increasing protein content in MPC. However, this trend
was not observed in this study.
All mixes were adequately homogenised as indicated
by primarily mono-modal fat particle size distribution
curves and the low D[4,3] and D(v,0.9) values (Table 6). In
the case of the ice creams, the fat particle size distribution curves were bimodal, indicating that partial coalescence took place as was also evident by the change in the
D(v,0.9) and % fat globules >2.3 lm (Table 6). However, there was no signicant dierence among the
dierent treatments. Fat destabilisation values also
showed no statistical dierence among all ice creams
(Table 5).
The time for the rst drip of the melting test took
3040 min for all samples, as this is mainly dependent
on hardness (amount of water frozen) rather than the
internal structure of the ice cream (Koxholt et al. 2001).
All ice creams melted at the same rate as indicated by
their slopes, which ranged between 4.7% and 6.2% mass
loss per minute (Table 5). All ice creams had equally
good stand-up properties, with no ice cream appearing
to have greater shape retention (results not shown).
Many factors can aect the melting rate of ice cream: fat
destabilisation, ice crystal size, and the consistency
coecient of the mix (Muse & Hartel 2004). Except for
Table 6 Fat particle size distribution in mix and ice cream from
formulations described in Table 2
Mix
Control
Mix 1
Mix 2
Mix 3
Mix 4
Mix 5
Ice cream
D[4,3] lm
D(v,0.9) lm
D(v,0.9) lm
% 2.3 lm
2.2ab
1.8ab
1.7b
2.2ab
2.9a
3.3ab
1.2a
2.5ab
1.2a
2.1ab
1.9b
4.3ab
47.5a
42.8a
40.3a
64.1a
44.3a
52.0a
45.8a
44.1a
36.4a
57.2a
44.5a
52.6a
1.08
0.77
0.32
0.39
0.61
1.79
0.16
2.48
0.11
1.47
0.27
4.20
a,b
22.46
3.26
11.03
17.88
8.89
10.98
17.28
9.26
7.43
15.71
9.64
10.46
Means with different letters in the same column are significantly

different. Means from duplicate measurements from three replications
(n = 6) (standard deviation).
the control and mix 3, all ice creams had similar % fat in
drip values (Table 5). An inverse correlation was evident
between % fat in the drip and % fat destabilisation in
the control ice cream and mix 3, as shown by Bolliger
et al. (2000).
All ice creams had similar hardness values, with only
mix 1 and mix 2 being statistically dierent from each
other (Table 5). Hardness in ice cream is inuenced by a
number of factors: ice phase volume, ice crystal size,
overrun, fat destabilisation, and the rheological properties of the mix (Muse & Hartel 2004). No correlation
was evident between hardness, meltdown and experimental freezing curve.
Sensory assessment of ice creams
A ve-member expert taste panel assessed the ice creams

from Table 2. The control was the most liked, with the
strongest dairy and vanilla avour. All experimental ice
creams lacked dairy and vanilla avour. Mix 4 was the
best of the three experimental formulations; however, it
had a slightly unnatural chemical taste. Ice creams
containing polydextrose at levels of 14% were very
bland in taste and were dominated by the polydextrose
with a slight alcohol-like aftertaste. Overall, the experimental ice creams were unacceptable from a sensory
perspective despite the optimisation of the physical
properties. It was concluded to reduce polydextrose
levels and to increase dairy avour to improve acceptability; thus, mixes described in Table 3 were formulated to optimise sensory characteristics. Sensory analysis
was conducted using an untrained panel (n = 70) and
results are presented in Table 7.
Mix A, as expected, was lacking in dairy avour as it
was formulated without milk and cream. In comparison,
the other ice creams were found to have a dairy avour
only slightly lower than the control. The slight loss in
dairy avour in mixes B, C and D was attributed to the
use of MPC instead of SMP (to reduce the lactose
content); therefore, these ice creams have less milk
solids, not fat. All ice creams were perceived as having
less vanilla avour than the control, with mix A the
most pronounced. As the same amount of vanilla was
added to each ice cream and they had the same fat
content, this may suggest that vanilla and dairy avour
are intrinsically linked. Even though all the ice creams
were formulated to have the same sweetness, all were
found to have lower sweetness intensity than the
control. Mix C was the closest to the control followed
by B, D and A. It was interesting that mix A was rated
as having the lowest sweetness intensity even though it
had a much higher SE compared with the others. This
may suggest that sweetness may also be linked to dairy
and vanilla avour.
The sweetness liking and overall liking (Table 7)
followed the same pattern as dairy avour with mix A
1525
Intensity attributes
Sweetness
intensity
Vanilla
flavour
A
B
C
D
)20.1a
)16.4a
)11.2a
)18.2a
)20.4a
)13.2a
)14.4a
)17.0a
26.92
20.77
17.93
18.63
Table 7 Intensity and liking attributes of ice

cream formulations from Table 3
Liking attributes
Formulation
31.99
19.41
21.41
20.68
Dairy
flavour
Sweetness
liking
Overall
liking
)36.7a 21.68
)8.3b 15.21
)5.0b 16.57
)11.4b 17.48
)36.3a 20.72
)4.3b 20.91
)7.3b 22.21
)11.8b 19.27
)44.6a 21.44
)6.5b 19.07
)10.8b 18.52
)14.8b 20.20
Subjects indicated their degree of liking on a 140 mm line scale (70 mm from the centre). Ice
creams were ranked as either less or more intense liked compared with the control, which was
given a fixed value of 0 anchored in the center.
a,b
Means with different letters in the same column are significantly different (n = 70) (standard
deviation).
70
vanilla avour, with vanilla avour perception lacking in the alternative formulations. From the formulations studied here, a combination of tagatose (6%),
polydextrose (6%) and maltitol (3%) or maltitol (15%)
and trehaolse (2.5%) in a fomulation with milk,
cream and MPC shows to be a potential formulation
that could satisfy both physicochemical and sensory
requirements.
61
Number of panelists
1526
35
35
29
24
24
16
10
19
17
2 3 4
27
0
1st
2nd
3rd
4th
Ranking
Figure 2 Ranking of preference between the four alternative
formulations described in Table 3.
Acknowledgements
This study was carried out with the nancial support

of Enterprise Ireland and the South Kerry Development Partnership. The contributions of Mr Dave
Waldron and Mr Denis McMahon are sincerely
appreciated.
References
least liked and mixes B, C and D all ranked as less liked

than the control. From overall liking, it was possible to
see the preference of each individual between the four
experimental ice creams (Fig. 2). When the ice creams
were ranked according to position, mix B was chosen as
rst by 35 of the 70 participants (i.e. 50% of the time),
followed by mix C (34.3%), D (14.3%) and A (1.4%).
Conclusions
This work demonstrated that matching the freezing

curve is thus the rst step to formulate low-GI ice
creams with alternative low-GI, low calorie sugars.
Most physicochemical parameters of interest to ice
cream structure texture will balance the control once the
freezing curve has been matched. The formulation can
then be easily adjusted to match sweetness. It is clear
from the aforementioned results that an acceptable lowGI ice cream cannot be produced without satisfying the
need for dairy and vanilla avour in addition to the
desired sweetness. A strong correlation was found
between perceived dairy avour, sweetness liking and
Abril, J.R., Stull, J.W., Taylor, R.R., Angus, R.C. & Daniel, T.C.
(1982). Characteristics of frozen dessert sweetened with xylitol and
fructose. Journal of Food Science, 47, 472475.
Alvarez, V.B., Wolters, C.L., Vodovotz, Y. & Ji, T. (2005). Physical
properties of ice cream containing milk protein concentrate. Journal
of Dairy Science, 88, 862871.
Bolliger, S., Go, H.D. & Tharp, B.W. (2000). Correlation between
colloidal properties of ice cream mix and ice cream. International
Dairy Journal, 10, 303309.
Bordi, P., Cranage, D., Stokols, J., Palchak, T. & Powell, L. (2004).
Eect of polyols versus sugar on the acceptability of ice cream
among a student and adult population. Foodservice Research
International, 15, 4150.
Bradley, R.L. Jr (1984). Plotting freezing curves for frozen desserts.
Dairy Record, 85, 8687.
Bradley, R.L. Jr & Smith, K. (1983). Finding the freezing points of
frozen desserts. Dairy Record, 84, 114115.
Carabin, I.G. & Flamm, W.G. (1999). Evaluation of safety of inulin
and oligofructose as dietary ber. Regulations in Toxicology and
Pharmacology, 30, 268282.
De Cindio, B., Correra, S. & Ho, V. (1995). Low temperature sugarwater equilibrium curve by a rapid calorimetric method. Journal of
Food Engineering, 24, 405415.
DuBois, G.E. (2000). Non-nutritive sweeteners. In: Encyclopaedia
of Food Science & Technology (edited by F.J. Francis). Pp. 2245
2265. New York, NY: John Wiley & Sons.
Foster-Powell, K., Holt, S.H.A. & Brand-Miller, J.C. (2002). International table of glycemic index and glycemic load values: 2002.
American Journal of Clinical Nutrition, 76, 556.
Go, H.D. & Jordan, W.K. (1984). Aspartame and polydextrose in a
calorie-reduced frozen dairy dessert. Journal of Food Science, 49,
306307.
Go, H.D. & Pearson, A.M. (1983). Aspartame and corn syrup solids
as sweeteners for ice cream. Modern Dairy, 62, 1124.
Go, H.D., Verespej, E. & Jermann, D. (2003). Glass transitions in
frozen sucrose solutions are inuenced by solute inclusions within
ice crystals. Thermochimica Acta, 399, 4355.
Guinard, J.X., Zoumas-Morse, C., Mori, L., Panyam, D. & Kilara, A.
(1996). Eect of sugar and fat on the acceptability of vanilla ice
cream. Journal of Dairy Science, 79, 19221927.
Guven, M. & Karaca, O.B. (2002). The eect of varying sugar content
and fruit concentration on the physical properties of vanilla and
fruit ice-cream-type frozen yogurts. International Journal of Dairy
Technology, 55, 2731.
Jenkins, D.J.A., Wolever, T.M.S. & Taylor, R.H. (1981). Glycemic
index of foods: a physiological basis for carbohydrate exchange.
American Journal of Clinical Nutrition, 34, 362366.
Koxholt, M.M.R., Eisenmann, B. & Hinrich, S.J. (2001). Eect of the
fat globule sizes on the meltdown of ice cream. Journal of Dairy
Science, 84, 3137.
Leighton, A. (1927). On the calculation of the freezing of ice cream
mixes and of the quantities of ice separation during the freezing
process. Journal of Dairy Science, 10, 300308.
Levin, G.V., Zehner, L.R., Saunders, J.P. & Beadle, J.R. (1995). Sugar
substitutes: their energy values, bulk characterises and potential
health benets. American Journal of Clinical Nutrition, 63, 1161
1168.
Livney, T., Verespej, E. & Go, H.D. (2003). On the calculation of ice
cream freezing curves. Milchwissenschaft, 58, 640642.
Marco, G.M. & Pearson, A.M. (1982). Xylitol as an ice cream
sweetener. Modern Dairy, 61, 1618.
Marie, S. (1991). Sweeteners. In: Food Additives Handbook (edited by
J. Smith). Pp. 4774. New York, NY: Blackie Academic.
Marshall, R.T., Go, H.D. & Hartel, R.W. (2003). Ice Cream, 6th edn.
New York, NY: Kluwer Academic Plenum Publishers.
Muse, M.R. & Hartel, R.W. (2004). Ice cream structural elements that
aect rate and hardness. Journal of Dairy Science, 87, 110.
Nabors, L.O. (2001). Alternative Sweeteners, 3rd edn. New York:
Marcel Dekker, Inc.
Newsome, R.L. (1986). Sweeteners nutritive and non-nutritive. Food
Noda, K., Nakayama, K. & Oku, T. (1994). Serum glucose and insulin
levels and erythritol balance after oral administration of erythritol in
healthy subjects. European Journal of Clinical Nutrition, 48, 286292.
Ozdemir, C., Dagdemir, E., Celik, S. & Ozdemir, S. (2003). An
alternative ice cream production for diabetic patients. Milchwissenschaft, 58, 164166.
Patil, D.H., Grimble, G.K. & Silk, D.B.A. (1987). Lactitol, a new
hydrogenated lactose derivative: intestinal absorption and laxative
threshold in normal human subjects. British Journal of Nutrition, 57,
195.
Portmann, M.O. & Birch, G. (1995). Sweet taste and solution
properties of a, a-trehalose. Journal of the Science of Food and
Portmann, M.O. & Kilcast, D. (1996). Psychophysical characterization
of new sweeteners of commercial importance for the EC food
industry. Food Chemistry, 56, 291302.
Rothwell, J. (1985). Diabetic ice cream. Ice Cream and Frozen
Confectionery, 36, 442.
Smith, D.E., Bakshi, A.S. & Lomauro, C.J. (1984). Changes in freezing
point and rheological properties of ice cream mix as a function
of sweetener system and whey substitution. Milchwissenschaft, 39,
455457.
Specter, S.E. & Setser, C.S. (1994). Sensory and physical properties of
a reduced-calorie frozen dessert system made with milk fat and
sucrose substitutes. Journal of Dairy Science, 77, 708717.
Stampanoni Koeferli, C.R., Piccinali, P. & Sigrist, S. (1995). The
inuence of fat, sugar and non-fat milk solids on selected taste,
avour and texture parameters of a vanilla ice cream. Food Quality
and Preference, 7, 6979.
1527
1528
Original article
Effect of glutamate and inosinate on sensory and instrumental
texture of extruded products
Renata Cassar,1 Fatima A. Sardinha2 & Jose A. G. Areas1*
1 Departamento de Nutricao, Faculdade de Saude Publica da Universidade de Sao Paulo, Av. Dr. Arnaldo, 715, 01246-904, Sao Paulo, SP, Brazil
2 Departamento de Nutricao, Universidade Paulista, Rod. Presidente Dutra, KM 157.5, CEP 12240-420, Sao Jose dos Campos, SP, Brazil
(Received 4 April 2005; Accepted in revised form 15 January 2007)
Expanded products have been developed by extrusion of non-conventional highly nutritious raw materials
such as amaranth and chickpea blended with bovine lung. As sensory acceptance of these snacks is restricted,
this study aimed at improving their texture, through the addition of monosodium glutamate (MSG) and
disodium inosinate (IMP) avor enhancers to the feeding material, or to the avor added after the extrusion.
Sensory and mechanical analyses showed that both enhancers aected texture, assessed by sensory and
instrumental methods. Addition of IMP together with MSG to the chickpea-based snacks presented the best
results. This benecial eect was not observed in the amaranth-based snack, suggesting that IMP and MSG
can favorably impact texture of extruded products depending on the amount and type of protein present.
Summary
Keywords
Amaranthus, chickpea, extrusion, ribonucleotides, sodium glutamate, snack-foods.
Introduction
In food production, several unconventional raw materials and by-products of high nutritive value are poorly
used for human feeding because of the lack of tradition
in their consumption, the unfamiliarity of their use or
the lack of basic requirements for their acceptability.
Animal by-products, such as bovine lung that is rich in
proteins of high biological value, bioavailable iron and
vitamins (Areas, 1993; Areas & Lawrie, 1984; Bastos &
Areas, 1990; Bastos et al., 1991; Campos and Areas,
1993; Pinto et al., 1997), some grains as chickpea (Cicer
arietinum L.), that presents the best nutritional quality
among the legume seeds (Chavan & Salunkhe, 1986;
Chavan et al., 1986; Newman et al., 1987; Sotelo et al.,
1987; Fernandez & Berry, 1988; Batistuti et al., 1991;
Pardez-Lopez et al., 1991; Avancini et al., 1992), and
amaranth (Amaranthus caudatus, L), a commodity that
presents great nutritional potential and the capacity for
reducing blood cholesterol (Saunders & Becker, 1984;
Teutonico & Knorr, 1985; Bressani, 1988; Breene, 1991;
Kauman, 1992; Lehmann, 1992, 1996; Bressani et al.,
1993; Plate & Areas, 2002), are good examples.
Extrusion cooking has been one of the most suitable
technological processing for these raw materials,
*Correspondent: Fax: +55-11-30667705;
e-mail: jagareas@usp.br
yielding a variety of ready-to-eat products and food

ingredients. Two new ready-to-eat extruded products
made of chickpea and bovine lung blend (CardosoSantiago & Areas, 2001) and of amaranth (ChavezJauregui et al., 2000) were developed, which proved
useful for nutritional intervention (Castro-Ferreira,
1999; Moreira-Araujo et al., 2002; Cardoso-Santiago
& Areas, 2001; Cardoso-Santiago et al., 2001). Acceptance of these products needs to be improved, mainly
regarding avour and texture that still constituted a
restriction to their wider acceptance.
There are many citations in the literature regarding
the impact of monosodium glutamate (MSG) and
disodium inosinate (IMP) on the avour, acceptance
or preference of snacks (Bellisle et al., 1989, 1991; Filer
& Stegink, 1994; Schiman et al., 1994). However, there
is no record of their eects on the texture of extruded
products. The mechanism by which the supramolecular
structure is formed and therefore the texture is obtained
in the extrusion process is still unclear (Areas, 1992;
Mitchell and Areas, 1992), and there is no way to
forecast the impact of the MSG and IMP on the
network that responds for extrudate texture.
This study evaluated, by means of sensory and
instrumental analyses, the texture characteristics of
extruded products of amaranth and chickpea blended
with bovine lung, to which MSG and MSG + IMP had
been added.
doi:10.1111/j.1365-2621.2007.01548.x
Sensory and instrumental texture of extruded products R. Cassar et al.
Materials
Ajinomoto Inter-American Industry and Commerce Ltd

(Sao Paulo, Brazil) supplied MSG and IMP.
Chickpea (Cicer Arietinum, L) Kabule type, was
acquired from a distributor of foods CEAGESP (Sao
Paulo, Brazil). The chickpea was ground in a knife mill
(Mod. Termomatic, Marconi, Brazil) to produce our.
Bovine lung was supplied by Sadia-Frigobras S.A.
(Parana, Brazil). It was collected from healthy animals
in proper hygienic and sanitary conditions for human
consumption and inspected by the Brazilian Federal
Inspection Service (SIF). The tissue was minced and
frozen immediately after slaughter, and later lyophilised
and milled (Nutribras, Brazil) to produce a powder.
Amaranth (Amaranthus caudatus, L) CAC-43A
variety (Oscar Blanco), was obtained from the germ
plasm bank of the Center of Investigation in Andean
Cultures (CICA), of the National University of San
Antonio Abad del Cusco, Peru. The grains were cleaned
by sieving (20 mesh or 0.84 mm hole), and milled in a
ne-mesh hammer mill (Wilye-MA-090) for the our
production.
The ours of chickpea, amaranth and the lyophilised
bovine lung were individually mixed for 30 min and
defatted (Batistuti et al., 1991) for 12 h in a pilot-scale
glass Soxhlet apparatus, by using ethyl ether for the
chickpea and the lung, and hexane for the amaranth as
the extractor solvents with intermittent dripping. After
defatting, the ingredients were periodically mixed in a
fume cupboard at ambient temperature until full
evaporation of the solvent (12 h), and after that sieved
in ne mesh (0.25 mm).
Methods
Chemical analysis
A triplicate determination was carried out by conventional methods (overnight desiccation to constant
weight at 105 C for moisture; calcination to constant
weight at 550 C for ash; defatting with ethyl ether in a
Soxhlet apparatus for lipids) (Instituto Adolfo Lutz,
1985).
Preparing of raw materials
The MSG and IMP were added to the raw materials

either before the extrusion, being dissolved in the water
used for moisture adjustment, or after the extrusion, in
the snack together with the avours. Based on the
traditional amount used for these ingredients, 0.5%
(w w) of MSG and 0.025% of IMP (5.0% of IMP in
relation to the amount of MSG) were added to the
product, according to Maga (1983). The composition
was balanced to guarantee that these ingredients were
used in the same ratios in the nal products, independent of the addition moment (before or after the
extrusion).
Extrusion
Extrusion of samples was carried out as described by

Cardoso-Santiago & Areas (2001) for chickpea and
bovine lung mixture and by Chavez-Jauregui et al.
(2000) for the amaranth, in a laboratory single-screw
extruder (Miotto Ltda., Sao Paulo, Brazil), with a L D
ratio of 20:1; 20 mm barrel diameter. The conditions for
extrusion were as follows: 200 r.p.m. screw rotation;
3.55:1 screw compression ratio; 4-mm die diameter;
30 r.p.m. feed screw (70 g min)1); 130 C and 150 C
central section temperature for the chickpea and lung
mixture, and for the amaranth, respectively; 13% feed
moisture for the chickpea with lung mixture and 15%
feed moisture for the amaranth.
According to previous studies carried out by CardosoSantiago & Areas (2001) and Chavez-Jauregui et al.
(2000), both the chickpea lung snacks and the amaranth
snacks were avoured with 2.5% of bacon and cheese
avours, respectively (Vittataste Ltda., Sao Paulo,
Brazil). Five samples of each type of snack (chickpea lung; and amaranth) were produced, according to
the moment of addition of the avour enhancers: MSG
added to the feeding material before the extrusion
(MSG-dough); MSG + IMP added to the feeding
material before the extrusion (IMP-dough); MSG added
with the avour after the extrusion (MSG-avour);
MSG + IMP added with the avour after the extrusion
(IMP-avour); and the original standard sample, with
no addition of avour enhancers (control).
Sensory evaluation
The test of multiple comparisons against a control was

chosen for the sensory evaluation. In this test, the
standard sample and the codied samples are presented
to the panel that give values to attributes of texture
typical for snack products (crispness and gumminess)
comparing the codied samples to a standard (Moraes,
1993). The tests were carried out in the Laboratory of
Sensory Analysis of Ajinomoto Interamericana Ltda.
(Sao Paulo, Brazil), with selected panelists recruited
among the company employees, whose sensory accuracy
was determined previously by the threshold test for the
four basic tastes and for the taste umami linked to the
avour enhancers MSG and IMP (Yoshida & Sato,
1969; Yamaguchi & Kimizuka, 1979; Kurihara, 1987;
Yamaguchi, 1979, 1991; Sugita, 1990; Filer & Stegink,
1994; Hettinger et al., 1996). Although taste threshold
tests do not predict perception performance above
threshold, this screening method was used to select the
panelists especially regarding unami perception. Panelists would assess samples with MSG and IMP addition
and they would have to distinguish any avour
1529
enhancing from texture alteration in the samples. The

previous experience of Ajinomoto Co. in this area shows
that this selection is essential to avoid biased texture
reports. The panelists were trained to evaluate specic
texture attributes, but in the texture sensory evaluation
they globally graded texture as a sum of all attributes,
assigning a total texture score in a 17 scale in
comparison to the control sample at the centre of this
scale. Thirty panelists with similar accuracy and ability
to evaluate the samples were selected among Ajinomoto
personnel and trained. To avoid imprecise results
because of repetition and saturation of the panel, the
sensory analyses for this work were performed in four
distinct days in a time span of 2 months. The number of
participants in each session ranged from 24 to 27. After
the texture evaluation in each session, the panelists were
asked to point out the preferred sample and also to write
comments freely. The preferences were calculated as
percentages, considering the number of times that a
sample was preferred (total of panelists = 100%).
Mechanical evaluation
The maximum strength to completely shear the extrudates, which is an important predicting factor of the
sensory acceptance of the product (Voisey & Stanley,
1979), was determined using an Instron universal testing
machine (Mod. 1000, Instron Corp., USA) equipped
with a Warner-Bratzler device. The cross-head speed
was 100 mm min-1 and the pressure transducer maximum capacity was 5 kg (50 N). The expansion ratio was
calculated as the ratio of extruded product diameter and
the diameter of the die hole. Averages of 33 to 66
determinations for each sample were reported, being the
number of replicates (n) indicated when necessary.
before extrusion was the one that received the highest

grade in the global evaluation of texture (Fig. 1).
It has been reported in the literature that the
simultaneous use of IMP and MSG has a synergistic
eect that makes the avour enhancement more ecient
(Sato & Akaike, 1965; Yamaguchi, 1991; Yamamoto
et al., 1991; Kemp & Beauchamp, 1994; Proudlove,
1996). The present results indicate that sensory perception of texture is also signicantly aected by their
addition.
The panelists comments suggest that both the MSG
and the IMP, besides aecting texture, promote also a
signicant improvement in the taste, when compared
with the original product. Analysing the sensory panel
comments, one observes that the IMP-dough amaranth
sample was the most frequently considered spicier and
more tasteful when compared with the other groups,
showing that there was among the panelists a clear
preference to the IMP-dough group, both for the
chickpea lung and amaranth snacks.
To our knowledge, there is no report in the literature
of the eect of avour enhancers on texture of extruded
products. The sensory evaluation of the texture of
chickpea lung snacks showed that the use of MSG and
IMP really promoted a signicant dierence in the
texture when compared with the original standard
samples (Fig. 1). The group of chickpea lung snacks
with MSG + IMP added to the dough, particularly,
presented an average grade for texture (5.63) signicantly higher than the control (4.22) and also signicantly higher than all the other groups averages
(Fig. 1). Table 1 also shows that the texture of the
group of chickpea lung snacks with IMP plus MSG
added in the feeding material (IMP-dough) is
The results were used to calculate the average grade of

taste and texture, to perform the one-way analysis of
variances (P < 0.05), and the multiple comparison tests
(Dunnett t test and Least Signicant Dierence test)
(Stone & Sidez, 1985; Meilgaard et al., 1991).
The MSG and IMP avour enhancing capacity is widely

reported in the literature (Filer & Stegink, 1994; Maga,
1994; Chaudhari et al., 1996; Proudlove, 1996). The
impact of these substances on the acceptance of foods,
including their role as facilitators of the introduction,
rise of preference levels and increase in the consumption
of products has been reported by many authors, for
decades (Kuninaka, 1967; Kato & Nishimura, 1987;
Bellisle et al., 1989; Yamaguchi, 1991; Schiman et al.,
1994; Schiman, 1996).
In this study, the chickpea lung sample added with
IMP plus MSG in the feeding material (IMP-dough)
cb
5.63 (1.33)
b
4.67 (1.54)
Texture scale (control = 4)
1530
a
4.22 (1.09)
b
4.44 (1.53)
4.44 (0.93)
5
4
3
2
1
Control
MSG-dough MSG-aroma IMP-dough
IMP-aroma
Figure 1 Sensory evaluation of 9:1 chickpea bovine lung snacks added

with MSG and IMP in the feeding material or after extrusion. Results
represent averages (n = 27 panelists) of global texture evaluation in a
17 point scale, relative to a standard at the centre value. Dierent
letters indicate signicant dierence among the groups (anova
P < 0.05).
Group (I)
Group (J)
Average
dierence (I-J)
Standard
error
P-value
MSG-dough
MSG-aroma
IMP-dough
IMP-aroma
Control
Control
Control
Control
0.44
0.22
1.41*
0.22
0.36
0.36
0.36
0.36
0.273
0.545
0.000
0.545
*The average difference is significant at the level of 0.05.

The Dunnett t test compares all the groups with the control.
signicantly dierent from the others. In the majority of

the panelists comments, the control group was considered the stickiest and the less crispy among all the
samples.
For the amaranth snacks, the addition of avour
enhancers did not promote signicant change in the
texture as perceived by the sensory analysis, and
therefore, it should not have an impact on the acceptance (Fig. 2). Multiple comparison test presented in
Table 2 shows signicant dierence between samples
where the avour enhancers were added to the feeding
a
b
b
160
a
a
119.2 (16.21)
117.67 (20.35)
106.03
(21.78)
140
108.91 (11.94)
109.55 (16.20)
120
100
80
60
40
20
0
Control MSG-dough MSG-aroma IMP-dough IMP-aroma
n = 58
n = 30
n = 33
n = 33
n = 32
Figure 3 Instrumental texture evaluation of 9:1 chickpea lung snacks

added with MSG and IMP in the feeding material or after extrusion.
Force (N) necessary to completely shear the samples in a WarnerBratzler cell. Number of assays (n) indicated in each bar. Dierent
P < 0.05).
a
250 195.67 (28.08)
c
b
200
Force (N)
Table 1 Multiple comparisons Dunnett t test for the sensory evaluation of the 9:1 chickpea lung snacks added with MSG and IMP in the
feeding material or after extrusion. [Dependent variable: grade;
Dunnett t (<control)]
Force (N)
174.97 (25.38)
c
159.88 (25.17)
134.42 (21.29)
131.84 (20.56)
150
100
50
0
Control MSG-dough MSG-aroma IMP-dough IMP-aroma
Texture scale (control = 4)
n = 66
7
6
5
4
3
2
1
0
4.38 (1.66)
4.33 (1.27)
4.46 (1.32)
4.38 (1.10)
3.88 (1.12)
Control
MSG-dough MSG-aroma IMP-dough
IMP-aroma
Figure 2 Sensory evaluation of pure amaranth snacks. Results represent averages (n = 24 panelists) of global texture evaluation in a 17
point scale, relative to a standard at the centre value. There was no
signicant dierence among the samples (anova P < 0.05).
Table 2 Multiple comparisons Dunnett t test for the mechanical
textural evaluation of the 9:1 chickpea lung snacks added with MSG
and IMP in the feeding material or after extrusion. [Dependent
variable: grade; Dunnett t (<control)]
Group (I)
Group (J)
Average
difference (I-J)
Standard
error
P-value
MSG-dough
MSG-aroma
IMP-dough
IMP-aroma
Control
Control
Control
Control
13.17*
3.51
11.63*
2.87
4.12
3.99
3.99
4.03
0.006
0.823
0.015
0.907

n = 32
n = 38
n = 33
n = 33
Figure 4 Instrumental texture evaluation of pure amaranth snacks

added with MSG and IMP in the feeding material or after extrusion.
Force (N) necessary to completely shear the samples in a WarnerBratzler cell. Number of assays (n) indicated in each bar. Dierent
P < 0.05).
material. However, as far as the texture became less

resistant, the snack became stickier. The samples in the
other groups kept a more resistant structure, being
described sometimes as harder, sometimes as crispy,
which indicates that greater resistance is not always
positive.
Mechanical evaluation of the chickpea lung and
amaranth snacks are presented in Figs 3 and 4, respectively. The strength to shear the chickpea bovine lung
snacks showed that the use of avour enhancers in the
feeding material promoted a signicant physical dierence in the resistance of the product (Fig. 3). The same
was observed through the multiple comparisons t test
(Table 2).
Comparing sensory and mechanical evaluation of
texture (Fig. 1, Table 1, Fig. 3, Table 2), one can observe
that the groups of chickpea bovine lung snacks that
presented the greatest mechanical resistance (MSGdough and IMP-dough) had also been the ones that
presented the highest texture rates (although in the
sensory evaluation, only the IMP-dough has presented
1531
1532
Table 3 Multiple comparisons Dunnett t test for the mechanical

evaluation of pure amaranth snacks added with MSG and IMP in the
feeding material and after extrusion. [Dependent variable: grade;
Dunnett t (<control)]
Group (I)
Group (J)
Average
difference (I-J)
Standard
error
P-value
MSG-dough
MSG-aroma
IMP-dough
IMP-aroma
Control
Control
Control
Control
)63.82*
)20.69*
)61.24*
)35.79*
5.38
5.09
5.33
5.33
0.000
0.000
0.000
0.000

signicant results in relation to the other groups). This

increased resistance was perceived by the sensory panel
only in the group IMP-dough and was considered positive
by the panelists comments. The best global texture
assessment for these snacks occurred in the samples in
which higher strengths were necessary to shear the
product. Batistuti et al. (1991), in previous studies,
reported a linear correlation between the sensory preference and the shear strength for snacks formulated
exclusively with chickpea.
Although the dierences observed in the sensory
evaluation of the texture for amaranth snacks have not
been signicant, the mechanical assessment of this
attribute by the shear strength test showed that addition
of MSG or IMP, despite occurring either to the feeding
material or to the avouring mixture, caused a signicant physical change in these snacks (Fig. 4 and
Table 3), making them less resistant than the control.
The greatest decrease was observed in the samples that
received the avour enhancers in the feeding material.
As far as the resistance decreased (MSG-dough and
IMP-dough groups), the sensory texture assessment was
positive. The panelists comments indicate that in this
product, a less resistant texture was regarded as more
adequate.
The observation that while the addition of MSG and
IMP to the chickpea bovine lung snacks feeding materials resulted in a more resistant structure, a less
resistant snack was observed for the same addition to
amaranth (Figs 3 and 4, respectively), suggesting the
existence of an interaction between the enhancer substances and the raw material chosen that inuences
texture characteristics of the nal product. The amaranth snack network is amylaceous (70% of starch and
less than 15% of proteins), highly expanded with a
fragile three-dimensional structural network, desegregating itself easily with the humidity. The protein network
formed upon extrusion cooking of amaranth does not
aect signicantly the nal texture (Chavez-Jauregui
et al., 2000). On the contrary, the chickpea bovine lung
snack (20% of protein and 60% of starch) is less
expanded, presenting a more entangled and resistant
protein network and a more complex structure, whose

mechanism of formation has not yet been completely
claried (Areas, 1992; Mitchell and Areas, 1992;
Cardoso-Santiago & Areas, 2001).
We can assume from the results that the MSG and the
IMP aect the way by which the protein binds in the
chickpea bovine lung snacks network, favourably modifying their structure (the IMP is a ribonucleotidephosphate; and polyphosphates usually improve the
structure of protein nets), and increasing its resistance.
The eects of the addition of MSG and IMP observed
on the structure of the studied products, as well as of the
expanded products (snacks) in general, need further
clarications, deserving specic further studies.
Acknowledgments
This study was supported by FAPESP (grants 98 080959 and 03 10246-5) and CNPq (grant 83 1231). The
authors are thankful to Ajinomoto Interamericana Ltd
for supplying raw materials and sensory analysis installations; to Dr Raquel de A. C. Santiago and Dr Rosa N.
Chavez-Jauregui for the technical support; and to Jose
P. Silva and Luciana Zanella for the support on snack
production.
References
Areas, J.A.G. (1992). Extrusion of food proteins. CRC Critical
Reviews in Food Science and Nutrition, 32, 365392.
Areas, J.A.G. (1993). Uso de materias-primas nao convencionais na
composicao de dietas especiais. Cadernos de Nutricao, 6, 1115.
Areas, J.A.G. & Lawrie, R.A. (1984). Eect of lipid-protein interactions on extrusion of oal protein isolates. Meat Science, 11, 275
299.
Avancini, S.R.P., Sales, A.M., Aguirre, J.M. & Mantovani, D.M.B.
(1992). Composicao qu mica e valor nutricional de cultivares de
grao-de-bico produzidos no Estado de Sao Paulo. Coletanea ITAL.,
22, 145153.
Bastos, D.H.M. & Areas, J.A.G. (1990). Lung proteins: eect of
defatting with solvents and extrusion cooking on some functional
properties. Meat Science, 28, 223235.
Bastos, D.H.M., Domenech, C.H. & Areas, J.A.G. (1991). Optimization of extrusion cooking of lung proteins by response surface
methodology. International Journal of Food Science and Technology,
26, 403408.
Batistuti, J.P., Barros, R.M.C. & Areas, J.A.G. (1991). Optimization
of extrusion cooking process for chickpea (Cicer arietinum, L.)
defatted our by response surface methodology. Journal of Food
Science, 56, 16951698.
Bellisle, F., Tounier, A. & Louis-Sylvestre, J. (1989). Monosodium
glutamate and the acquisition of food preferences in a European
context. Food Quality and Preference, 1, 103108.
Bellisle, F., Monnause, M.O., Chabert, M., Larue-Achagiotis, C.,
Lanteaume, M.T. & Louis-Sylvestre, J. (1991). Monosodium
glutamate as a palatability enhancer in the European diet.
Physiology & Behavior, 49, 869873.
Breene, W.M. (1991). Food uses of grain amaranth. Cereal Foods
World, 36, 426430.
Bressani, R. (1988). Amaranth: the nutritive value and potential
uses of grain and by products. Food and Nutrition Bulletin, 10,
4959.
Bressani, R., De Martell, E.C.M. & De Gordinez, C.M. (1993). Protein

quality evaluation of amaranth in adult humans. Plant Foods for
Human Nutrition, 43, 123143.
Campos, M.A. & Areas, J.A.G. (1993). Protein nutricional value
of extrusion-cooking defatted lung our. Food Chemistry, 47,
6166.
Cardoso-Santiago, R.A. & Areas, J.A.G. (2001). Nutritional evaluation of snacks obtained from chickpea and bovine lung blends. Food
Chemistry, 74, 3540.
Cardoso-Santiago, R.A., Moreira-Araujo, R.S.R., Pinto e Silva,
M.E.M. & Areas, J.A.G. (2001). The potential of extruded chickpea,
corn and bovine lung for malnutrition programs. Innovative Food
Science & Emerging Technologies, 2, 203209.
Castro-Ferreira, T.A.P. (1999). Avaliacao nutricional do amaranto
(Amaranthus caudatus L.) extrusado em diferentes condicoes de
umidade. Pp. 154. Sao Paulo: PhD Thesis, Faculdade de Ciencias
Farmaceuticas, Universidade de Sao Paulo.
Chaudhari, N., Yang, H., Lamp, C., Delay, E., Cartford, C. & Than,
T. (1996). The taste of monosodium glutamate: membrane receptors
in taste buds. The Journal of Neuroscience, 16, 38173826.
Chavan, J.K. & Salunkhe, D.K. (1986). Biochemistry and technology
of chickpea (Cicer arietinum L.) seeds. CRC Critical Reviews in Food
Science and Nutrition, 25, 107.
Chavan, J.K., Kadan, S.S. & Salunkhe, D.K. (1986). Biochemistry and
technology of chickpea (Cicer arietinum L.) seeds. Critical Reviews in
Food Science and Nutrition, 25, 107158.
Chavez-Jauregui, R.N., Pinto e Silva, M.E.M. & Areas, J.A.G. (2000).
Extrusion cooking process for amaranth (Amaranthus caudatus L.).
Journal of Food Science, 65, 10091015.
Fernandez, M.L. & Berry, J.W. (1988). Nutritional evaluation of
chickpea and germinated chickpea ours. Plant Foods for Human
Nutrition, 38, 127132.
Filer, L.J. & Stegink, L.D. (1994). A report of the proceedings of an
MSG workshop held august 1991. Critical Reviews in Food Science
and Nutrition, 34, 159174.
Hettinger, T.P., Frank, M.E. & Meyers, W.E. (1996). Are the tastes of
polycose and monosodium glutamate unique? Chemical Senses, 21,
341347.
Instituto Adolfo Lutz (1985). Normas analticas do Instituto Adolfo
Lutz, Vol. 1, 3rd edn Pp. 533. Sao Paulo: Instituto Adolfo Lutz.
Kato, H. & Nishimura, T. (1987). Taste components and conditioning
of beef, pork, and chicken. In: Umami: A Basic Taste (edited by
Y. Kawamura & M.R. Kare). Pp. 289. New York, NY: Marcel
Dekker.
Kauman, C.S. (1992). The status of grain amaranth for the 1990s.
Food Reviews International, 8, 165185.
Kemp, S.E. & Beauchamp, G.K. (1994). Flavor modication by
sodium chloride and monosodium glutamate. Journal of Food
Science, 59, 682686.
Kuninaka, A. (1967). Flavor potentiators. In: The Chemistry and
Physiology of Flavors (edited by H.W. Schultz, E.A. Day & L.M.
Libbey). Pp. 515535. Westport: AVI Publ.
Kurihara, K. (1987). Recent progress in the taste receptor mechanism.
In: Umami: a basic taste (edited by Y. Kawamura & M.R. Kare).
Pp. 339. New York, NY: Marcel Dekker.
Lehmann, J.W. (1992). Grain amaranth developing 21st century
products from the legendary sister crop of maize. International Food
Ingredients, 3, 2635.
Lehmann, J.W. (1996). Specialty grains: history of grain amaranth as
an alternative crop. American Association of Cereal Chemists, 41,
399411.
Maga, J.A. (1983). Flavor potentiators. Critical Reviews in Food
Maga, J.A. (1994). Flavor potentiators. In: Food additive toxicology
(edited by J.A. Maga & A.T. Tu). Pp. 379412. New York, NY:
Marcel Dekker.
Meilgaard, M., Civille, G.V. & Carr, B.T. (1991). . Sensory Evaluation
Techniques, 2nd edn, Pp. 354 Boca Raton, FL: CRC Press.
Mitchell, J.R. & Areas, J.A.G. (1992). Structural changes in

biopolymers during extrusion. In: Food Extrusion Science and
Technology, Chapter 22. (edited by J.L. Kokini, C.T. Ho & M.V.
Karwe). Pp. 345360. New York, NY: Marcel Dekker, Inc.
Moraes, M.A.C. (1993). Metodos para avaliacao sensorial dos alimentos, 8th edn Pp. 85. Campinas, SP: UNICAMP.
Moreira-Araujo, R.S.R., Araujo, M.A.M., Silva, A.M.S., Carvalho,
C.M.R. & Areas, J.A.G. (2002). Impacto de salgadinho de alto valor
nutricional de criancas de creches municipais de Teresina-PI.
Nutrire, 23, 721.
Myers, R.H. (Ed) (1971). Response Surface Methodology, 1st edn.
Boston, MA: Allyn and Bacon Publishers.
Newman, C.W., Roth, N.J., Newman, R.K. & Lockerman, R.H.
(1987). Protein quality of chickpea (Cicer arietinum L.). Nutrition
Reports International, 36, 18.
Pardez-Lopez, O., Ordorica-Falomir, C. & Olivares-Vasquez, M.R.
(1991). Chickpea protein isolates: physicochemical, functional and
nutritional characterization. Journal of Food Science, 56, 726729.
Pinto, T.A., Colli, C. & Areas, J.A.G. (1997). Eect of processing on
iron bioavailability of extruded bovine lung. Food Chemistry, 60,
459463.
Plate, A.Y.A. & Areas, J.A.G. (2002). Cholesterol-lowering eect of
extruded amaranth (Amaranthus caudatus L.) in hypercholesterolemic rabbits. Food Chemistry, 76, 16.
Proudlove, R.K. (1996). Os alimentos em debate: uma visao equilibrada.
Pp. 19, 207, 209, 22, 224. Trad. de AT Giova. Sao Paulo: Varela.
Sato, M. & Akaike, N. (1965). 5ribonucleotides as gustatory stimuli in
rats: electrophysiological studies. Japanese Journal of Physiology, 15,
5370.
Saunders, R.M. & Becker, R. (1984). Amaranthus: a potential food
and feed resource. Advances in Cereal Science and Technology, 6,
357396.
Schiman, S.S. (1996). Update on monosodium glutamate: sensory
properties and safety. Nutrition, 12, 451452.
Schiman, S.S., Sattely-Miller, E.A., Zimmerman, I.A., Graham, B.G.
& Erickson, R.P. (1994). Taste perception of monosodium
glutamate (MSG) in foods in young and elderly subjects. Physiology
& Behavior, 56, 265275.
Sotelo, A., Flores, F. & Hernandez, M. (1987). Chemical composition
and nutritional value of Mexican varieties of chickpea (Cicer
arietinum L.). Plant Foods for Human Nutrition, 37, 299306.
Stone, H. & Sidez, J.L. (1985). Sensory Evaluation Practices. Pp.
243270. New York, NY: Academic Press.
Sugita, Y. (1990). Flavor enhancers. In: Food Additives (edited by A.L.
Branen, P.M. Davidson & S. Salminen). Pp. 259. New York, NY:
Marcel Dekker.
Teutonico, R.A. & Knorr, D. (1985). Amaranth: composition properties and applications of a rediscovered food crop. Food Technology,
39, 4460.
Voisey, P.W. & Stanley, D.W. (1979). Interpretation of instrumental
results in measuring bacon crispness and brittleness. Canadian
Institute of Food Science and Technology Journal, 12, 712.
Yamaguchi, S. (1979). The umami taste. In: Food Taste Chemistry.
(edited by J.C. Boudreau). Pp. 33. Washington, DC: American
Chemical Society.
Yamaguchi, S. (1991). Basic properties of umami and eects in
humans. Physiology & Behavior, 49, 833841.
Yamaguchi, S. & Kimizuka, A. (1979). Psychometric studies on the
taste of monosodium glutamate. In: Glutamic Acid: Advances in
Biochemistry and Physiology (edited by L.J. Filner, S. Garattini,
M.R. Kare, W.A. Reynolds & R.J. Wurtman). Pp. 35. New York,
NY: Raven Press.
Yamamoto, T., Matsuo, R., Fujimoto, Y., Fukunaga, I., Miyasaka, A.
& Imoto, T. (1991). Electrophysiological and behavioral studies on
the taste of umami substances in the rat. Physiology & Behavior, 49,
919925.
Yoshida, M. & Sato, S. (1969). Multidimensional sealing of the taste of
amino acids. The Japanese Psychological Research, 11, 149153.
1533
1534
Original article
Physicochemical and sensory characteristics of cookies containing
residue from king palm (Archontophoenix alexandrae) processing
Manoela A. Vieira, Karina C. Tramonte, Rossana Podesta, Sandra R. P. Avancini,
Renata D. de M. C. Amboni & Edna R. Amante*
Departamento de Ciencia e Tecnologia de Alimentos, Centro de Ciencias Agrarias, Universidade Federal de Santa Catarina, Rodovia Admar
Gonzaga 1346, Itacorubi, 88034-001 Florianopolis, SC, Brazil
(Received 13 July 2006; Accepted in revised form 8 February 2007)
Summary
The aim of this study was to evaluate the chemical properties of the ours prepared with residues from
organic king palm processing, and also to determine the cookie-making performance of residue blends from
organic king palm processing and wheat our, as well as the eect of the blends on the consumers
acceptance and purchase intent of high-bre cookies. The king palm ours (PFs) contained high contents of
total dietary bre and total ash. Blends containing 0%, 10%, 15%, 20% and 25% of either PF or sieved
king-palm our (SPF) replacing wheat our were prepared. The total dietary bre content of the cookies
ranged from 4 to 7 g (100 g))1 on a dry-matter basis. The level of these components improved with increased
amounts of PF and SPF in the blends. All the cookies were acceptable and approved in relation to purchase
intent.
Keywords
Acceptability, chemical composition, cookies, dietary bre, king palm, residue.
Introduction
Archontophoenix alexandrae (Wendl. & Drude), commonly known as king palm or as Alexandra palm or
even as Seafortia palm, is endemic to the tropical forests
of east Australia. It produces heart-of-palm or palmito
of the noble type, with higher quality and superior
avour compared with that of Euterpe oleracea Mart.
(aca , or acaizeiro), which supplies more than 80% of
the palmito traded in the international market (Bovi,
1998).
As an exotic plant, the king palm has a strong
ecological connotation, because it reduces the pressure
over other palmito producer species which are often
illicitly extracted, such as the Euterpe edulis. Additionally, it can be cultivated without pesticides and chemical
products, with good adaptation to organic cultivation
system (Tagliari, 1999).
A wide variety of bre sources have been developed
for use in various foods to provide more bre. There are
medical studies about the benets of total dietary bres
consumption such as falling serum cholesterol concentration, lowering the risk of coronary heart disease,
reducing blood pressure, aiding weight control, improving glycemic control, reducing the risk of certain types of
*Correspondent: Fax: +55 48 3331 99 43;
e-mail: eamante@cca.ufsc.br
cancer and improving gastrointestinal functions

(Bonithon-Kopp et al., 2000; Terry, 2001; Bingham,
2003; Ferguson & Harris, 2003; Peters, 2003; Tames
et al., 2003; Skea et al., 2005). As a result, bres from
dierent sources and compositions are being obtained,
and the fortication of foods with dietary bres is
increasing.
A considerable increase of king palm production as a
result of palmito industrialisation produces a big quantity of residues constituted of brous material from
leaves and leaf sheaths, which constitute from 80% to
90% of total raw material, with some variations
according to species (Vieira et al. 2003). These residues
could be a source of natural dietary bres. However,
there is no information available in the to literature on
any chemical composition of this material, or about its
application as a food ingredient. One potential food
application for this residue could be its use in composite
ours for the production of bakery products, such as
bread and cookies.
The concept of using composite ours is not new and
has been the subject of numerous studies. Studies have
reported that acceptable wheat products can be made
with as much as 10% Hymenaea stigonocarpa Mart. and
Hymenaea stilbocarpa Mart. ours (Silva et al., 2001),
50% fonio our (McWatters et al., 2003), 525% African
breadfruit kernel our (Akubor & Badifu, 2004), 515%
substitution with uted pumpkin our (Giami et al.,
doi:10.1111/j.1365-2621.2007.01568.x
Chemical properties of cookies containing king palm residue M. A. Vieira et al.
2005) and 515% soy protein isolate (Lee & Brennand,

2005). Cookies have been suggested as a better use for
composite our than bread because of their ready-to-eat
form, wide consumption, relatively long shelf life and
good eating quality (Tsen et al., 1973). No reports of
studies were found for combinations of residues from
king palm processing and wheat our. Thus, the aim of
this study was to evaluate the chemical properties of the
ours prepared with residue from organic king palm
processing, and also to determine the cookie-making
performance of blends of residue from organic king palm
processing and wheat our and the eect of the blends
on the consumers acceptance and purchase intent of
high-bre cookies.
Residues from organic king palm

processing
Washing
and drying (60 C)
Milling
(42 mesh)
Sieving
(60 mesh)
King palm flour
(PF)
Sieved king palm flour

(SPF)
Figure 1 Preparation of king palm our.
Materials
Residues from organic king palm processing were

supplied by the king palm processing industrial plants
in Florianopolis (Brazil). Commercial wheat our having 11.2% moisture, 10.3% protein, 2.0% total dietary
bre and 0.45% ash on a dry-matter (d.m.) basis was
used in the study. Wheat our and the other ingredients
were purchased at a local supermarket in Florianopolis,
SC, Brazil. All the reagents used were of analytical
grade.
Preparation of king palm flours
The residues from organic king palm processing were

washed and oven-dried (60 C, 48 h) in an air-forced
oven (Model 171, FABBE, Sao Paulo, Brazil). The dried
raw material was then milled in a hammer mill (MunchEdelstahl, Wuppertal, Germany, licensed by Bliss, USA)
to a 42-mesh sieve powder (Bertel, Sao Paulo, Brazil),
producing king palm our (PF). Part of the PF was
screened to pass through a 60-mesh sieve (Bertel, Sao
Paulo, Brazil), producing sieved king-palm our (SPF).
The PF and SPF samples were packaged in an airtight
plastic bag and stored in a freezer ()18 2 C) until
required (Fig. 1).
High-fibre cookies formulation and preparation
Blends containing 0%, 10%, 15%, 20% and 25% PF or

SPF replacing wheat our were prepared by gradual
mixing in a rotary mixer at full speed for 5 min.
Cookies were prepared using the American Association of Cereal Chemists method 1050 D (AACC,
2000), with slight modications. Each treatment (our
blend) was baked in three batches. The same measuring
apparatus, mixing bowls and electrical mixers were used
for preparing the dierent formulations. The cookies
were prepared by creaming together hydrogenated
vegetable shortening (52.0 g), brown sugar (50.0 g) for

5 min in a high speed mixer. The mixture of the ours
(112.5 g), salt (1.1 g), baking powder (2.5 g), vanilla
(1.0 g), cocoa powder (2.5 g) and water (110.0 mL) were
added to the creamed mixture to form a soft dough. The
cookie dough was sheeted to a thickness of 0.4 cm and
cut into circular shapes of 3.5 cm diameter using a
circular scone cutter. The cut-out pieces of dough were
baked on greased pans at 200 C for 10 min in a
conventional electric oven. The cookies were cooled at
room temperature (28 2 C), weighed and packed in
high density polyethylene bags. Cookies from the each
batch were then analysed for physical and chemical
characteristics, and consumers acceptance.
Chemical analyses
All the our samples and cookies were analysed, in

triplicate (n = 3), for moisture (AOAC Method 925.09),
lipid (AOAC Method 920.85), crude protein (N 6.25)
(AOAC Method 920.87), total ash contents (AOAC
Method 923.03), soluble dietary bre and insoluble
dietary bre content (AOAC Method 991.43) (AOAC,
2005). The total carbohydrate was calculated by dierence (% moisture + % ash + % lipid + % protein).
Energy values (Kcal) were calculated by applying factors
4, 9 and 4 for each gram of protein, lipid and carbohydrate, respectively (Watt & Merrill, 1999).
The amino acid proles of the our samples were also
evaluated. Amino acids were determined by separation
with ion exchange chromatography, followed by postcolumn reaction with ninhydrin, using a Dionex DX 300
analyser according to the method described by Spackman & Stein (1958).
The water absorption capacities of wheat our, PF
and SPF were determined as described by Sosulski et al.
(1976). A gram of the our sample was mixed with
1535
1536
10 mL of distilled water in a centrifuge tube and allowed

to stand at 30 2 C for 1 h. Thereafter, it was
centrifuged (200 g, 30 min.). Water absorption capacity was expressed as per cent water absorbed by 1 g
our on a dry-matter basis.
Physical characteristics of cookies
Ten cookies from each batch were weighed (g); their

height (cm) and diameter (cm) were measured with
Vernier callipers (Metrica, Germany). Spread ratio was
expressed as diameter height. The cookies yield was
calculated by dierence between the weight of cookies
before and after cooking. Specic bulk volume of the
cookies was determined by the rape seed displacement
method (AACC, 1995).
Consumer acceptance and purchase intent of
high-fibre cookies
Consumer acceptance test was performed according to

the methods described by Meilgaard et al. (1999) to
evaluate the overall acceptability of cookie samples that
had been stored for 7 days in high density polyethylene
bag at room temperature (28 2 C). A nine-point
hedonic scale ranging from like extremely to dislike
extremely, corresponding to the highest and lowest
scores of 9 and 1, respectively, was used. This test was
carried out by 100 volunteers who were habitual cookie
consumers (men and women, between 17 and 62 years
old).
The Purchase Intent was also evaluated on a vepoint scale, denitely would buy to denitely would
not buy corresponding to the highest and the lowest
scores of 5 and 1, respectively (Meilgaard et al., 1999).
All analytical determinations were carried out in triplicate. Mean SD values were calculated and the data
were subjected to analysis of variance. If a signicant
F-test was noted, means were separated using Tukey
multiple range test. Signicance was accepted at
P 0.05.
Chemical analysis of king palm flour
The chemical composition of PF and SPF are shown in

Table 1. The higher moisture content of PF in relation
to SPF may be because of the greater water binding
properties of PF. The water absorption capacities for
PF, SPF and wheat our were 61, 58 and 23 g water
per 100 g sample, respectively. Rossel et al. (2001)
reported that the dierences in water absorption is
Table 1 Chemical composition of king palm our (PF) and sieved

king-palm our (SPF) on a dry-matter basis
Components
Means SD (n = 3)
PF
Moisture (%)
Total ash (%)
Lipid (%)
Crude protein (n 6.25) (%)
Total dietary fibre (%)
Soluble fibre (%)
Insoluble fibre (%)
Carbohydrate* (%)
Energy (kcal (100 g))1)
SPF
8.66
5.01
1.56
6.43
70.56
3.20
67.22
78.35
353
0.03a
0.02a
0.04a
0.06a
0.01a
0.02a
0.02a
0.03a
0.5a
8.02
5.44
1.92
6.78
65.42
2.39
63.05
77.85
356
0.03b
0.03b
0.06b
0.06b
0.05b
0.03b
0.02b
0.03b
0.4b
Mean values in the same row followed by different superscript alphabets

are significantly different (P 0.05). Mean values SD of triplicate
determinations.
*Values calculated by difference.
mainly caused by a greater number of hydroxyl groups

which exist in the bre structure and allow more water
interaction through hydrogen bonding. Similar results
were found by Silva et al. (1998), Kruger et al. (2003)
and Sudha et al. (2006).
Sieved king-palm our contained higher amounts of
protein, lipid and total ash than PF, whereas PF
contained higher amounts of bre and carbohydrate.
SPF and PF showed higher contents of total dietary
bre than those reported of other ours, such as whole
wheat our (10.7% d.m.) (Chaudhary & Weber, 1990),
rye (21.90% d.m.) (Filizette-Cozzi & Lajolo, 1991), oat
bran (20.4% d.m.), wheat bran (47.5% d.m.), rice bran
(40.0% d.m.) and barley bran (45.0% d.m.) (Sudha
et al., 2006). The consumption of about 10 g of SPF or PF
would provide about 26.17% and 28.22%, respectively, of
the recommended daily requirement for dietary bre
(25 g day)1), as recommended by FAO WHO (1973).
The total ash content of PF and SPF samples was
higher than the ash content reported for wheat our
(0.7% d.m.) (Giami et al., 2005), oat bran (4.00% d.m.)
and barley bran (5.00% d.m.) (Sudha et al., 2006).
Lipid contents of PF and SPF were lower when
compared with cereals as wheat (2.63.8% d.m.), corn
(3.95.8% d.m.), barley (3.34.6% d.m.) and rye
(2.73.5% d.m.) according to the values reported by
Morrison (1998).
The energy values of SPF and PF are higher than those
of the conventional ours, such as whole wheat our and
defatted soy our, 339 and 329 kcal (100 g))1, respectively, and similar to that of whole yellow corn our, which
is 365 kcal (100 g))1,according to the Nutrient Database
for Standard Reference (USDA, 2002).
Amino acid values of PF and SPF are shown in
Table 2. The data for each amino acid are presented on
a dry-matter basis. Aspartic acid, glutamic acid and
leucine were the predominant amino acids for two kinds
Table 2 Amino acid composition of King palm our (PF) and

sieved king-palm our (SPF) in g (100 g))1 protein and g (100 g))1
our on a dry-matter basis
PF
SPF
g (100 g)
Sample
Aspartic acid
Threoninec
Serine
Glutamic acid
Proline
Glycine
Alanine
Valinec
Cystine
Methionine
Cystine + Methionine
Isoleucinec
Leucinec
Tyrosinec
Phenylalaninec
Lysinec
Histidinec
Arginine
)1
Proteina
0.71
0.36
0.39
0.90
0.42
0.42
0.47
0.40
0.03
0.01
0.04
0.30
0.64
0.21
0.40
0.41
0.14
0.43
Proteinb
0.74
0.38
0.41
0.94
0.44
0.45
0.49
0.41
0.05
0.03
0.08
0.33
0.68
0.21
0.44
0.42
0.17
0.44
Sample
11.04
5.60
6.06
14.00
6.53
6.53
7.31
6.22
0.47
0.15
0.62
4.47
9.95
3.27
6.22
6.38
2.18
6.69
10.91
5.91
6.04
13.86
6.49
6.64
7.23
6.04
0.77
0.47
1.24
4.87
10.03
3.10
6.49
6.19
2.51
6.49
6.43% protein level.

6.78% protein level.
c
Essential amino acid.
b
of ours from palmito production solid residues. SPF

contained relatively higher concentrations of dierent
amino acids when compared with PF. In the comparison
of the amino acid composition of PF and SPF to the
FAO WHO (1990) pattern, all samples contained higher
amino acids in g (100 g))1 protein than the standard,
except for a deciency in sulphur amino acids
(cystine + methionine).
Chemical composition of cookies
The proximate composition of cookies prepared by

using blends of wheat our with any of the two dierent
types of king palm our (PF and SPF) is shown in
Table 3. The fat, ash and total dietary bre contents of
the cookies signicantly increased (P < 0.05) with
respect to the control. This was due to an addition
eect, as PF and SPF are higher in these nutrients in
comparison with wheat our.
The ash content found in the cookies was higher than
that reported for conventional cookies (Shrestha &
Noomhorm, 2002; Giami, et al. 2005). Fortication of
cookies with PF and SPF signicantly (P < 0.05)
increased the levels of total ash.
The energy values of cookies ranged from 466 to
468 kcal (100 g))1, about 22% of the requirement for
energy (17902500 kcal day)1), as recommended by
FAO WHO (1973), for children aged between 5 and
19 years. The energy values found in the cookies
supplemented with PF and SPF were within the range
reported for cookies prepared with wheat-uted
pumpkin seed (Giami et al., 2005), and wheat-soy
protein isolate our blends (Shrestha & Noomhorm,
2002).
Substitution of wheat our for PF and SPF caused a
signicant increase in bre content of cookies
(P < 0.05). The consumption of 100 g day)1 of these
cookies would represent around 22% of the recommended daily requirement for dietary bre
(25 g day)1), as recommended by FAO WHO (1973).
Total dietary bre values of all types of cookies ranged
from 3.73 to 7.10 g (100 g))1. These levels were within
the range reported for high-bre cookies (Silva et al.,
2001; Shrestha & Noomhorm, 2002) and higher than
the levels reported for other cookies (McWatters et al.,
2003; Giami et al., 2005; Guilherme & Jokl, 2005).
Table 3 Chemical composition of cookies fortied with king palm our (PF) and sieved king-palm our (SPF) on a dry-matter basis
Types of cookies
Lipid
Total ash
TDF
Moisture
Carbohydrate*
Kcal (100 g))1
0.02a
19.33 0.03e
2.14 0.03h
2.70 0.02i
5.13 0.03g
69.25 0.09a
468 0.1a
0.03bc
0.03c
0.02d
0.02e
19.43
19.46
19.57
19.61
0.03c
0.03bc
0.03a
0.03a
2.22
2.32
2.42
2.52
0.01g
0.03ef
0.01cd
0.03ab
3.89
4.99
6.03
7.10
0.02g
0.01e
0.03c
0.02a
5.20
5.32
5.41
5.50
0.03ff
0.03de
0.02bc
0.03a
69.09
68.90
68.66
68.49
0.08b
0.05c
0.02d
0.02e
467
467
467
466
0.2ab
0.3ab
0.2ab
0.2c
0.01ab
0.01c
0.02d
0.01de
19.37
19.41
19.50
19.63
0.02de
0.01cd
0.02b
0.01a
2.26
2.36
2.47
2.57
0.02fg
0.02de
0.02bc
0.03a
3.73
4.72
5.71
6.71
0.03h
0.02f
0.03d
0.01b
5.19
5.28
5.37
5.45
0.02fg
0.01e
0.02cd
0.03ab
69.08
68.91
68.71
68.44
0.03b
0.02c
0.05d
0.01e
467
466
466
466
0.1ab
0.2c
0.4c
0.5c
Crude protein
Control
4.16
King palm flour (%)
10% PF
4.06
15% PF
4.01
20% PF
3.94
25% PF
3.88
Sieved king-palm flour (%)
10% SPF
4.11
15% SPF
4.05
20% SPF
3.96
25% SPF
3.92
Mean values in the same column followed by different superscript alphabets are significantly different (P 0.05).
Mean values SD of triplicate determinations.
TDF, total dietary fibre.
*By difference.
1537
1538
Table 4 Physical characteristics of cookies fortied with king palm our (PF) and sieved king-palm our (SPF)
Types of cookies
Yield
Control
0.74 0.02a
King palm flour
PF 10%
0.74 0.03a
PF 15%
0.73 0.02a
PF 20%
0.73 0.02a
PF 25%
0.73 0.01a
Sieved king-palm flour
SPF 10%
0.74 0.02a
SPF 15%
0.73 0.02a
SPF 20%
0.73 0.02a
SPF 25%
0.73 0.02a
Specific bulk
volume [cm3 (g))1]
Weight (g)
Diameter
(D) (cm)
Height (H) (cm)
Spread radio
(D H)
1.44 0.03a
3.62 0.02g
3.43 0.02a
0.63 0.02a
5.42 0.02a
1.39
1.39
1.24
1.25
0.03b
0.03b
0.01d
0.03d
3.64
3.65
3.67
3.70
0.02de
0.03cd
0.02bc
0.01a
3.17
3.06
3.00
2.84
0.01cd
0.02ef
0.01f
0.03g
0.60
0.58
0.57
0.55
0.02bc
0.03de
0.01ef
0.04f
5.28
5.27
5.26
5.21
0.02cd
0.02cde
0.01cde
0.05f
1.43
1.43
1.31
1.25
0.03a
0.02a
0.03c
0.06d
3.63
3.64
3.65
3.68
0.01fg
0.03de
0.03cde
0.04b
3.41
3.33
3.20
3.10
0.02a
0.01b
0.02c
0.06de
0.63
0.62
0.60
0.59
0.02a
0.05ab
0.02bc
0.02cd
5.41
5.37
5.33
5.25
0.06a
0.02ab
0.05bc
0.01de
Mean values in the same column followed by different superscript alphabets are significantly different (P 0.05).
Mean values SD of ten samples from each batch.
Physical characteristics of the cookies
Data on the physical characteristics of the cookies are

shown in Table 4. There were no signicant (P > 0.05)
dierences between the values obtained for the yield of
cookies supplemented with PF and SPF.
In general, the cookies made by blending PL with
wheat our had reduced heights, diameters, spread
radios and specic bulk volume; and they also had
increased weights. These eects increased when the level
of replacement of wheat our with PL was increased.
These results were similar to those reported for cookies
prepared with wheat-jatoba (Silva et al., 1998),
wheat-cowpea (McWatters et al., 2003) and wheatpumpkin seed (Giami et al., 2005) our blends.
The spread ratio of cookies gradually decreased to
5.21 for PF and to 5.25 for SPF as the content of these
ours was increased to 25% in the blends. There are
several views on the mechanisms by which the spread
ratio of cookies is reduced when wheat our is supplemented with non-wheat ours (Silva et al., 1998;
McWatters et al., 2003; Giami et al., 2005; Sudha et
al., 2006). It has been established that cookie spread is
strongly correlated to the water absorption capacities of
our (Doescher et al., 1987). The higher water absorption capacities values for PF and SPF compared with
wheat our could have contributed to the lower spread
radio. McWatters (1978) reported that rapid partitioning of free water to hydrophilic sites during mixing
increased dough viscosity, thereby limiting cookie
spread. However, it has been suggested that spread
ratio is aected by the competition of ingredients for the
available water our or any other ingredient which
absorbs water during dough mixing will decrease spread
ratio (Fuhr, 1962).
The specic bulk volume of cookies gradually
decreased to 1.25 for PF and SPF as the content of
these ours was increased to 25% in the blends.
Consumer acceptance and purchase intent of high-fibre

cookies
The results of overall acceptability are shown in Table 5.

The nine dierent types of cookies showed acceptability
scores higher than the minimum acceptable score, i.e. 7
(like moderately).
Cookies made with 0% and 10% PF received the
highest hedonic ratings for overall acceptability (7.98
and 7.95, respectively), diering statistically from other
formulations, which proves the inuence of Brazilian
eating habits over product preference, i.e. Brazilians like
foods with little or no bre addition. Studies revealed
Table 5 Mean sensory rating for consumer acceptance and purchase
intent for cookies containing varying levels of king palm our (PF)
and sieved king-palm our (SPF)
Types of cookies Overall acceptability* Purchase intent scale (%)
1
Control
King palm flour
PF 10%
PF 15%
PF 20%
PF 25%
Sieved king-palm
SPF 10%
SPF 15%
SPF 20%
SPF 25%
2
7.98a
3
0
4
5
2.04 23.54 43.58 30.84
7.50b
7.49b
7.35b
7.33b
flour
7.95a
7.39b
7.38b
7.50b
2.59
1.78
1
1.03
2.59
1.78
2.63
4.18
22.07
25
25.44
24.74
42.28
41.86
40.58
39.14
30.47
29.58
30.05
31.91
2.65
1.83
2.88
1.14
1.77
2.50
2.44
3.41
23.47
23.17
26.27
21.91
40.74
42.50
39.53
42.27
31.37
30.00
29.88
31.27
Mean values in the same column followed by different superscript

alphabets are significantly different (P 0.05).
Mean of 100 judgments.
*Overall acceptability: a nine-point hedonic scale with 1 = dislike
extremely, 5 neither like nor dislike, and 9 = like extremely was used.
Purchase intent scores: 5 = definitely would buy, 4 = probably would

buy, 3 = maybe would buy maybe would not buy; 2 = probably would
not buy; 1 = definitely would not buy.
low dietary bre consumption in the common diet of the

Brazilian population. Although the recommended
intake of dietary bre is 38 g day)1 for men and
25 g day)1 for women under 50 years, the average percapita consumption of dietary bre is approximately
20 g day)1 (De Mattos & Martins, 2000; De Francisco
et al., 2006). All the other formulations received overall
acceptability ratings above seven (like moderately) and
were therefore acceptable.
Purchase intent responses are shown in Table 5. All
the cookies were rated with Denitely would buy and
probably would buy scores, conrming the overall
acceptability results.
Conclusion
This study shows that PFs amounts high contents of total

dietary bre, with predominance of insoluble dietary bre
content. Wheat our supplemented with PF and SPF at
1025% levels produce acceptable cookies in relation to
the overall acceptability and purchase intent, with
increased total dietary bre. Therefore, PLs could be
used as a bre supplement in human diet and as a
functional ingredient in formulated foods such as cookies.
Further studies could investigate strategies to improve the sensory quality of cookies containing high
levels of PL.
Acknowledgments
The authors would like to thank the Conselho Nacional

de Desenvolvimento Cientico and Tecnologico
(CNPq, Brazil) for nancial support and all volunteer
panellists.
References
Akubor, P. & Badifu, G.I. (2004). Chemical composition, functional
properties and baking potential of breadfruit kernel and wheat our
blends. International Journal of Food Science and Technology, 39,
223229.
American Association of Cereal Chemists (AACC) (1995). Approved
Methods of American Association of Cereal Chemists, 9th edn.
St Paul, MI.
American Association of Cereal Chemists (AACC) (2000). Approved
Methods of American Association of Cereal Chemists, 10th edn.
Pp. 1050D. St Paul, MI.
Association of Ocial Analytical Chemists (AOAC) (2005). Ofcial
Methods of Analysis of the Association Analytical Chemists, 18th
edn. Pp. 925.89, 920.85, 920.87, 923.03, 991.43. Washington, DC:
AOAC International Ocal Method.
Bingham, S.A. (2003). Dietary ber in food and protection against
cholesterol cancer in the European prospective investigation into
cancer and nutritional (EPIC): an observational study. The Lancet,
57, 914.
Bonithon-Kopp, C., Kronborg, O., Giacosa, A., Rath, U. & Faivre, J.
(2000). Calcium and bre supplementation in prevention of colorectal adenoma recurrence: a randomized intervention trial. The
Lancet, 356, 13001306.
Bovi, M.L.A. (1998). Cultivation of Alexandra Palm Aiming Heart-ofPalm Production. 26 p. Campinas, Brazil: Instituto Agronomico
(Boletim Tecnico 172).
Chaudhary, V.K. & Weber, F.E. (1990). Barley our evaluated as
dietary ber ingredient in wheat bread. Cereal Food, 35, 560562.
De Francisco, A., Funck, L., Borges, M.R., et al. (2006). Brazils whole
grain market. World Grain, 24, 4247.
De Mattos, L.L. & Martins, I.S. (2000). Dietary ber consumption in
an adult population. Revista de Saude Publica, 34, 5055.
Doescher, L.C., Milliken, R.C. & Rubenthaler, G.L. (1987). Eect of
sugar and ours on cookies spread evaluated by time-lapse
photography. Cereals Chemistry, 64, 163167.
Ferguson, L.R. & Harris, P.J. (2003). The dietary bre debate: more
food for thought. The Lancet, 361, 14871488.
Filizette-Cozzi, T.M.C.C. & Lajolo, F.M. (1991). Fibra alimentar
insoluvel, soluvel e total em alimentos brasileiros. Revista Farmacia
e Bioquimica, 27, 8399.
Food and Agriculture Organization of the United Nation (FAO
WHO) (1973). Energy and Protein Requirements. Food and
Agriculture Organization Nutrition Meeting Report Series 52,
Rome: World Health Organization Technical Report Series, 522.
Food and Agriculture Organization of the United Nation (FAO
WHO) (1990). Amino-Acid Content of Foods and Biological Data on
Proteins. Pp. 6263. Rome: FAO.
Fuhr, F.R. (1962). Cookie spread: its eects on production and
quality. Bakers Digest, 36, 5660.
Giami, S., Achinewhu, S.C. & Ibaakee, C. (2005). The quality and
sensory attributes of cookies supplemented with uted pumpkin
(Telfairia occidentalis Hook) seed our. International Journal of
Guilherme, F.P. & Jokl, L. (2005). Use of quality-protein maize in
our blends and in the production of cookies. Ciencia e Tecnologia
de Alimentos, 25, 6371.
Kruger, C.C.H., Camassetto, M.C.G., Candido, L.M.B., et al. (2003).
Sweet and salted biscuits (snacks) enriched, respectively, with
enzyme coagulated casein and sodium caseinate. Ciencia e Tecnologia de Alimentos, 23, 8186.
Lee, K. & Brennand, C.P. (2005). Physico-chemical, textural and
sensory properties of a fried cookie system containing soy protein
isolate. International Journal of Food Science and Technology, 40,
501508.
McWatters, K.H. (1978). Cookie baking properties of defatted peanut,
soybean, and eld pea ours. Cereal Chemistry, 55, 853863.
McWatters, K., Ouedraogo, J.B., Ressurreccion, A.V.A., Hung, Y.C.
& Phillips, R.D. (2003). Physical and sensory characteristics of sugar
cookies containing mixtures of wheat, fonio (Digitaria exilis) and
cowpea (Vigna unguiculata) ours. International Journal of Food
Science and Technology, 38, 403410.
Meilgaard, M., Civille, G.V. & Carr, B.T. (1999). Sensory Evaluation
Techniques, 3rd edn. 387 p. Boca Raton, FL: CRC Press.
Morrison, W.R. (1998). Cereal lipids. In: Advances in Cereal Science
and Technology (edited by Y. Pomeranz). Pp. 221288. Vol. 2.
St Paul.
Peters, U. (2003). Dietary bre and colorectal adenoma in a colorectal
cancer early detection programme. The Lancet, 361, 14911495.
Rossel, C.M., Rojas, J.A. & de Benedito, B. (2001). Inuence
of hydrocolloids on dough rheology and bread quality. Food
Hydrocolloids, 15, 7581.
Shrestha, A.K. & Noomhorm, A. (2002). Comparison of physicochemical properties of biscuits supplemented with soy and kinema
ours. International Journal of Food Science and Technology, 37,
361368.
Silva, M.R., Silva, M., Aparecida, A.P. & Chang, Y.K. (1998). Use of
jatoba (Hymenaea stigonocarpa mart.) our in the production of
cookies and acceptance evaluation using univariate and multivariate
sensory tests. Ciencia e Tecnologia de Alimentos, 18, 2534.
1539
1540
Silva, M.R., Silva, M.S. & Martins, K.A. (2001). Studies on the use of
jatoba our in biscuits as a source of dietary bre containing no
added simple sugars. Ciencia e Tecnologia de Alimentos, 21, 176182.
Skea, C. M., Thoma, C., Mann, J., Chisholm, A., Williams, S. &
Richmond, K. (2005). Isocaloric substitution of plant sterolenriched fat spread for carbohydrate-rich foods in a low-fat, brerich diet decreases plasma low-density lipoprotein cholesterol and
increases high-density lipoprotein concentrations. Nutrition,
Metabolism and Cardiovascular Diseases, 15, 337344.
Sosulski, F.W., Garatt, M.O. & Slinkart, A.E. (1976). Functional
properties of several sources of dietary bre. Journal of Food
Science, 47, 14721479.
Spackman, D.C. & Stein, W.H. (1958). Automatic recording apparatus
for use in the chromatography of aminoacids. Analytical Biochemistry, 30, 11901206.
Sudha, M.L., Vetrimani, R. & Leelavathi, K. (2006). Inuence of bre
from dierent cereals on the rheological characteristics of wheat
our dough and on biscuit quality. Food Chemistry, 100, 13651370.
Tagliari, P.S. (1999). King-palm stimulates production of palmito in
Santa Catarina. Revista Agropecuaria Catarinense, 11, 6165.
Tames, S.L., Muir, J.G., Curtis, S.L. & Gibson, P.R. (2003). Dietary
bre: a roughage guide. International Medicine Journal, 33, 291296.
Terry, F.C. (2001). Fruit, vegetables dietary ber, and risk of colesterol
cancer. Journal of National Cancer Institute, 93, 525533.
Tsen, C.C., Peters, E.M., Shaer, T. & Hoover, W.J. (1973). Highprotein cookies. 1. Eect of soy fortication and surfactants. Bakers
Digest, 47, 34.
United States Department of Agriculture (USDA) (2002). Nutrient
Database for Standard. Reference. Release 15, Nutrient. Data
Laboratory, Beltsvile Research Center, US. Department of
Agriculture.
Vieira, M.A., Tramonte, K.C., Podesta, R., Braun, A., Amboni,
R.D.M.C. & Amante, E.R. (2003). Estudos preliminares para a
valorizacao dos res duos da producao de palmito a partir da king
palm (Archontophoenix alexandrae). In: Proceedings of the Fourth
Simposio Latino Americano de Ciencia e Tecnologia de Alimentos
(edited by SBCTA). Pp. 571. Campinas, SP.
Watt, B. & Merrill, A.L. (1999). Composition of foods: raw,
processed, prepared. Maryland: US. Department of Agricultural,
Agricultural Research Service, USDA Nutrient Data Laboratory.
1999. USDA Nutrient Database for Standard Reference, Release,
13. Available on: http://www.nal.usda.gov/fnic/foodcomp, accessed
on 25 January 2006).
Original article
Impact of ripening stages of banana flour on the quality of extruded
products
Shirani Gamlath1,2
1 School of Exercise and Nutrition Sciences, Deakin University, Victoria 3125, Australia
2 Department of Food Science & Technology, University of Peradeniya, Sri Lanka (Research was carried out)
(Received 24 July 2006; Accepted in revised form 19 February 2007)
Summary
The study investigated the physical, nutritional and sensory properties of dierent ripening stages of banana
during extrusion processing in combination with rice our to develop quality snack products. Dehydrated
banana ours at ripening stages 4, 5 and 6 (peel colour) were mixed separately at 40% banana to 60% rice
our levels. The mixtures were extruded through a twin-screw extruder at 120 C barrel temperature, 220
and 260 r.p.m, screw speed and 12% feed moisture. Increase in ripeness indicated negative eect on
expansion and water absorption capacity while increasing the water solubility index and moisture retention
(wet basis) of the products. Protein and mineral (except for zinc and copper) content of the products were
signicantly dierent (P < 0.05) from 4 to 6 of the ripening stages. Most of the essential amino acids in the
extruded products increased signicantly (P < 0.05) at the ripening stage of 6. All the products were within
the acceptable range in the 9-point Hedonic scale showing the best texture and avour scores for stage 4 and
6, respectively. The extruded products show potential as snack products because of their nutritional quality
and sensory acceptability.
Keywords
Extrusion, fruits, nutritional aspects, physical, rice, sensory.
Introduction
Extrusion processing plays an important role in food

processing as it provides versatility in handling wide
range of raw materials and creating novel products.
Extrusion is one of the widely used processing methods
in snack, breakfast cereal, confectionary and pet food
industries (Frame, 1994; Fellows, 2000; Guy, 2001)
because of a signicant number of advantages such as
continuous process, high productivity and no process
euents. The high-temperature short-time conditions in
extrusion processing have minor eects on natural
colours and avours of foods while inactivating undesirable enzymes, lowering anti-nutritional factors and
sterilising the product (Harper, 1981; Fellows, 2000;
Guy, 2001). Despite the increased use of extrusion
processing on cereals, legumes and other starchy materials, there is very limited information available on
extrusion of perishable crops. Understanding the behaviour of unexplored raw materials such as fruits during
extrusion processing would be benecial to provide
scientic knowledge in the gaps and to identify the
Correspondent: E-mail: shirani.gamlath@deakin.edu.au
doi:10.1111/j.1365-2621.2007.01574.x
feasibility of using some valuable raw materials in future

product development.
Banana is the fruit of the plant Musa spp. which
grows under tropical and sub tropical conditions. It is
the most extensively grown fruit with a world total
production of 69 832 000 Mt in 2002 (FAO, 2003). In
tropical and sub tropical countries the warm and humid
climate accelerates the decaying of fruits leading to
2050% losses in fruits. Considerable quantities of
banana surplus to the requirement of fresh trade are
available for processing in all the banana exporting
countries in the world (FAO, 2003).
The nutritional composition of banana is comparable
to potato containing about 20% of carbohydrate, 1%
protein, less than 1% fat, 100130 calories and micronutrients; 0.4 mg iron, 0.6 mg niacin, 10 mg ascorbic
acid in 100 g of the edible pulp of the ripe banana
(Marriott, 1980). The low lipid and high energy content
of banana combined with high satiety value makes it
very useful in low-fat diets. Its high potassium content
[400 mg (100 g)1)] has a protective eect against excess
sodium in diets. Banana is a preferred item in the diet as
it contains considerable amount of bre to reduce
intestinal disorders and cholesterol absorption (Nagy &
1541
1542
Ripening stages of banana during extrusion processing S. Gamlath
Shaw, 1980; Stover & Simmonds, 1987) According to

Horigome et al. (1992) freeze-dried banana pulp showed
a marked cholesterol-lowering eect when incorporated
into a diet at the level of 300 or 500 g kg)1. The soluble
and insoluble components of dietary bre have participated in the hypocholesterolaemic eect of banana pulp.
High starch content of banana compared with the
other fruits could be favourable for the extrusion
process. However, high levels of sugars and moisture
in the composition could lead to undesirable eect
during the process and also on the product quality
(Noguchi et al., 1982). Dierent ripening stages of
banana have slightly dierent compositions. In bananas,
changes during ripening are mainly the conversion of
starch to non-reducing and reducing sugars (Salunke
et al., 1991). The changes in composition in unripe to
fully ripened Montecristo banana were 6771% water,
0.816.2% sugars, 273% starch (Marriott, 1980).
As a complex multivariate process, extrusion requires
careful control if product quality is to be maintained.
Changes to the product quality during extrusion primarily depends on the raw material properties (composition and moisture content) and the extruder conditions
mainly as barrel temperature, screw speed and feed
moisture (Harper, 1981; Lawton et al., 1985; Guy, 2001;
Gamlath & Balendran 2004; Gamlath & Clarke, 2004).
Any change in feed composition, inuenced by the type
of ingredients, as well as the amount of each ingredient,
and any process variable that aects physical or
chemical transformation of macromolecules during
extrusion can inuence extrusion performance and
quality of the extrudate (Rhee, et al., 2004). As dierent
ripening stages of banana poses varied amounts of
starch and sugars in the composition, it might lead to
variation in product qualities and would be important to
identify the correct ripening stage for acceptable quality
products.
In this context, the main objective of this study was to
investigate the impact of dierent ripening stages of
banana on development of high-quality extruded products. Further, the study aimed at providing scientic
knowledge in the changes in physico chemical, nutritional and sensory properties of the extruded banana
products. The study would also provide a fundamental
basis to enhance the utilisation of banana in product
development using extrusion technology.
Material and methods
Raw materials
Three dierent ripening stages of banana variety

Cavendish; No. 4 (more green than yellow), No. 5
(green tips and necks) and No. 6 (all yellow, no green
tips) according to the Banana Ripening Index (Anon.,
1985) were obtained from Agricultural research station,
Gannoruwa, Sri Lanka. Raw red rice (Oryza sativa) was

collected from local market.
Preparation of raw materials
Bananas were peeled, pulped and dehydrated in a drum

drier (Agriculture Research Station, Gannoruwa, Sri
Lanka) at 120 C for 1 min. Rice and dehydrated
banana were powdered using the cross beater Mill tted
with a screen of 3 mm. Samples were prepared with rice
and banana powder at the proportion of 60:40 (wet
basis), respectively, 23 h before the extrusion using the
Hobart mixer for 10 min. Each sample was made to
2 kg and the moisture content of the feed mixture was
calculated. Raw materials and feed mixtures were also
analysed for their proximate composition and mineral
contents.
Extruder
Extrusion was performed with a pilot plant scale

Bersto co-rotating twin-screw extruder with a 44 cm
length and 4 cm diameter. Barrel consisted with a three
zones separated by three kneading elements at equal
distance. The temperature of each section was monitored using a thermocouple probe. A variable speed
automatic feed hopper was attached to ensure a uniform
ow into the barrel. Moisture content in the feed was
automatically adjusted by using a moisture pump
attached in the feed section.
Experiments
Preliminary tests were conducted to identify the optimum range of the extruder (temperature, screw speed
and die diameter) and feed (moisture and proportions of
banana to rice) variables. Based on these trials, further
experiments were conducted using 120 C (nal zone)
barrel temperature. Zones 1 and 2 were kept constant at
50 and 80 C, respectively. The optimum screw speeds
were maintained at 220 and 260 r.p.m, and 3 mm die
diameter. The feed moisture contents were xed at 12%
and the feed rate was maintained at 260 g min)1. The
proportion of 60:40 rice to banana our was used as this
was the highest proportion of banana could go through
the extruder without any feeding diculty.
Detail tests were conducted using two-way factorial
design using dierent ripening stages of banana and
screw speeds, and the extruded products were analysed
for physical properties. Based on the changes in the
physical properties, one screw speed was selected
(220 r.p.m). The products obtained at 220 r.p.m showed
uniform and smooth surface compared with the uneven
and rough surface of the products extruded at
260 r.p.m.
Therefore,
products
obtained
at
220 r.p.m was analysed for proximate composition,
mineral content and the essential amino acid prole.

Sensory evaluation was also conducted to assess the
acceptability of those products. Overall, the eect of
ripening stages on the quality of extruded products was
evaluated.
Analytical methods
Physical properties
Expansion ratio
Expansion ratio (ER) was dened as the ratio of the

diameter of the extrudates and the diameter of the die
(Fan et al., 1996). A vernier calliper was used to
measure the diameter of the extrudates. Measurements
were taken on ten randomly selected pieces of extrudate
and ER was calculated using eqn 1.
ER
Total soluble solids
Total soluble solids (TSS) content was determined using

the refractometric method; 940.09 (AOAC, 1995). Five
grams of dehydrated banana powder was measured in to
a centrifuged tube and 25 ml of distilled water was
added. The tubes were shaken for 3 h, centrifuged and
supernatant was decanted. The extraction was repeated
with another 25 ml of water and the total supernatant
was used to test the TSS content by measuring the
refraction index (ATAGO, digital refractometer
PR1, USA Inc., Washington, DC, USA) as given by
the eqn 5.
TSS% B D
where B and D refer to Brix value and the dilution

factor, respectively.
Nutritional analysis
D
D1
1
Proximate composition
where D and D1 are the diameters of the extrudate and

the die nozzle, respectively.
Water absorption capacity and water solubility index
The water absorption capacity (WAC) and water

solubility index (WSI) were determined using the procedures described by Ascheri et al. (1998). Ground
extrudate (3 g) was mixed with 70 ml of distilled water
in a centrifuge tube for 3 s using a vortex mixture. The
tubes were shaken in a water bath at 30 C for 30 min.
The samples were centrifuged for 10 min at 3000 r.p.m
and the supernatant was decanted. The weight of gel
formed in the centrifuged tube was recorded. The
supernatant solution was heated in a water bath, dried
at 80 C and the weight was recorded. The WAC and
WSI were calculated using eqns 2 and 3, respectively.
Moisture content (oven dry method; 945.15), crude

protein (Kjeldhal method), crude fat (ether extract) and
crude bre (digestion and gravimetric) and total ash
content (gravimetric) were measured in triplicate on raw
materials, feed mixtures and extruded products using
AOAC standard methods; 945.18 (1995). Total carbohydrate content was calculated by the dierence.
Essential amino acid prole
where W, W1 W2 are weight of dry sample solids, weight

of sediment and weight of dissolved solids in the
supernatant, respectively.
Amino acid contents were determined according to the

Pico-Tag method in a Manual of advanced techniques
for amino acid analysis (1987) using the waters Pico-Tag
amino acid analyser; model, 510 (Millipore, Bedford,
MA, USA). Samples were powdered to pass through
1 mm screen and 40 mg of sample was hydrolysed with
6-N hydrochloric acid at 110 C for 24 h. The solution
was cooled and hydrolysate was ltered through a nylon
membrane lter. The 20 lL of sample was dried with
addition of ethanol, water and trimethylamine and then
derivatised using phenyl isothiocyanate. The sample was
injected into the liquid chromatography system. Amino
acid standards prepared as the above procedure were
injected before the samples. Each of the amino acid in
the sample was calculated and was expressed as
mg (100 g)1) dry weight.
Moisture retention
Mineral contents
Moisture contents of feed mixtures and extruded products were measured using automatic moisture analyser
(CSC Mettler, instruments, Greifensee, Switzerland) and
moisture retention (MR) was calculated using eqn 4.
Mineral contents of feed mixtures and extrudates were

conducted according to Atomic Absorption Spectrometric (AAS) method manual (1985) as follows.
One gram of ground sample was ignited in a Mue
furnace at 450 C, cooled and dissolved in 10 ml 2 N
conc. Hydrochloric acid, boiled gently and ltered in to a
100 ml volumetric ask and made up to the volume with
de-ionised water. This solution was used to determine the
mineral contents using an AAS (Perkins Elmer model,
WACg=g
WSI%
MR%
W1
W
W2
100
W
Mf Mp
100
W
2
3
where Mf, Mp and W denote feed moisture, extruded

product moisture and weight of dry sample, respectively.
1543
1544
2380, Waltham, MA, USA) with corresponding

lamps. Absorption was measured directly within the
ranges after the standard solutions were made.
Sensory evaluation
Quantitative aective analysis (Meillgard et al., 1999)

was used and the tests were conducted in the Food
Science laboratory at University of Peradeniya. Sensory
evaluation was conducted by a panel of thirty judges
consisting of faculty sta and students who were
experienced with the product and terminology. A
questionnaire developed according to the 9-point Hedonic scale (9-Like extremely and 1-Dislike extremely)
was used to evaluate the acceptability for colour,
avour, texture and overall acceptability of samples.
The panelists were also asked to comment on the colour,
avour, texture and overall properties of the products.
Data on physical properties were analysed statistically

using analysis of variance (anova) for a randomised
two-way factorial design (ripening stages and screw
speeds) and Mean comparison was done by least
signicant dierence (LSD) at P < 0.05. One-way
anova was performed for proximate composition,
mineral contents and amino acid contents. Data on
sensory evaluation was analysed using Freedmans test
using MINITAB statistical package.
Physical properties
Physical properties of extruded products, processed at

220 and 260 screw speeds are presented in Table 1.
Increase in ripeness signicantly (P < 0.05) increased
the MR and WSI while decreasing the ER and WAC.
This could be related to the increase in TSS during the
ripening of bananas. Addition of sugars, more than 10%
may have a large eect on the product characteristics
and higher levels of sugar reduce the energy input and

mass temperatures to fall thus reducing expansion
(Frame, 1994; Jintian et al., 1996). The low levels of
starch gelatinisation at high sugar contents in the feed
may have also contributed to a less expansion. The
lower expansion of the products at the ripening stage 6
could also be related to the increased protein content
during ripening (Table 2) and also having higher levels
of protein in the products at Stage 6 (Table 3). Several
researches have reported that proteins tend to reduce the
extensibility of the starch polymer during its expansion
at the die, reducing the degree of expansion (Li & Lee,
1996; Iwe, 1998; Guy, 2001).
Increased MR with more ripened banana in the
products could be related to the less ashing of moisture
at lower product expansion (Harper, 1981; Badrie &
Mellows, 1991). Higher screw speed (260 r.p.m) also
slightly increased the expansion and WSI, while
decreased the MR and WAC. However, the surface of
the product obtained at 260 r.p.m was uneven. Higher
screw speeds create high shear within the extruder barrel
which could lead to breakdown of gelatinised starch to
more soluble compounds such as dextrin or sugars
lowering the WAC and increasing the WSI. Gautam &
Choudhury (1999) indicated that screw conguration
also an important factor in starch breakdown.
Nutritional contents
The proximate composition, TSS and mineral contents

of dierent ripening stages of banana are presented in
Table 2. Among the proximate composition, protein
and carbohydrate contents were signicantly dierent
(P < 0.05) in dierent ripening stages of banana. The
ripening stage 6 had signicantly dierent (P < 0.05)
total ash content to stage 4. The minerals, calcium,
magnesium and potassium were also signicantly
increased during ripening. Compositional changes in
bananas during ripening have been discussed by several
authors (Marriott, 1980; Stover & Simmonds, 1987;
Ripening
stage
Screw
speed(r.p.m)
Expansion
ratio
Moisture
retention%
WAC
WSI%
220
260
220
260
220
260
3.11 0.04a
3.23 0.07b
2.48 0.05c
2.61 0.12d
2.21 0.07e
2.38 0.08f
0.08
38.6 0.05a
37.7 0.03b
40.3 0.05c
39.4 0.06d
50.5 0.03e
51.0 0.0f
0.41
5.68 0.13a
5.63 0.17a
5.47 0.08b
5.28 0.73c
5.23 0.06c
5.14 0.08d
0.06
23.35 0.72a
24.10 0.08a
24.30 0.05b
25.81 0.18c
25.40 0.40c
26.20 0.21c
0.84
5
6
LSD (P < 0.05)
Table 1 Physical properties of different

ripening stages of banana at 220 and
260 r.p.m screw speed
WAC = g of gel formed per gram.

Values are means of three determination SD.
Means followed by different superscript alphabets in each column are significantly different
(P < 0.05).
Table 2 The proximate composition

[g (100 g)1)] and mineral contents
[mg (100 g)1)] of edible portion of dierent
ripening stages of banana and rice
Banana pulp
No. 4
Protein
Fat
Fibre
Ash
Carbohydrate*
Total soluble solids
Calcium (Ca)
Phosphorus (P)
Magnesium (Mg)
Potassium (K)
Iron (Fe)
Zinc (Zn)
Copper (Cu)
3.3
2.3
2.7
3.8
34.1
16.3
16.4
45.2
25.7
323.2
0.8
0.2
0.1
Banana pulp
No. 5
0.2
0.3a
0.5a
0.2a
1.2a
0.2a
0.2a
0.3a
0.3a
1.2a
0.2a
0.1a
0.1a
3.1
3.1
2.4
4.5
31.2
21.4
14.3
38.4
34.2
331.1
0.8
0.2
0.2
Banana pulp
No. 6
0.1
0.5a
0.3a
0.3a
1.3b
0.1b
0.4b
0.2b
0.4b
1.6b
0.1a
0.1a
0.1a
3.9
2.7
2.7
4.7
29.8
23.5
17.8
38.6
32.1
347.0
0.5
0.2
0.2
Rice
0.2
0.3a
0.2a
0.2b
0.8c
0.3c
0.1c
0.1b
0.5c
1.8c
0.0b
0.2a
0.0a
7.3
0.8
1.2
1.1
78.9
0.3
0.1
0.1
0.1
1.3
62.1
122.4
32.7
158.1
1.3
1.9
0.4
0.5
0.6
1.2
0.4
0.2
0.2
0.1
Values are means of three determination SD.

Means followed by different superscript alphabets in each row are significantly different (P < 0.05).
*Calculated by the difference.
Table 3 The proximate composition

[g (100 g)1) of dry weight] of feed mixtures
(40% banana:60% rice) and extruded products of dierent ripening stages of banana at
220 r.p.m screw speed
Protein
Feed mixture No. 4
No. 5
No. 6
Product No. 4
No. 5
No. 6
6.8
7.0
7.3
6.4
6.8
6.9
Fat
c
0.2
0.5c,d
0.2d
0.1e
0.2c
0.2d
3.2
2.6
2.9
2.7
2.0
2.5
Fibre
0.4
0.3c
0.3c
0.3c
0.5c
0.3c
2.3
2.8
2.5
2.4
2.9
2.3
Ash
c
0.3
0.2c
0.3c
0.3c
0.4c
0.5c
2.5
3.1
2.6
2.2
2.6
2.4
Carbohydrate*
c
0.2
0.3c
0.2c
0.4c
0.3c
0.3c
86.5
85.7
84.7
84.1
83.3
82.6
0.8c
1.0c
1.2c
1.1d
1.1d
0.7d
Values are means of three determinations SD.

Means followed by different superscript alphabets in each column are significantly different
(P < 0.05).
Salunke et al., 1991). In bananas, changes are mainly

conversion of starch to sugars. As the fruit ripens, the
moisture content of the pulp increases as a result of the
penetration of water from the peel, increasing the value
of pulp:peel ratio. Marriott (1980) also noted that there
is no change in protein content relative to nitrogen
content in pulp and peel during ripening. In contrast
Stover & Simmonds (1987) noted that protein synthesis
including enzymes increases during ripening. The
increase in all essential amino acids in bananas during
ripening has been reported (Askar, 1973).
Similar pattern of dierence was observed in the feed
mixtures and the extruded products showing increased
levels of protein with increase of ripeness (Table 3). The
insignicant change in carbohydrate content may be
attributed to the use of 60% of rice our with the
mixtures as the rice our contains almost 80% carbohydrate. The ripening stage 6 had the higher levels of
calcium, magnesium, potassium and were signicantly
dierent from the ripening stage 4 (Table 4) However,
as a result of the presence of 60% rice in the mixtures
and loses during extrusion may have contributed to the
insignicant dierences in the mineral contents of the

extruded products. The feed mixtures for all the ripening
stages had higher levels of calcium, phosphorus, magnesium and potassium compared with the extruded
products; however, the changes were insignicant
(P < 0.05). Singh et al. (2000) and Alonso et al.
(2001) indicated increased in mineral contents during
extrusion because of the contamination of metals by
wear and tear of the extruder components and water
used during extrusion.
Extruded products containing the most ripened
banana had the highest levels of all essential amino
acids except for Threonine, Methionine and Lysine
(Fig. 1). The results are comparable with that of Askar
(1973). However, lower levels of Lysine in the extruded
product may be attributed to the Maillard reactions
during processing with sugars and amino acids as sugar
levels increased with the ripening. The increase in
soluble solids also an indication of the conversion of
starch to sugars during ripening.
According to Iwe et al. (2004) lower die diameters
and higher screw speeds reduces the lysine content and
1545
)1
Table 4 The mineral contents [mg (100 g
dry basis] of feed mixtures (40 banana:60 rice) and extruded products at 220 r.p.m screw speed
Ca
Feed Mixture No. 4
No. 5
No. 6
Product No. 4
No. 5
No. 6
65.3
64.3
68.4
64.2
61.5
67.1
1.2
1.3 a
1.4b
1.2 a
1.5 c
1.4b
Mg
148.5
146.7
143.8
146.2
140.2
139.6
1.5
1.7a
1.3b
1.5a
1.3c
1.7c
64.8
66.9
75.2
58.3
65.8
67.1
0.5
1.2a
0.8b
1.2c
1.1a
1.5a,e
Fe
428.4
416.1
438.5
425.2
415.0
432.7
1.4
2.3b
2.1c
1.2a
1.6b
1.5d
2.3
2.2
1.4
2.6
2.1
1.8
Zn
0.5
0.2a
0.5a
0.2a
0.1a
0.4a
1.1
0.7
0.8
0.8
0.5
1.1
CU
0.2
0.4a
0.1a
0.2a
0.3a
0.2a
0.5
0.2
0.2
0.3
0.2
0.5
0.3a
0.4a
0.3a
0.2a
0.1a
0.1a
Values are means of three determinations SD.

Means followed by different superscript alphabets in each column are significantly different (P < 0.05).
400
Amino acid content (mg 100 g1)
b
350
300
a
b
250
bb
bb
bb
a
200
a b
a
a aa
a
150
100
50
0
His
Thr
Tyr
Val
Meth
Ile
Essential amino acids
Leu
the reductions could be related to the Maillard

reaction as evident by the increased Browning index.
Noguchi et al. (1982) has also supported the argument
that lysine losses are mainly because of the Maillard
browning.
Phe
Lys
Figure 1 Essential amino acid contents of extruded products of different ripening stages (Nos
4, 5 and 6) of banana and rice (40:60, respectively) at 220 r.p.m, screw speed. Vertical bars
represent SE of mean (n = 6). Mean values
followed by dierent letters in each amino acid
are signicantly dierent (P < 0.05). His =
Histidine, Thr = Threonine, Tyr = Tyrosine,
Val = Valine, Meth = methionine, Ile = Isolucine, Leu = Leucine, Phe = Phenylalanine,
Lys = Lysine.
Sensory properties
The median scores for sensory acceptability of all

ripening stages of products (Fig. 2) were within the
acceptable range indicating ranks above 5. However,
9
a
7
Median scores
1546
b b
4
5
6
5
4
3
2
1
0
Colour
Flavour
Texture
Sensory properties
Overall
Figure 2 Median scores of sensory properties

for extruded products of different ripening
stages (Nos 4, 5 and 6) of banana and rice at
40:60 proportions, respectively. Median values
followed by different letters in each sensory
property are signicantly different (P < 0.05).
there is a signicant dierence in all the attributes as the

ripeness increases from stage 4 to 6. Increase in ripeness
had a negative eect on texture while a positive eect on
colour and avour of the products. The products
containing banana, stage 6, were darker in colour and
may be related to the Maillard reaction of sugars and
amino acids during extrusion. Based on the comments
made by panellists on specic attributes, the product
containing the ripening stage 6 was preferred than the
lighter or dull colour of the stage 4. Increased ripening
also contributed to the more avour in the product. The
opposite eect of ripeness on texture would be related to
the less expansion with increased levels of sugars leading
to less crispier and denser products. Most of the panellists
were commented that the products containing stage 6
were less crispier and denser than the products made with
stage 4. This has also supported by the decreased WAC,
indicating less starch gelatinisation with increasing sugar
levels in the extrudates as ripening proceeds. Overall,
ripening stages 5 and 6 were not signicantly dierent for
avour, texture and overall properties. However, colour
acceptability was signicantly dierent indicating, the
ripening stage 6 as the more preferred product regardless
of the slightly lower textural scores.
Conclusion
In summary, dierent ripening stages of banana from

stage 4 to 6 showed signicant dierences in physical,
nutritional and sensory properties of extruded products.
The products containing banana number 6 had higher
protein and amino acid levels (except, lysine, methionine
and threonine) and signicantly preferred for overall
sensory properties among the three ripening stages. The
product of 100 g (dry basis) contained 6.9 g of protein,
2.5 g of fat and 2.3 g of bre, 432.7 mg potassium and
139.6 mg phosphorus. The study suggests that banana
could be incorporated at 40% dry basis with rice our to
produce acceptable products as snack or breakfast
cereals with signicant nutritional properties. The physical, nutritional and sensory qualities of banana products
would be further enhanced by replacing rice our by
legume ours in the mixtures. Further investigations are
required to establish optimum extruder conditions and
raw material proportion for desired products.
Acknowledgments
Author would like to thank CISIRO and Agriculture

Research Station, Gannoruwa, Sri Lanka for the
technical support.
References
Alonso, R., Rubio, L.A., Muzquiz, M. & Marzo, F. (2001). The
eect of extrusion cooking on mineral bioavailability in pea and
kidney bean seed meals. Animal Feed Science and Technology, 94,
113.
Anon. (1985). The Ripening of Banana Standard Colour Index for
Banana Ripening. Pp. 1. Collingwood, Victoria: CSIRO.
AOAC (1995). Association of Ofcial Analytical Chemists. Ofcial
Methods of Analysis, 16th edn. Arlington, VA: AOAC Inc.
Ascheri, J.L.R., Bernal-Gomez, M.E., Carvalho, C.W.P. & Modesta,
R.C.D. (1998). Production of snacks from mixtures of rice and sweet
potato ours by thermoplastic extrusion. Functional and sensory
properties. Alimentaria, 293, 7986.
Askar, A. (1973). The importance of amino acids in fruits. Behaviour
of amino acids during ripening of bananas. Giordian, 73, 12, 14, 16.
Badrie, N. & Mellows, W.A. (1991). Eect of extrusion variables on
cassava extrudates. Journal of Food Science, 56, 13341337.
Fan, J., Mitchell, J.R. & Blanchard, J.M.V. (1996). The eect of sugars
on the extrusion of maize grits: II. Starch conversion. International
Journal of Food and Technology, 31, 6776.
FAO (2003). The World Banana Economy 19852002. Pp. 6. Rome:
Food & Agriculture Organisation.
Fellows, P. (2000). Food Processing Technology: Principles and
Practice. Pp. 294305. Cambridge: Woodhead Publication Ltd.
Frame, N.D. (1994). The Technology of Extrusion Cooking. Pp. 151.
London: Blackie Academic & Professional Publishers.
Gamlath, G.G.S. & Balendran, S. (2004). Physical characteristics of
extruded products from maize, nger millets and legumes. In:
Proceedings of the Hawaii International Conference on Sciences
(edited by A. Burge). Pp. 288. Honolulu: Hawaii International
Conference on Sciences.
Gamlath, G.G.S. & Clarke, B. (2004). Physical, nutritional and
sensory properties of extruded banana, rice and green gram
mixtures. In: Proceedings of the Hawaii International Conference
on Sciences (edited by A. Burge). Pp. 289. Honolulu: Hawaii
International Conference on Sciences.
Gautam, A. & Choudhury, G.S. (1999). Screw conguration eects
starch breakdown during twin screw extrusion of rice our. Journal
Food Processing and Preservation, 23, 355375.
Guy, R. (2001). Extrusion Cooking: Technologies and Application. Pp.
528. London: Woodhead Publishing Limited.
Harper, J.M. (1981). Extrusion of Foods, Vol. 1 and Vol. 2. Pp. 4160,
89112. FL: CRC Press.
Horigome, T., Sakaguchi, E. & Kishimoto, C. (1992). Hypocholesterolaemic eect of banana (Musa sapientum L. var. Cavendishii) pulp
in the rat fed on a cholesterol-containing diet. The British Journal of
Iwe, M.O. (1998). Eects of extrusion cooking on functional properties
of mixtures of full-fat soy and sweet potato. Plant Foods for Human
Iwe, M.O., Zuilichem, D.J., Van Stlop, W. & Ngoddy, P.O. (2004).
Eect of extrusion cooking soy-sweet potato mixtures on available
lysine content and browning index of extrudates. Journal of Food
Engineering, 62, 143150.
Jintian, F., Mitchell, J.M. & Blanshard, J.M.V. (1996). The eect of
sugars on the extrusion of maize grits: starch conversion. International Journal of Food Science and Technology, 31, 6776.
Lawton, J.W., Davis, A.B. & Behnke, K.C. (1985). High temperature
short time extrusion of wheat gluten and a bran like fraction. Cereal
Li, M. & Lee, T.C. (1996). Eect of cysteine on the functional
properties and microstructures of wheat our extrudates. Journal of
Agriculture and Food Chemistry, 44 (7), 18711880.
Marriott, J. (1980). Bananas Physiology and Biochemistry of Storage
and Ripening for Optimum Quality. Critical Reviews in Food Science
and Nutrition. Pp. 13, 4148. Boca Raton, FL: CRC Press.
Meillgard, M., Civille, G.V. & Carr, B.T. (1999). Sensory Evaluation
Techniques. Pp. 201226. Boca Raton, FL: CRC Press.
Nagy, S. & Shaw, P.E. (1980). Tropical and Subtropical Fruits.
Compositional Properties and Users. Pp. 258278. Westport, CT:
AVI Publishers.
1547
1548
Noguchi, A., Mosso, K., Aymanrd, C., Jevnink, J. & Cheftel, J.C.
(1982). Millard reactions during extrusion cooking of protein
enriched biscuits. Lebensmittel Wissenschaft & Technology, 15,
105110.
Rhee, K.S., Kim, E.S., Kim, B.K., Jung, B.M. & Rhee, K.C.
(2004). Extrusion of minced catsh with corn and defatted soy
ours for snack foods. Journal of Food Processing Preservation,
28, 288301.
Salunke, D.K., Bolin, H.R. & Reddy, N.R. (1991). Storage Processing
and Nutritional Quality of Fresh Fruits and Vegetables, Vol. 2. Pp.
5694. Boca Raton, FL: CRC Press.
Singh, D., Chauhan, G.S., Suresh, I. & Tyagi, S.M. (2000). Nutritional
quality of extruded snacks developed from composite of rice brokens
and wheat bran. International Journal of Food Properties, 3, 421431.
Stover, R.H. & Simmonds, N.W. (1987). Bananas. Pp. 386395.
New York, NY: Longman Scientic & Technical Publishers.
Original article
Quality assessment of whole and gutted sardines
(Sardina pilchardus) stored in ice
zkan Ozden
Nuray Erkan* & O
Department of the Seafood Processing and Quality Control, Faculty of Fisheries, Istanbul University, Istanbul, Turkey
(Received 11 March 2006; Accepted in revised form 19 February 2007)
Summary
The quality and shelf life of whole ungutted and gutted sardines (Sardina pilchardus) stored in ice were
studied. The changes in the sh were investigated by sensory assessments, chemical analyses and
microbiological analyses. The sensory scores of uneviscerated and gutted sardines stored in ice at +4 C
were 7 days. The chemical indicators of spoilage, total volatile basic nitrogen and trimethylamine values of
gutted sardine increased very slowly, whereas for whole ungutted samples higher values were obtained
reaching a nal value of 15.0329.23 mg per 100 g and 2.364.16 mg per 100 g, respectively (day 9). Peroxide
and thiobarbituric acid values remained lower for whole ungutted sardine samples until day 9 of storage,
whereas for gutted sh were higher. The level of histamine exceeded the legal limit in whole ungutted sardine
after 7 days of storage in ice, during which sardines were rejected by the sensory panel. Mesophilic aerobic
bacteria count, H2S-producing bacteria, sulphide reducing anaerobe Clostridias, Enterobacteriaceae count of
whole ungutted sardine samples are higher than gutted sardine samples during the storage. Psychrotrophic
bacteria counts of the two groups were not dierent. The limits of microbiological data were not exceeded
throughout the storage in both the groups samples.
Keywords
Fish quality, freshness indicator, sardine, shelf life.
Introduction
Sardine is an important food species in Turkey; total

catch was 12 883 tons in 2004. It is generally consumed
as fresh, canned or salted and also utilised as sh meal
and oil. Sardine is known to be high protein and oil sh.
Fresh sh and other sh sea products are highly
susceptible to spoilage from post-mortem microbial
growth and enzymatic activity. During handling and
storage, quality deterioration of fresh sh occurs rapidly
and limits the shelf life of the product. The shelf life of
sh in ice can be extended depending on raw material,
the storage conditions (temperature and atmosphere),
intrinsic factors such as species, age and size, fat content,
feeding and physiological status, and the qualitative and
quantitative composition of the initial microora, related to the environment where the sh live and are
caught, the seasonal period and the shing method
(Huss, 1994; Tulsner, 1994; Ashie et al., 1996; Erkan,
2003).
Information on quality and shelf life of sardine
(Sardine pilchardus) stored in ice are available in
literature (El Marrakchi et al., 1990; Nunes et al.,
*Correspondent: E-mail: nurerkan@istanbul.edu.tr
1992; Gokoglu et al., 1998; Gennari et al., 1999;

Pacheco-Aguilar et al., 2000; Gokoglu & Yerlikaya,
2004). Most of them studied quality and shelf life of
sardine in ice and or in cold 4 C, but results were not
always in agreement and in particular the limit of
acceptability for whole sardine ranged from 4 to
9 days.
The present paper reports on quality assessment of
whole and eviscerated sardines stored in ice by evaluation of chemical, microbiological and sensory changes.
Furthermore, the shelf life of whole and eviscerated
sardines during ice storage and under refrigeration is
also reported.
Raw material, processing and sampling
Fresh sardine (Sardina pilchardus), a Marmara Sea

pelagic species, was obtained from a shing port. Fresh
sh with ice in wooden boxes were transported to the
laboratory in 11 h. Samples (20 kg) were used for
the experiment. The mean and standard deviations of
the weight and length of the sh studied were
23.07 0.2 g and 13.5 0.05 cm, respectively. Fish
doi:10.1111/j.1365-2621.2007.01579.x
2007 The Authors. Journal compilation 2007 Institute of Food Science and Technology
1549
1550
. O
zden
Gutting effect on the quality in fish N. Erkan and O
were individually washed, and were divided into two

lots; whole ungutted and gutted samples. Gutting of the
shes was carried out manually in the sh processing
plant. The two groups of shes were kept in polystyrene
boxes with ice in cold storages at +4 1 C. During
storage, ice was replenished when necessary. Boxes had
perforated bottoms to allow drainage of crushed ice. All
analytical determinations were done by triplicate at day
1, 3, 5, 7 and 9. The ice sh ratio (1:1) was maintained
constant throughout the experiment.
Sensory assessment
Raw sh
Sensory analysis was conducted by a taste panel

consisting of ve experienced judges from the laboratory
sta, trained in grading sh, on the whole and gutted
raw sardines according to the European Community
(EC) grading scheme (Table 1) for bluesh (EC Regulation, 1996). The evaluation depends on the table
created because bluesh concept includes tuna, mackerel, sardine, horse mackerel, anchovy and herring. The
appearance and the odour of the skin, eyes, gills, mucus
and texture of each sh (whole and gutted) were assessed
into four quality grades. In this EC grading scheme,
excellent quality (perfect condition), high quality (slight
loss of excellent characteristics), good quality (some
deterioration, but t for sale) and unt for sale were
assigned as E (better extra), A (good quality), B
(moderate quality) and C (unt) grades, respectively.
For the EC scheme, the mean points of each sh were
calculated and the sh were classied according to the
following correspondence between points and quality
bands: E 2.7, 2.0 A < 2.7, 1.0 B < 2.0 and C < 1.0
(Huss, 1988). The total points of each sh were

estimated from the grades attributed by each panellist
and the nal grade of each sh species (whole and
gutted) was estimated from the sh samples examined at
each day of evaluation.
Quality index method
Raw sh were evaluated using the quality index method

(QIM) shown in Table 2. This structured category scale
is based on the freshness quality grading system for
whole-iced herring developed by Ozogul et al. (2000).
The QIM involves specifying the characteristics of
appropriate sensory attributes of the raw sh. Once
the characteristic of a sensory attribute is determined, it
is assigned for a demerit score ranging from 0 to 3 (see
Table 2). The scores for all characteristics are then
summed to give an overall sensory score, the so-called
quality index. The scale gives zero score for absolutely
fresh sh, while increasingly larger totals result as sh
deteriorate. The limit of acceptability for round sardine
on this freshness scale is 18 2.
Cooked sh
The attributes of cooked sh were evaluated by a panel

of ve experienced judges on each day of sampling. Fish
samples were cooked individually in a microwave oven
at full power (600 W) for 5 min time and immediately
presented to the panellists. Panel lists were laboratory
trained. Sensory evaluation was conducted in individual
booths under controlled conditions of light, temperature
and humidity. Panellists were asked to score odour, taste
and texture of sh using a 010 acceptability scale (scale:
109 = excellent, 8.98 = very good, 7.96 = good,
5.94 = sucient, 4.93.1 = limit of acceptable
Table 1 EC freshness assessment scheme
Skin
Skin mucus
Consistency
of flesh
Gill covers
Eye
Gills
Smell of gills
Extra
Bright pigmentation, bright,

shining iridescent colours;
clear distinction between
dorsal and central surfaces
Aqueous, transparent
Very firm, rigid
Loss of lustre and shine;

duller colours; less
difference between dorsal
and ventral surfaces
Slightly cloudy
Fairly rigid, firm
Dull, lustreless,
insipid colours; skin
creased when fish curved
Very dull pigmentation;

skin coming away from flesh
Milky
Slightly soft
Yellowish grey, opaque mucus

Soft
Silvery
Silvery, slightly red

or brown
Convex and slightly sunken;
dark pupil; slightly
opalescent cornea
Less bright colour,
paler at edges.
Transparent mucus
No smell or seaweed.
Neutral smell
Brownish and extensive

seepage of blood from vessels
Flat; blurred pupil;
blood seepage around the eye
Yellowish
Convex, bulging; blue-black

bright pupil, transparent
eyelid
Uniformly dark red to purple.
No mucus
Fresh seaweed;
pungent; iodine
Concave in the centre;

grey pupil; milky cornea
Becoming thick discoloured

opaque mucus
Yellowish; milky mucus
Slightly sulphurous fatty

smell, rancid bacon cuttings
or rotten fruit
Rotten sour
.O
zden
Table 2 Quality index scheme for sardine

Parameters being
assessed
0 point
1 point
2 points
3 points
Appearance
Skin
Slime
Stiffness
Eyes clarity
Eyes shape
Eyes iris
Very bright
Firm or elastic
Absent
Pre-rigor
Clear
Normal
Visible, black
Slightly dull
Dull
Gills colour
Gills mucus
Dark red
Absent
Bright
Soft
Slightly slimy
Rigor
Slightly cloudy
Slightly sunken
Not visible,
black
Red
Slight
Gills smell
Fresh oily,
metallic seaweed
Seaweed
Belly cavity odour
and 3 = unacceptable) (Karl et al., 2001). The scores

of the ve assessors were averaged to give the panel
mean score. The sh were judged unt for consumption
when the mean value for sensory score was below 3.
Chemical analysis
Proximate analyses
The sh samples were analysed in triplicate for proximate composition: lipid content of sardine by acid
hydrolysis method of AOAC 948.15 (AOAC, 1998a),
moisture content by the Mattissek et al. (1992) method,
the ash content by the AOAC (1998b) method, total
crude protein by the Kjeldhal method (AOAC, 1998c)
and the carbohydrate content of sh by the Merrill &
Watt (1973) method.
pH
Aliquots of ten sardines were lleted from each sampling

lot and homogenised using a food processor. Samples
were prepared according to Manthey et al. (1988) by
blending 5 g of the homogenate with 5 mL distilled water
for 1 min at room temperature in an Ultra-Turrax (IKA T
25 Basic, Staufen, Germany). pH was monitored using a
WTW-pH-meter (inoLab pH Level 1 model, Weilheim,
Germany). Determinations were carried out in triplicate
at days 1, 3, 5, 7 and 9 during the storage period.
Determination of total volatile basic nitrogen
Total volatile basic nitrogen (TVB-N) was determined

according to the method Antonocoupoulos & Vyncke
(1989). For TVB-N, sh muscle (10 g) was homogenised
with 6% perchloric acid (90 mL) for 1 min in an UltraTurrax . The homogenates were ltered through a lter
paper (Whatman No. 1, Middlesex, UK) and ltrates
alkalised by NaOH (20%) before distillation duplicate
ltrates were distilled in a Velp Mark apparatus (Model
Slimy
Post-rigor
Cloudy
Sunken
Bloody, grey
Dark brown or grey
Moderate
Fishy
Stale
Fishy
Metallic
Excessive
or sticky
Spoilt
UDK 140, Milan, Italy). The distillate titrated with

0.01 N HCl.
Determination of trimethylamin nitrogen
Trimethylamin nitrogen (TMA-N) was determined by

the method of AOAC (1998d). Homogenised samples
(10 g) were weighed, blended with 90 mL of 7.5%
trichloroacetic acid (TCA) solution and ltrated. Blended solution was xed with formaldehyde (20%). Four
millilitres of extract were transferred into test tubes and
1 mL formaldehyde, 10 mL anhydrous toluene and
3 mL K2CO3 solutions were added. The tubes were
shaken and 5 mL toluene layer was pipetted. Five
millilitres of picric acid working solution (0.02%) were
added. The contents were mixed and transferred to a
spectrophotometric cell. Absorbance at 410 nm against
the blank was measured. At the same time, standards
were prepared and measured. Results of TVB-N and
TMA-N were expressed as mg per 100 g of muscle.
Determination of lipid oxidation
Peroxide value (PV) was determined by a titrimetric

Wheeler method (Mattissek et al., 1992) after fat
extraction (Ozden et al., 2001). Results were expressed
as millimole of O2 per kilogram of lipid. The thiobarbituric acid (TBA) was determined according to Weilmeier & Regenstein (2004) and Khan et al. (2006). Two
grams of the homogenised sample were placed in a
50-mL centrifuge tube to which 16 mL of a 5-% (w v)
solution of TCA and 100 lL butylated hydroxytoluene
were added and in an Ultra-Turrax, homogenised at
high speed for 2 min. The mixture was ltered through a
Whatman No. 1 lter paper. One millilitre of a 0.01-m
aqueous solution of 2-TBA and 5 mL of the ltrate were
mixed. The mixture was heated in boiling water bath for
40 min, cooled to room temperature, and the absorbance of the resultant coloured solution was read at
1551
1552
. O
zden
532 nm using a Shimadzu spectrophotometer Model

1601 (Shimadzu, Kyoto, Japan). TBA values (expressed
as milligram malonaldehyde equivalents per kilogram of
sh meat) were calculated by multiplying the absorbance
readings by a factor of 10.2, which was obtained from a
standard line prepared using 1,1,3,3-tetraethyoxypropane as a precursor of malonaldehyde.
regression analysis was performed to determine the

correlation between analyses (Sumbuloglu & Sumbuloglu, 2002).
Determination of histamine
For sardine, the chemical composition values were

determined as follows: moisture 63.63 2.68%, ash
2.62 0.11%, total protein 21.82 0.86%, total fat
2.29 0.69% and total carbohydrate 1 0.34%. This
is in agreement with the conclusions made by literature
data (Gokoglu et al., 1998; Pacheco-Aguilar et al., 2000;
Ozogul et al., 2004).
Histamine content was determined in triplicate by the

enzyme-linked immunosorbent assay (ELISA) methods
(Staruskiewiez & Rogers, 2001), with commercial rapid
histamine test kits (Veratox histamine test kits from
Neogen, Lansing, USA and Canada). Results were read
using a microwell reader (Neogen Stat Fax 303, Palm
City, Florida, USA) with a 650-nm lter. Results were
expressed as milligram per kilogram.
Microbiological analyses
Samples (25 g) were mixed with 225 mL of pepton water

(Merck, Cat No. 107228, Darmstadt, Germany) diluents
in a Stomacher (Stomacher, IUL Instrument, Barcelona,
Spain). Further serial dilutions were made in tubes
before plating. The media and incubations were: plate
count PCA agar (PCA, Merck, Cat No. 105463), 37 C,
12 days for mesophilic aerobic; PCA agar (PCA,
Merck, Cat No. 105463), 7 C, 10 days for psychrotrophic bacteria (Baumgard, 1986); cetrimide agar
(Merck, Cat No. 105284), 37 C, 2 days for Pseudomonas spp. (Pichhardt, 1998); Iron agar (Peptone from
casein 5 g; yeast extract 2.5 g; glucose 1 g; agaragar
14 g; 0.3 g iron III citrate; 0.48 g sodium thiosulphate;
3 g NaCl), 25 C, 2 days for H2S-producing bacteria
(typical of Shewanella putrefaciens) (Lagrange et al.,
2003); and VRGB agar (Merck, Cat No. 110275), 37 C,
1 day for Enterobacteriaceae. Results are expressed as a
logarithm of colony forming units (log CFU) per gram
of sample. Sulphide reducing Clostridias were determined in Dierential Reinforced Clostridial Broth
(Merck, Cat No. 111699) after incubation at 37 C for
2 days, by the most probable number (MPN, 3 tubes
per dilution) method. Results were expressed as
log MPN per gram of samples. Anaerob Clostridias
were investigated in the same way, except that an
anaerobic atmosphere kit (Merck, Cat No. 113829) was
placed, together with the plates, inside the anaerobiosis
jar (Merck, Cat No. 116387) (Pichhardt, 1998; Merck
Microbiology Manual, 2002).
Signicant dierences between the samples were calculated by Excel XP 2003 by one-way analysis of variance
(anova) using a signicance level of P < 0.05 by
Tukeys honestly signicant dierence test. Pearsons
Proximate analysis
Sensory analysis
The results obtained with the EC and quality index

scheme score are presented in Table 3. Appearance of
sh was rm and bright on the rst day of the storage,
whereas on the following days, especially the 4th day,
they became soft, pale and o-odoured. It can be seen
(Table 3) that excellent and very good grades (E and A)
were scored during the rst 3 days of storage for whole
and gutted sh. Moderate grades (B) were obtained
between days 5 and 7 of storage for both the groups.
Unt for sale raw sh samples (C) were obtained by
9 days of storage for whole ungutted and gutted sardine
samples. In the present study, however, whole ungutted
and gutted sardines were given a grade of E for up to
1 day, a grade of A for up to 5 days and a grade of B for
up to 9 days of storage. The observed shelf life of
sardine, as determined by panellists who indicated that
the sh were unacceptable, was 7 days in whole ungutted sardines (demerit score: 20.10) and 7 days in gutted
sardines (demerit score: 20.23). Not signicant dierence (P > 0.05) was observed between whole and
gutted sardine samples during storage. Acceptability
scores for odour and taste of whole ungutted and gutted
sardines decreased (signicant, P < 0.05) during the
storage. The corresponding limit for taste and odour
was reached after 7 days for the two groups samples.
Not signicant dierence (P > 0.05) was observed
between the groups after 7 days of storage. Other
researchers have reported that sardine could not be
accepted after 6 days at +4 C (Gokoglu et al., 1998),
9 days at 0 C (El Marrakchi et al., 1990), 6 days at
0 C and 4 days at 4 C (Gokoglu & Yerlikaya, 2004).
Similar sensory results (8 days) have been reported for
iced sardine (Gennari et al., 1999). Nunes et al. (1992)
and Pacheco-Aguilar et al. (2000) reported that the limit
of acceptability for whole sardine was 5 days. Chemical
composition of sh, catching method, the condition of
the sh before catching or harvesting, storage temperature and seasons may aect the results of sensory
.O
zden
Table 3 Changes in sensory indices in sardine during the storage period in ice
Storage time (days)
1
EC score
Whole
2.76 0.15 (E quality)
Gutted
2.79 0.19 (E quality)
Quality index score
Whole
2.36 0.15
Gutted
2.4 0.09
Odour score of cooked fish
Whole
9.86 0.09 (excellent)
Gutted
9.86 0.10 (excellent)
Taste score of cooked fish
Whole
10 0 (excellent)
Gutted
10 0 (excellent)
2.1 0.09 (A quality)

2.12 0.11 (A quality)
5.53 0.88
5.6 0.59
1.72 0.08 (B quality)

1.77 0.12 (B quality)
13.26 1.12
13.8 1.95
1.28 0.10 (B quality)

1.32 0.06 (B quality)
0.43 0.07 (C quality)

0.49 0.08 (C quality)
20.10 1.16 (unacceptable)


8.53 0.45 (very good)

8.03 0.51 (very good)
7.46 0.61 (good)

7.53 0.35 (good)
5.66 0.43 (sufficient)


8.4 0.71 (very good)

8.16 0.63 (very good)
7.53 0.29 (good)

6.93 0.37 (good)


(Schormuller, 1968; Kietzmann et al., 1969). Papadopoulos et al. (2003) determined the eect of gutting on
microbiological, chemical and sensory properties of
aquacultured sea bass stored in ice. Results of this
study indicate that the shelf life of whole ungutted and
gutted sea bass stored in ice as determined by the overall
acceptability sensory scores and microbiological data is
13 and 8 days, respectively. Similar results for assessment of shelf life of other sh have been reported.
Chytiri et al. (2004) reported a shelf life for ungutted
and lleted trout of 1510 days, whereas Taliadourou
et al. (2003) reported a shelf life for lleted and ungutted
sea bass of 812 days. Paleologos et al. (2004) reported
the unacceptable quality for whole sea bass of 13 days,
gutted sea bass of 11 days and lleted sea bass of 9 days.
Karl et al. (2001) determined the inuence of gutting on
Table 4 Changes in chemical indices in
sardine during the storage period in ice
the sh quality of tench (Tinca tinca) stored in ice. These

authors found for tench no dierences between gutted
and ungutted sh shelf life and quality. Similar results
have been reported for iced sea bass (Erkan & Ozden,
2006). Our ndings agree with these literatures. Results
of this study indicate that shelf life of whole ungutted
and gutted sardines stored in ice as determined by the
sensory scores is 7 days. Not signicant dierence
(P > 0.05) was observed between whole ungutted and
gutted groups entire storage.
Chemical analysis
The changes in pH, TVB-N and TMA-N for whole

ungutted and gutted sardines during the 9 days of
storage period in ice are shown in Table 4. At the
Storage time (days)

1
pH
Whole
6.01 0.003
6.08 0.003
Gutted
6.02 0.01
6.09 0.003
TVB-N (mg per100 g of muscle)
Whole
11.11 0.59
17.40 1.15
Gutted
10.18 1.01
15.81 0.30
TMA-N (mg per100 g of muscle)
Whole
2.5 0.07
3.1 0.13
Gutted
2.6 0.02
2.1 0.04
Histamin (mg kg)1)
Whole
12.30 0.29
48.35 1.84
Gutted
12.30 0.71
29.44 2.62
PV (mmol of O2 per kilogram)
Whole
3.34 0.13
6.43 0.21
Gutted
3.22 0.09
6.67 0.12
TBA (mg MDA equivalents per kilogram)
Whole
2.86 0.16
8.23 0.26
Gutted
2.7 0.06
10.91 0.36
6.18 0.01
6.07 0.01
6.16 0.01
6.16 0.01
6.27 0.003
6.20 0.003
20.83 1.26
18.05 0.02
22.03 0.03
12.87 0.66
29.23 2.73
15.03 0.72
3.44 0.09
3.64 0.5
3.86 0.02
1.87 0.02
4.16 0.21
2.36 0.06
49.48 2.12
44.56 3.01
51.73 1.26
41.06 1.46
52.80 2.91
45.80 2.41
6.67 0.15
11.14 0.28
12.68 0.36
17.51 0.46
14.87 0.54
22.8 0.41
18.74 0.14
19.28 0.16
17.53 0.27
20.12 0.01
21.54 0.01
21.54 0.01
1553
. O
zden
beginning of the storage, pH values of whole ungutted

and gutted sardines were determined as 6.016.02. pH
values increased (P < 0.05) according to time of
storage. At the end of the storage period of 9 days,
TVB-N values reached 6.27 and 6.20 for whole ungutted
and gutted sardines, respectively. Increases in pH
indicate the accumulation of alkaline compounds, such
as ammonia mainly derived from microbial action. El
Marrakchi et al. (1990) found that the pH value of
sardine (Sardina pilchardus) stored in ice was 5.8 at day
0 and 6.36 at day 9. The pH increases are in agreement
with the ndings of Scott et al. (1986), Manthey et al.
(1988), Nunes et al. (1992), Ryder et al. (1984) and
Simeonidou et al. (1998) for other sh species stored in
ice. pH value increased according to storage time, but
pH value is not a criterion of spoilage. It has to be
supported by other chemical and sensory analyses
(Schormuller, 1968; Ludor & Meyer, 1973; Scott et al.,
1992).
TVB-N was found to be a good index quality decay of
iced sh: at the time of sensory rejection (Kietzmann
et al., 1969; Oehlenschlager, 1981), the value for TVB-N
coincided with limit of acceptability (20 mg per 100 g)
xed for Sikorski et al. (1990) for certain categories of
sh, such as fatty sh (sardine, mackerel and herring).
At the beginning of the storage, TVB-N values of
ungutted and gutted sardines were determined as
11.11 mg per 100 g and 10.18 mg kg)1, respectively. In
the whole ungutted group samples, TVB-N values
increased according to time of storage. At the end of
the storage period of 9 days, TVB-N values reached
29.23 mg per 100 g, respectively. The legal limits set for
these indexes at 20 mg per 100 g for TVB-N were not
exceeded throughout storage in gutted sardine samples.
Similar results have been reported by Ozogul et al.
(2004) for sardine under air conditions as 4 days of
storage at 4 1 C. A correlation was found between
total microbial count and TVB-N for whole ungutted
and
gutted
samples
(rwhole
ungutted = 0.9633,
rgutted = 0.5248 for psychrotrophic bacteria count;
rwhole ungutted = 0.7684, rgutted = 0.7369 for mesophilic
aerobic bacteria count). In this study, we determined the
TVB-N value of whole sardine higher than the eviscerated sardine esh because of enzymatic and bacterial
activity in internal organs. The Monterey sardine, like
other small pelagic sh, is susceptible to rapid autolytic
degradation of abdominal tissue after capture; this
process is caused mainly by proteases from the digestive
tract (Castillo-Yanez et al., 2004)
Trimethylamine oxide (TMAO), found in a large
number of marine sh, is broken down to trimethylamine (TMA) by either endogenous enzymes or the
bacterial enzyme trimethylamine oxidase (Ashie et al.,
1996; Debevere et al., 2001). The quantitative level of
TMA in sh is considered as a major index of the quality
of marine sh (Zhang et al., 2003). Initial average values
were 2.5 and 2.6 mg of TMA-N per 100 g of muscle for

whole ungutted and gutted samples, respectively, with
nal values of 4.16 and 2.36 mg of TMA-N per 100 g of
muscle. Values of TMA-N were statistically signicant
(P < 0.05) on the 7th and 9th day of storage. The
rejection limit is usually 510 mg of TMA-N per 100 g
of muscle; however, in numerous fatty sh, the
concentration of TMA-N never reaches the limit of
5 mg of TMA-N per 100 g of muscle (El Marrakchi
et al., 1990; Sikorski et al., 1990; Ababouch et al.,
1996). El Marrakchi et al. (1990) reported a moderate
freshness when muscle of Morocco sardine (Sardina
pilchardus) reached 4.8 mg of TMA-N per 100 g of
muscle after 9 days of ice storage. The lower TMA-N
value of gutted sardine samples after 5 days of storage
related to the low level of bacterial counts (Figs 15). In
this study, the TMA-N value of gutted sardine was
found to be lower than that of whole ungutted sardine
esh because of the eviscerated of internal organs.
Biogenic amine formation in seafood is important as
histamine, and possibly other biogenic amines are
responsible for scombrotoxic sh poisoning. Furthermore, biogenic amines have been used as chemical
indicators of seafood quality (Erkan, 2004). Initial
histamine values were 12.30 mg kg)1 for both the
groups of sardine in ice. The maximum levels were
52.80 mg kg)1 for whole sardine and 45.80 mg kg)1 for
gutted sardine after 7 days of storage in ice. There were
signicant dierences (P < 0.05) in histamine levels
between the two groups on days 3 and 7. EC has set a
maximum mean value of 100 mg kg)1 among nine
samples of fresh or canned sh and 200 mg kg)1 among
the same number of samples of ripened product
(91 493 EWG). Histamine limit of acceptability of
decomposed sh in oil was established at 50 mg kg)1
Psychrotrophic bacteria count
5.6
5.4
Whole
Gutted
5.2
5.0
log cfu g1
1554
4.8
4.6
4.4
4.2
4.0
3.8
3.6
3.4
1
Storage time (days)
Figure 1 Changes in psychrotrophic bacteria count of sardine during

the storage period of ice.
.O
zden
Total aerob mesophilic bacteria count
6.2
5.8
6.0
5.8
Whole
Gutted
5.4
5.2
5.4
log cfu g1
log cfu g1
5.6
Whole
Gutted
5.6
Enterobacteriaceae
6.0
5.2
5.0
4.8
5.0
4.8
4.6
4.4
4.6
4.2
4.4
4.0
4.2
3.8
4.0
3.6
3.8
3.4
3.2
3.6
1
Storage time (days)

Figure 2 Changes in total aerobic mesophilic bacteria count of sardine
during the storage period in ice.
4.9
4.8
Figure 4 Changes in Enterobacteriaceae of sardine during the storage

period in ice.
H2S-producing bacteria counts
5.0
5
7
Storage time (days)
Pseudomonas spp.
4.4
Whole
Gutted
4.2
4.6
4.0
log cfu g1
log cfu g1
4.7
4.5
4.4
4.3
3.8
3.6
4.2
Whole
Gutted
4.1
3.4
4.0
3.9
1
Storage time (days)

Figure 3 Changes in H2S-producing bacteria count of sardine during
the storage period in ice.
(Veciana-Nogues et al., 1997). Many results were reported to inhibit histamine formation signicantly in sh
stored at low temperatures (Yamanaka et al., 1985).
Ababouch et al. (1991) reported that histamine reached
toxic levels after 610 days in four out of seven
experiments of sardines stored at 8 C and on ice.
Jhaveri et al. (1982) found initial histamine level of
20 ppm in Atlantic mackerel, respectively. They also
stated that the levels increased to value of 100 ppm in
15 days in ice. El Marrakchi et al. (1990) reported that
amount of histamine in sardine esh at the time of
rejection (12 days) in ice was 162 ppm. Jeya Shakila
et al. (2003) found a similar of histamine level (40
50 ppm) in sardine muscle stored at ambient temperature storage 12 h.
3.2
1
Storage time (days)

Figure 5 Changes in Pseudomonas spp. of sardine during the storage
period in ice.
Lipid oxidation is a major quality problem. It leads to

the development o-odours and o-avours in edible
oils and fat-containing foods, called oxidative rancidity
(Ashie et al., 1996). High-fatted sardine muscles are rich
in polyunsaturated fatty acids which are susceptible to
peroxidation. Free radicals react with oxygen to produce
fatty acid peroxides. The fatty acid peroxides are free
radicals which can attack another lipid molecule,
resulting in peroxide and a new free radical (Hamre
et al., 2003). The primary product of lipid oxidation is
fatty acid hydroperoxide, measured as peroxide value.
Peroxides are unstable compounds and they break down
to aldehydes, ketones and alcohols that are volatile
products causing o-avour in products. Peroxide and
1555
1556
. O
zden
TBA values are the major chemical indices of oxidative

rancidity (Kolokowska & Deutry, 1983; Lubis & Buckle,
1990; Perez-Villarreal & Howgate, 1991).
Initial peroxide values of whole ungutted and gutted
sardines were 3.343.22 mmol of O2 per kilogram,
respectively (day1). Literature reports initial PV for
small pelagic sh, of 2.12 mmol of O2 per kilogram of
lipid for sardine (Ozden & Gokoglu, 1997), from 0.1 to
0.15 mmol of O2 per kilogram of lipid for herring
(Clupea harengus) (Smith et al., 1980) and from 0.36 to
1.11 mmol of O2 per kilogram of lipid for fresh sardine
(Pacheco-Aguilar et al., 2000). Peroxide values reached
14.87 mmol of O2 per kilogram for whole ungutted
sardine and 22.8 mmol of O2 per kilogram for gutted
sardine after 9 days of storage. Ludor & Meyer (1973)
proposed the following PV scale as a basis for
determining the freshness of sh: PV = 02 mmol of
O2 per kilogram of very good, PV = 25 mmol of O2
per kilogram of good, pH = 58 mmol of O2 per
kilogram of acceptable, pH = 810 mmol of O2 per
kilogram of spoiled. Peroxide acceptable levels of
whole ungutted sardine samples increasing to
12.68 mmol of O2 per kilogram after 7 days and of
gutted sardine 11.14 mmol of O2 per kilogram after
5 days of storage at +4 C.
Thiobarbituric acid is the second breakdown product
of lipid oxidation. The temperature, the storage conditions and processing methods aected lipid oxidations
degree (Lie, 2001). TBA values of whole ungutted and
gutted sardines are shown in Table 4. At the beginning
of the storage period, TBA values of ungutted and
gutted sardines were determined as 2.86 and 2.7 mg of
MDA equivalents per kilogram, respectively. During the
storage period of 3 days, TBA values of ungutted and
gutted sardines determined 8.23 mg of MDA equivalents per kilogram and 10.91 mg of MDA equivalents
per kilogram of esh, respectively. At the end of the
storage period of 9 days, TBA values of the two groups
samples found to be 21.54 mg of MDA equivalents per
kilogram. A highly signicant correlation was found
between sensory data and PVTBA value (Peroxideodour rwhole ungutted = -0.9595, rgutted = -0.9635; Peroxide-taste rwhole ungutted = -0.9691; rgutted = -0.9784;
TBA-odour rwhole ungutted = -0.8520, rgutted = -0.8027;
TBA-taste rwhole ungutted = -0.8608, rgutted = -0.8671).
The concentration of TBA in freshly caught sh is
typically between 3 and 5 mg of MDA equivalents per
kilogram esh, but levels of 58 mg of MDA equivalents per kilogram of esh are generally regarded as the
limit of acceptability for sh stored in ice (Schormuller,
1968; Beltran & Moral, 1990; Nunes et al., 1992; Ozden
& Gokoglu, 1997). Lipid oxidation is one of the most
important factors responsible for quality deterioration
of fatty sh during the refrigerated storage. Lipid
oxidation in muscle foods can be initiated by nonenzymatic and enzymatic reactions. The rate and extent
of oxidative deterioration depends on factors such as the

storage period and temperature, saturation degree of
fatty acids, presence of antioxidants or pro-oxidants and
availability of oxygen. The highly unsaturated fatty
acids commonly found in sea food especially fatty sh
are particularly sensitive to oxidative changes during
storage (Serdaroglu & Felekoglu, 2005). Gutting process
inhibits microbial spoilage and increases lipid oxidation.
Gutting seems to aect rancidity levels in sea bass stored
in ice probably because of exposure of the lipid to
the atmospheric oxygen, which accelerates oxidation
(Papadopoulos et al., 2003).
Microbiological analysis
Bacterial spoilage in refrigerated sh and sh products under aerobic storage conditions is caused by
Gram-negative psychrotrophic organisms such as
Pseudomonas, Alteromonas, Flavobacterium spp. and
H2S-producing bacteria (including Shewanella putrefaciens) (Huss, 1994). Microbiological results are shown in
Figs 16. Initial psychrophilic bacteria counts of ungutted and gutted sardines were 3.53.8 log CFU g)1,
respectively (day 1). Initial mesophilic bacteria counts
of the two groups samples were found 43.8 log
CFU g)1. Mesophilic counts reached 6 log CFU )1 for
whole ungutted sardine and 5.25 log CFU g)1 for gutted sardine after 9 days of storage (Fig. 2). On day 9 of
storage, psychrophilic counts of ungutted and gutted
sardines were 5.37 and 5.32 log CFU g)1, respectively.
The microbial limit of acceptability at 106107 CFU g)1
for mesophilic aerobic bacteria (Gobantes et al., 1998)
and 105 CFU g)1 for psychrotrophic bacteria (LapaGuimaraes et al., 2000) were exceeded 7 days of storage.
Similar results for mesophilic aerobic counts have been
reported by Rehbein et al. (1994) for redsh as 12 days
of storage in ice. Eifert et al. (1992) reported hybrid
striped bass llets stored at 4 C did not reach
107 CFU g)1 until 12 days.
Initial H2S-producing bacteria counts were 4 log
CFU g)1 for two groups sh sampling periods, with
nal counts of 4.90 log CFU g)1 (whole sardine ) and
4.30 log CFU g)1 (gutted sardine), respectively (Fig. 3).
Counts of sulphide producers in the range of log 6
log 6.7 are normally present on rejectable sh from
temperate and tropical waters (Hanna, 1992). Kyrana &
Lougovois (2002) and Lougovois et al. (2003) reported
the time of rejection for sea bass and sea bream after 14
15 days of ice storage H2S-producing bacteria counts of
5 log CFU g)1. Papadopoulos et al. (2003), Taliadourou et al. (2003) and Paleologos et al. (2004) found the
H2S-producing bacteria count of whole ungutted sea
bass muscle lower than gutted and lleted sea bass
muscle.
At the beginning of the storage period, Enterobacteriaceae counts for ungutted and gutted sardines were
.O
zden
Sulphide reducing Clostridia's

1.6
Whole
Gutted
log MPN g1
1.4
1.2
1.0
0.8
0.6
0.4
1
(22.03 mg per 100 g), histamine (51.73 mg kg)1), PV

(12.68 mmol of O2per kilogram) the shelf life of whole
ungutted sardine stored in ice at +4 1 C was
considered less than 7 days. The acceptability limits of
TVB-N, TMA-N and histamine were not exceeded
throughout storage in gutted sardine samples. Gutted
sh are highly perishable because of presence of fatty
oxidation. The oxidation parameters of fat determined
of gutted sardine samples were higher than the whole
ungutted sardine samples. The principle of sh eviscerate process is to delay, reduce or inhibit the microbial
spoilage. In this study, we found the microbial count
reducing eects of gutting sardine stored in ice. The legal
limits of microbiological counts were not exceeded
throughout storage in whole ungutted and gutted
sardine samples.
Storage time (days)

Figure 6 Changes in sulphide reducing Clostridias during the storage
period in ice.
determined as 3.5 log CFU g)1. At the end of the

storage period of 9 days, Enterobacteriaceae counts
increased to 5.08 log CFU g)1 and 4.30 log CFU g)1,
respectively (Fig. 4). The contribution of Enterobacteriaceae to the microora of sh and its spoilage potential
must be taken into consideration especially in the case of
polluted water or delay in chilling after catch (Chouliara
et al., 2004).
Initial Pseudomonas spp. counts of whole ungutted
and gutted sardine samples were 3.3 log CFU g)1. A
content of 4.30 log CFU g)1 was reached for the whole
samples after 5 days of storage, whereas the ungutted
samples reached this content (4 log CFU g)1) after
5 days of ice storage (not signicant, P > 0.05)
(Fig. 5). At the beginning of the storage, sulphidereducing Clostridias counts of ungutted and gutted
sardines were determined as 0.48 log MNP g)1. In the
two groups, sulphide-reducing Clostridias counts
increased according to time of storage. At the end of
the storage period of 9 days, sulphide-reducing Clostridias counts reached 1.36 log MNP g)1 and 0.56 log
MNP g)1 for ungutted and gutted sardines, respectively
(Fig. 6).
Conclusions
The aim of this study is to determine the sensory,

chemical and microbiological assessment of whole
ungutted and gutted sardines stored in ice. Results of
this study indicate that shelf life of whole ungutted and
gutted sardines stored in ice as determined by the
sensory scores is 7 days, respectively. Not signicant
dierence (P > 0.05) was observed between ungutted
and gutted groups in entire storage. According to TVB-N
Acknowledgment
This study was supported by the Research Fund of the

University
of
Istanbul (Project
No.
UDP590 12072005).
References
91 493 EWG. 31991L0493. Richtlinie 91 493 EWG des Rates vom 22.
Juli 1991 zur Festlegung von Hygienevorschriften fur die Erzeugung
und die Vermarktung von Fischereierzeugnissen.
Ababouch, L., Alal, M.E., Benabdeljelil, H. & Busta, F.F. (1991).
Quantitative changes in bacteria, amino acids and biogenic amines
in sardine (Sardina pilchardus) stored at ambient temperature (25
28C) and in ice. International Journal of Food Science and
Ababouch, L.H., Souibri, L., Rhaliby, K., Ouadhi, O., Battal, M. &
Busta, F.F. (1996). Quality changes in sardines (Sardine pilchardus)
stored in ice and at ambient temperatures. Food Microbiology, 13,
123132.
Antonocoupoulos, N. & Vyncke, W. (1989). Determination of volatile
basic nitrogen in sh. Zeitschrift fur Lebensmittel Untersuchung und
Forschung, 189, 309316.
AOAC (1998a). Ocial Method 948.15 Fat (crude) in Seafood Acid
Hydrolysis Method. Fish and other marine products, Chapter 35,
p. 9. (edited by J.M. Hungerford). In: Ofcial Methods of Analysis
of AOAC International (edited by P. Cunni). ISBN 0-93558454-4.
AOAC (1998b). Ocial Method 938.08 Ash of Seafood. Fish and
other marine products, Chapter 35, p. 6. (edited by J.M. Hungerford). In: Ofcial Methods of Analysis of AOAC International,
(edited by P. Cunni). ISBN 0-935584-54-4.
AOAC (1998c). Ocial Method 955.04 Nitrogen (Total) in
Seafood. Fish and other marine products, Chapter 35, p. 6.
(edited by J.M. Hungerford). In: Ofcial Methods of Analysis
of AOAC International, (edited by P. Cunni). ISBN 0-93558454-4.
AOAC (1998d). Ocial Method 971.14 Trimethylamine Nitrogen in
Seafood Colorimetric Method. Fish and other marine products,
Chapter 35, p. 7. (edited by J.M. Hungerford). In: Ofcial Methods
of Analysis of AOAC International (edited by P. Cunni). ISBN
0-935584-54-4.
Ashie, I.N.A., Smith, J.P. & Simpson, B.K. (1996). Spoilage and shelflife extension of fresh sh and shellsh. Critical Reviews in Food
1557
1558
. O
zden
Baumgard, J. (1986). Lebensmittel tierischer Herkunft, Feinkosterzeugnisse, gefrorene, tiefgefrorene und getrocknete lebensmittel,
Fertiggerichte, hitzekonservierte Lebensmittel, Speiseeis, Zucker,
Kakao, Zuckerwaren, Rohmassen. Mikrobiologische Untersuchung
von Lebensmitteln. Edt: Jurgen Baumgart, unter Mitarbeit von
Jurgen Firnhaber, Gottfried Spicher. ISBN: 3-922528-91-0 Hamburg: Behrs Verlag.
Beltran, A. & Moral, A. (1990). Gas chromatographic estimation of
oxidative deterioration in sardine during frozen storage. Lebensmittel-Wissenschaft und Technologie, 23, 499504.
Castillo-Yanez, F.J., Pacheco-Aquilar, R., Garcia-Garreno, F.L. &
Navarrete-Del Toro, M. (2004). Characterization of acidic proteolytic enzymes from Monterey sardine (Sardinops sagax caerulea)
viscera. Food Chemistry, 85, 343350.
Chouliara, I., Sawaidis, LN., Riganakos, K. & Kontaminas, M.G.
(2004). Preservation of salted, vacuum-packaged, refrigerated sea
bream (Sparus aurata) llets by irradiation: microbiological,
chemical and sensory attributes. Food Microbiology, 21, 351359.
Chytiri, S., Chouliara, I., Savvaidis, I.N. & Kontominas, M.G. (2004).
Microbiological, chemical and sensory assessment of iced whole and
lleted aqua cultured rainbow trout. Food Microbiology, 21, 157
165.
Debevere, J., Devlieghere, F., Von Sprundel, P. & De Meulenear, B.
(2001). Inuence of acetate and CO2 on the TMAO-reduction
reaction by Shewanella baltica. International Journal of Food
Microbiology, 68, 115123.
EC Regulation (1996). Council Regulation (EC) N 2406 96 of 26
November 1996 laying down common marketing standards for certain
shery products. (OJ L 334, 23.12.1996, Pp. 115).
Eifert, J.D., Hackney, C.R., Libey, G.S. & Flick, G.J. Jr (1992).
Aquacultured hybrid striped bass llet quality resulting from postharvest cooling or CO2 treatments. Journal of Food Science, 57,
10991102.
El Marrakchi, A.E., Bennour, M., Bouchriti, N., Hamama, A. &
Tagafatit, H. (1990). Sensory, chemical and microbiological assessments of Moroccan sardines (Sardina pilchardus) stored in ice.
Journal of Food Protection, 53, 600605.
Erkan, N. (2003). Verderb in Fisch und Fischwaren, chemische und
mikrobiologische Veranderungen und Gefahren. Rundschau fur
Fleischhygiene und Lebensmitteluberwachung, 55, 254258.
Erkan, N. (2004). Biogenen Amin Bedeutung in Fischqualitat und
Fischhygiene. Rundschau fur Fleischhygiene und Lebensmitteluberwachung, 56, 8285.
Erkan, N. & Ozden, O. (2006). Eect of gutting on quality and shelflife of aquacultured sea bass (Dicentrarchus labrax) stored in ice.
International Journal of Food Properties, 9, 331345.
Gennari, M., Tomaselli, S. & Cotrona, V. (1999). The micro ora of
fresh and spoiled sardines (Sardina pilchardus) caught in Adriatic
(Mediterranean) Sea and stored in ice. Food Microbiology, 16,
1528.
Gobantes, I., Choubert, G. & Go`mez, R. (1998). Quality of pigmented
(astaxanthin and canthaxanthin) rainbow trout (Oncorhynchus
mykiss) llets stored under vacuum packaging during chilled
storage. Journal of Agricultural Food Chemistry, 46, 43584362.
Gokoglu, N. & Yerlikaya, P. (2004). Use of eye uid refractive index in
sardine (Sardina pilchardus) as a freshness indicator. European Food
Research and Technology, 218, 295297.
Gokoglu, N., Ozden, O. & Erkan, N. (1998). Physical, chemical and
sensory analyses of freshly harvested sardines (Sardina pilchardus)
stored at 4C. Journal of Aquatic Food Product Technology, 7, 515.
Hamre, K., Lie, . & Sandnes, K. (2003). Development of lipid
oxidation and esh colour in frozen stored llets of Norwegian
spring-spawning herring (Clupea harengus L.): eects of treatment
with ascorbic acid. Food Chemistry, 82, 447453.
Hanna, J. (1992). Rapid microbial methods and fresh sh quality
assessment. In: Fish Processing Technology (edited by G.M. Hall).
Pp. 252254. ISBN: 0-7514-0280-X. London, UK: Blackie Academic
& Professional.
Huss, H.H. (1988). Fresh sh quality and quality changes. FAO

Fisheries Series, No. 29. Rome: FAO.
Huss, H.H. (1994). Assurance of seafood quality. FAO Fisheries
Technical paper, No. 334. Roma: FAO.
Jeya Shakila, R., Vijayalakshmi, K. & Jeyasekaran, G. (2003).
Changes in histamine and volatile amines in six commercially
important species of sh of the Thoothukkudi coast of Tamil Nadu,
India stored at ambient temperature. Food Chemistry, 82, 347352.
Jhaveri, S.N., Leu, S.S. & Constantinides, S.M. (1982). Atlantic
mackerel (Scomber scombrus, L.): shelf life in ice. Journal of Food
Science, 47, 18081810.
Karl, H., Meyer, C. & Munkner, W. (2001). Vergleichende Eislagerung
von ausgenommenen und unausgenommenen Schleien. Information
fur die Fischwirtschaft aus der Fischereiforschung, 48, 139143.
Khan, M.A., Parrish, C.C. & Shahidi, F. (2006). Eects of mechanical
handling, storage on ice and ascorbic acid treatment on lipid
oxidation in cultured Newfoundland blue mussel (Mytilus edulis).
Kietzmann, U., Priebe, K., Rakou, D. & Reichstein, K. (1969).
Seesch als Lebensmittel. Hamburg-Berlin: Paul Parey Verlag.
Kolokowska, A. & Deutry, J. (1983). Some comments on the
usefulness of 2-thiobarbituric acid (TBA) test for the evaluation of
rancidity in frozen sh. Nahrung, 27, 513518.
Kyrana, V.R. & Lougovois, V.P. (2002). Sensory, chemical and
microbiological assessment of farm-raised European sea bass
(Dicentrarchus labrax) stored in melting ice. International Journal
of Food Science and Technology, 37, 319328.
Lagrange, F., Jark, U., Etzel, V. & Feldhusen, F. (2003). Einuss des
Ausnehmens auf die Qualitat von frischem Seesch. Archiv fur
Lebensmittelhygiene, 54, 6570.
Lapa-Guimaraes, J., Aparecida Azavedo da Silva, M., Eduardo de
Felicio, P. & Contreras Guzman, E. (2000). Sensory, colour and
psychrotrophic bacterial analyses of squids (Loligo plei) during
storage in ice. Lebensmittel Wissenschaft und Technology, 35, 2129.
Lie, . (2001). Flesh quality the role of nutrition. Aquaculture
Research, 32, 341348.
Lougovois, V.P., Kyranas, E.R. & Kyrana, V.R. (2003). Comporasion
of selected methods of assessing freshness quality and remaining
storage life of iced gilthead sea bream (Sparus aurata). Food
Research International, 36, 551560.
Lubis, Z. & Buckle, A.K. (1990). Rancidity and lipid oxidation of
dried-salted sardines. International Journal of Food Science and
Ludor, W. & Meyer, V. (1973). Fische und Fischerzeugnisse.
Hamburg, Berlin: Paul Parey Verlag.
Manthey, M., Karnop, G. & Rehbein, H. (1988). Quality changes of
European catsh (Silurus glanis) from warm-water aquaculture
during storage ice. International Journal of Food Science and
Technology, 23, 19.
Mattissek, R., Schnepel, M. F. & Steiner, G. (1992). Lebensmittelanalytik, Grundzuge. Methoden. Anwendungen. ISBN: 3-540-54684-7
Zweite, Korrigierte Auage Berlin, Heidelberg, New York: Springer
Verlag.
Merck Microbiology Manual (2002). Sulte-reducing clostiridia. Available from: http://service.merck.de/microbiology (accessed 1 March
2005).
Merrill, A.L. & Watt, B.K. (1973). Energy value of Foods, basis and
derivation. Agriculture handbook No. 74, p. 2. United States
Department of Agriculture: Agriculture research service.
Nunes, M.L., Batista, I. & Morao de Campos, R. (1992). Physical,
chemical and sensory analysis of Sardine (Sardine pilchardus) stored
in ice. Journal of Science of Food and Agriculture, 59, 3743.
Oehlenschlager, J. (1981). Variation der gehalte an uchtigen stickstofgehaltigen basen und TVB-N in Rotbarsch. Information fur die
Fischwirtschaft, 53, 3334.
Ozden, O. & Gokoglu, N. (1997). Sardalya Balgnn (Sardina
pilchardus (Wal., 1792)Sogukta Depolanmas Srasnda Yagnda
_
Olusan Degisimlerin Incelenmesi.
Gda, 22, 309313.
.O
zden
Ozden, O., Kruse, R. & Erkan, N. (2001). Ubersicht uber Ruckstande

von Organochlor-Pestiziden Nitromoschusverbindungen und Polychlorierten Biphenylen in Turkischen Fischdauerkonserven.
Deutsche Tierarztliche Wochenschrift, 108, 158163.
Ozogul, F., Taylor, K.D.A., Quantick, P. & Ozogul, Y. (2000).
Chemical, microbiological and sensory evaluation of Atlantic
herring (Clupea harengus) stored in ice, modied atmosphere and
vacuum pack. Food Chemistry, 71, 267273.
Ozogul, F., Polat, A. & Ozogul, Y. (2004). The eect of modied
atmosphere packing and vacuum packaging on chemical, sensory
and microbiological changes of sardines (Sardine pilchardus). Food
Pacheco-Aguilar, R., Lugo-Sanchez, M.E. & Roblos-Bugueno, M.R.
(2000). Postmortem biochemical and functional characteristics of
monterey sardine muscle stored at 0C. Journal of Food Science, 65,
4047.
Paleologos, E.K., Savvaidis, I.N. & Kontominas, M.G. (2004).
Biogenic amines formation and its relation to microbiological and
sensory attributes in ice- stored whole, gutted and lleted
Mediterranean sea bass (Dicentrarchus labrax). Food Microbiology,
21, 549557.
Papadopoulos, V., Chouliara, I., Badeka, A., Savvaidis, I.N. &
Kontominas, M.G. (2003). Eect of gutting on microbiological,
chemical, and sensory properties, of aqua-cultured sea bass
(Dicentrarchus labrax) stored in ice. Food Microbiology, 20, 414420.
Perez-Villarreal, B. & Howgate, P. (1991). Deterioration of European
hake (Merluccius merluccius) during frozen storage. Journal of the
Science Food and Agriculture, 55, 455469.
Pichhardt, K. (1998). Lebensmittelmikrobiologie. Grundlagen fur die
Praxis. ISBN: 3540633804 Berlin, Heidelberg: Springer-Verlag.
Rehbein, H., Martinsdottir, E., Blomsterberg, F., Valdimarsson, G. &
Oehlenschlager, J. (1994). Shelf life of ice-stored redsh, Sebastes
marinus und Sebastes mentella. International Journal of Food Science
Ryder, J.M., Buisson, D.H., Scott, D.N. & Fletcher, G.C. (1984).
Storage of New Zealand jack mackerel (Trachurus novaezelandiae) in
ice: chemical, microbiological and sensory assessment. Journal of
Schormuller, J. (1968). Handbuch der Lebensmittelchemie (Band
III 2). Berlin, Heidelberg, New York: Springer Verlag.
Scott, D.N., Fletcher, G.C., Hogg, M.G. & Ryder, J.M. (1986).
Comparison of whole and gutted orange roughy stored in ice:
sensory, microbiology and chemical assessment. Journal of Food
Science, 51, 7986.
Scott, D.N., Fletcher, G.C., Charles, J.C. & Wong, R.J. (1992).
Spoilage changes in the deep water sh, smooth Oreo dory during
storage in ice. International Journal of Food Science and Technology,

27, 577587.
Serdaroglu, M. & Felekoglu, E. (2005). Eects of using rosemary
extract and onion juice on oxidative stability of sardine (Sardna
plchardus) mince. Journal of Food Quality, 28, 109120.
Sikorski, Z.E., Kolakowska, A. & Burt, J.R. (1990). Post harvest
biochemical and microbial changes seafood. In: Resources Nutritional Composition and Preservation (edited by Zdzisaw E. Sikorski).
Pp. 5575. Boca Raton Florida: CRC Press-Inc.
Simeonidou, S., Govaris, A. & Vareltzis, K. (1998). Quality assessment
of seven Mediterranean sh species during storage on ice. Food
Smith, J.G.M., Hardy, R.S., McDonald, I. & Temoleton, J. (1980).
The storage of herring (Cuplea harengus) in ice, refrigerated sea
water and at ambient temperature. Chemical and sensory assessment. Journal of the Science of Food and Agriculture, 31, 375385.
Staruskiewiez, W.F. & Rogers, P.I. (2001). Performance of Histamine
Test Kits for Applications to Seafood. Presentation on October 26,
2001. Vancouver, BC: Fourth World Fish Inspection and Quality
Control Congress.
Sumbuloglu, K. & Sumbuloglu, V. (2002). Biyoistatistik. Hatipoglu
Basm ve Yaym San. Tic. Ltd. Sti. ISBN: 975-7527-12-2. 10.
Ankara: Bask.
Taliadourou, D., Papadopoulos, V., Domvridou, E., Savvaidis, I. &
Kontominas, G.M. (2003). Microbiological, chemical and sensory
changes of whole and lleted Mediterranean aquacultured sea bass
(Dicentrarchus labrax) stored in ice. Journal of the Science of Food
and Agriculture, 83, 13731379.
Tulsner, M. (1994). Fischverarbeitung. Bd.1 Rohstoeigenschaften von
Fisch und Grundlagen der Verarbeitungsprozesse. ISBN: 3-86022-1965. Hamburg: Behrs Verlag.
Veciana-Nogues, M.T., Marine-Font, A. & Vidal-Carou, M.C. (1997).
Changes in biogenic amines during the storage of Mediterranean
anchovies immersed in oil. Journal of Agrcultural. Food Chemistry;,
45, 13851389.
Weilmeier, D.M. & Regenstein, J.M. (2004). Cooking Enhances the
Antioxidant Properties of Polyphosphates. Journal of Food Sciences,
69, FCT16FCT23.
Yamanaka, H., Itagaki, K., Shiomi, K., Kikuchi, T. & Okuzumi, M.
(1985). Inuences of the concentration of sodium chloride on the
formation of histamine in the meat of mackerel. Journal of Tokyo
University Fisheries, 72, 5156.
Zhang, Y., Lu, H. & Levin, R.E. (2003). Enhanced storage-life of fresh
haddock llets with stabilised sodium chlorite in ice. Food
Microbiology, 20, 8790.
1559
1560
Original article
Probiotic potential and sensory properties of coconut flan
supplemented with Lactobacillus paracasei
and Bifidobacterium lactis
Sabrina B. M. Correa,1 Inar A. Castro2 & Susana M. I. Saad1*
1 Department of Biochemical and Pharmaceutical Technology, Faculty of Pharmaceutical Sciences, University of Sao Paulo, Av. Prof. Lineu
Prestes, 580-B16, 05508-000 Sao Paulo, SP, Brazil
2 Department of Food and Experimental Nutrition, Faculty of Pharmaceutical Sciences, University of Sao Paulo, Av. Prof. Lineu Prestes, 580-B14,
05508-000 Sao Paulo, SP, Brazil
(Received 11 July 2006; Accepted in revised form 08 March 2007)
Summary
The eect of probiotic cultures on sensory performance of coconut an during storage at 5 C and the
viability of these micro organisms for up to 28 days were investigated. Sensory analyses of the product were
performed after 7, 14 and 21 days of storage. Coconut ans were produced with no addition of cultures (T1,
control), or supplemented with Bidobacterium lactis (T2), Lactobacillus paracasei (T3) and B. lactis + L. paracasei (T4). Populations of L. paracasei and B. lactis as single or in co-culture remained above
7 log CFU g)1 during the entire storage period. Viability of L. paracasei was higher for T3. All products
were well accepted and no signicant dierences (P > 0.05) were detected between the coconut ans studied.
The addition of L. paracasei and B. lactis to coconut an resulted in its having great potential as a functional
food, which has high sensory acceptability.
Keywords
Bidobacterium lactis, coconut an, interaction, Lactobacillus paracasei, probiotics, sensory acceptability, shelf life.
Introduction
The benecial eects of food with added live healthpromoting micro organisms (probiotics) on human
health, and in particular of milk products on children
and other high-risk populations, are being increasingly
promoted by health specialists (FAO WHO, 2001).
Foods that promote human health over and above the
provision of basic nutrition are called functional foods
(Halsted, 2003). A working denition which was
recently proposed describes functional foods as foods
that can be satisfactorily demonstrated to benecially
aect one or more target functions in the body, beyond
adequate nutritional eects, so as to lead to an improved
state of health and well-being and or a reduction of risk
of disease (Stanton et al., 2005). The concept has also
been directed towards food additives that may exert a
positive eect on the gut microbiota composition
probiotics and prebiotics (Ziemer & Gibson, 1998).
Gibson et al. (2004) reviewed the prebiotic concept as
a selectively fermented ingredient that allows specic
changes, both in the composition and or activity in the
*Correspondent: Fax: + 55-11-38156386;
e-mail: susaad@usp.br
gastrointestinal microbiota that confers benets upon

host well-being and health. Probiotics are dened as
live non-pathogenic microorganisms which, when consumed in adequate amounts, confer health benets on
the host (FAO WHO, 2001; Schirin & Blum, 2001;
Sanders, 2003), with an ongoing controversy as to
whether cultures must be viable for ecacy in all cases
(Stanton et al., 2005). Although it is recognised that
dead cells may mediate physiological benets, a dierent
term should be used to refer to these agents. Probiotics
are emerging as signicant dietary ingredients in the eld
of nutrition (Sanders, 2003). They have an important
growing global market, with recent estimates indicating
an annual turnover approaching US$50 billion. The
largest segment of this market in Europe, Japan and
Australia is made up of foods containing probiotics,
prebiotics and synbiotics (Talwalkar et al., 2004; Stanton et al., 2005).
Evidence for the role of probiotics in maintenance of
health or prevention of disease is increasing and is
supported, in some cases, by blind, placebo-controlled
human trials (Sanders, 2003). When ingested, probiotic
bacteria are resistant to gastric acidity and bile salts,
and, therefore, pass through the gastrointestinal tract,
where they may inuence the metabolism and the
doi:10.1111/j.1365-2621.2007.01585.x
Potentially probiotic coconut flan S. B. M. Correa et al.
equilibrium of the indigenous microbiota (Minelli et al.,

2004). The action of probiotics on intestinal microbiota
results in vital benets, including protection against
pathogens, immune stimulation, prevention and treatment of gastrointestinal disorders, metabolism of lactose
and positive eects on colonic health and host nutrition
(Charalampopoulos et al., 2002; Minelli et al., 2004).
Lactobacillus and Bidobacteria species are the microorganisms which are most commonly used as human
probiotics. In particular, strains of Bidobacterium
animalis, Bidobacterium lactis, Bidobacterium bidum,
Bidobacterium breve and Bidobacterium longum biotypes infantis and longum are often implemented in
probiotic products in combination with other lactic acid
bacteria (LAB) (Masco et al., 2005). Among the
Lactobacillus species, Lactobacillus casei is the most
widely studied, because of its eects on the reduction of
incidence and duration of various types of diarrhoea
and its capacity to modulate the immune system
(Kalantzopoulos, 1997). The majority of the strains of
L. casei were identied as members of the species
Lactobacillus paracasei, according to current nomenclature, based on DNA homology studies (Schillinger,
1999).
In an eort to expand the probiotic product range,
researchers and companies have endeavoured to manufacture probiotic foods that maintain viability of probiotic cultures (Stanton et al., 1998). Some examples are
yogurt and a number of types of cheeses (Hatting &
Viljoen, 2001; Mc Brearty et al., 2001; Akalin et al.,
2004; Boylston et al., 2004; Buriti et al., 2005b), among
other dairy products. Reports on non-fermented probiotic products are still few and far between but include
ice-cream, frozen soy dessert and chocolate mousse
(Alamprese et al., 2002; Heenan et al., 2004; AragonAlegro et al., 2007).
The market for functional foods has been growing
and is expected to grow further in a number of
countries. However, it is important to point out that
the real functional properties of these foods must be
reconciled with the sensory acceptance of the food that
is being developed (Castro et al., 2004). If the product
does not have or maintain acceptable organoleptic
qualities during its shelf life, it will not be supported
by consumers through repeated purchases. The commercial success of probiotic products ultimately depends
on taste and appeal to the consumer (Heenan et al.,
2004). Hedonic scales and preference tests are common
instruments used when decisions on market introductions are made (Hersleth et al., 2005).
Coconut an is one of the most traditional of
Brazilian desserts and has favourable properties for
the development of probiotic micro organisms, including pH and moisture content above 6% and 70%,
respectively. Moreover, bearing in mind the popularity
of the dessert, the benecial eects of probiotics and the
growing interest food companies have demonstrated in

developing new probiotic products, coconut an appears to be an excellent option as a probiotic nonfermented dairy product. No previous studies have
reported sensory changes during the shelf life of
probiotic coconut an. Therefore, the aim of the present
study was to examine the viability of probiotic cultures
(L. paracasei and B. lactis) added as single cultures or in
co-culture, during production of coconut an, and the
implications of their addition on the sensory performance of the product during storage at 5 C.
Chemical compounds and cultures
The following commercial ingredients were employed

for the production of coconut an: whole milk (Danone,
Sao Paulo, Brazil; ultra-high temperature [UHT]),
carrageenan gum LP-60 (CP Kelco, Limeira, Brazil),
sucrose (Uniao, Coopersucar Uniao, Limeira, Brazil),
corn starch (Unilever, Mogi Guacu, Brazil), grated
coconut (Ducoco, Itapipoca, Brazil) and coconut milk
(Ducoco). Bidobacterium lactis (BL-04 300 B; Danisco,
Dange, France) and L. paracasei subsp. paracasei (LBC
82; Danisco) cultures were employed, and both cultures
were freeze-dried commercial cultures for direct vat
inoculation (DVS).
Statistical design and coconut flan manufacture
A factorial (22) design was applied in this study, and the

variables involved in the production of coconut ans are
described in Table 1. Four pilot-scale coconut an
making trials, denoted T1, T2, T3 and T4, were
performed in triplicate (three genuine repetitions of
each trial were carried out on dierent days). The
B. lactis BL-04 (300 B; Danisco) and the L. paracasei
subsp. paracasei LBC 82 (Danisco) cultures (0.050 g)
were grown in 20 mL of milk for 2 h at 37 C, prior to
their addition to the product, either individually (T2 and
T3) or in co-culture (T4), in order to obtain concentraTable 1 Description of the four coconut an trials studied in the
present work
Coconut
flan trials
Lactobacillus
paracasei*
Bifidobacterium
lactis**
T1
T2
T3
T4
)
)
+
+
)
+
)
+
+, addition; ), no addition.
*L. paracasei subsp. paracasei culture (LBC 82; Danisco).
**B. lactis culture (BL-04 300 B; Danisco).
1561
1562
tions of a minimum of between 9 and 10 log CFU g)1 (6

and 7 log CFU g)1 in the nal product).
Each coconut an lot was produced to obtain 23 kg
of the nal product. Production was performed in 4-L
vats with three-fourths of the total weight of commercial
milk (total proportion 71% w w) heated to 5455 C,
and carrageenan gum was added (0.3% w w). The
mixture was heated to 6566 C, and sucrose (6.3%
w w) was added. The mixture was then heated to 91
92 C, following corn starch, previously diluted in the
remaining amount (one-fourth) of milk (4.0% w w),
addition. At this time, the temperature decreased to 85
86 C, and grated coconut was added (2.4% w w), and
heating was discontinued. When the temperature of the
mixture reached 77 C, coconut milk (16.0% w w) was
added, and the heating continued up to 80 C. In the
next step, the vat was cooled to 3637 C, the cultures
grown in milk (as described before) were added, and the
mixture was homogenised. The coconut an was then
packaged in individual plastic cups, each containing
40 g of the product, sealed with a metallic cover, and
stored at 5 C for up to 28 days.
Analysis
Coconut ans from each batch were used for the

analysis of the nal product (day 1), and after 7, 14,
21 and 28 days of storage. On each sampling day,
coconut ans from the same batch and trial were
unpacked and sampled for chemical composition, physico-chemical analysis and sensory evaluation of the nal
product. For the microbiological analysis, portions of
25 g were collected aseptically from two cups of coconut
an from each batch. For the physico-chemical analysis,
three cups of coconut an from each batch were
collected. Portions of each coconut an on day 1 of
storage were also collected for subsequent chemical
composition analysis of the nal product.
Physico-chemical analysis of coconut an
The pH values of coconut ans were determined in

duplicate samples with a pH meter Analyser Model
300 M (Analyser, Sao Paulo, Brazil) equipped with a
penetration electrode model 2A04-IF (Analyser). The
moisture content was determined from 5-g samples by
drying at 70 C under vacuum (Marconi MA030112,
Piracicaba, Brazil) for 24 h. The coconut an mean
compositions (ash, fat and protein content) were determined in the nal product (after one day of storage at
5 C) in triplicate grated samples. Ash was determined
according to AOAC 945.46. Protein (AOAC 991.20)
was estimated by measuring the coconut an nitrogen
content by the Kjeldahl method and multiplying it by a
conversion factor (6.38), after drying 5 g of coconut an
samples. Fat (AOAC 989.05) was determined through
lipids extraction with ethyl ether, using the Soxhlet
device (Association of Ocial Analytical Chemists,

2000). The carbohydrate content was calculated by
dierence to achieve 100% of total content.
Microbiological analysis of coconut an
The viability of L. paracasei and of B. lactis were

monitored, respectively, for coconut ans T3 and T4,
and for coconut ans T2 and T4, during the storage
period. For this purpose, 25-g portions of duplicate
coconut samples were blended with 225 mL of 0.1%
peptone water in a Bag Mixer 400 (Interscience, St.
Nom, France) and submitted to serial dilutions with the
same diluent.
Lactobacillus paracasei was counted by pour-plating
1 mL of each dilution in DeMan-Rogosa-Sharpe (MRS)
agar (Oxoid Inc., Ogdensburg, NY, USA), acidied to
pH 5.4 with acetic acid, after 3 days of anaerobic
incubation (Anaerobic System Anaerogen; Oxoid) at
37 C.
Bidobacterium lactis was counted by pour-plating
1 mL of each dilution in MRS agar to which lithium
chloride (2%) and sodium propionate (3%) were added
LP-MRS agar, according to Vinderola & Reinheimer
(1999), after 3 days of anaerobic incubation (Anaerobic
System Anaerogen; Oxoid) at 37 C for coconut ans
T2 and T4. For coconut an T4, B. lactis was also
counted by pour-plating 1 mL of each dilution in
modied MRS agar (Oxoid), to which a dichloxallin
(0.5% v v), lithium chloride (1.0% v v) and cysteine
hydrochloride (0.5% v v) solution was added after
3 days of anaerobic incubation (Anaerobic System
Anaerogen; Oxoid) at 37 C, in order to guarantee
absence of L. paracasei interference in B. lactis counts in
either medium.
Sensory analysis
Acceptability of all treatments was compared after 7,

14 and 21 days of storage, employing a nine-point
structured hedonic scale, where one represented
dislike intensely and nine like greatly. Twenty-four
untrained consumers from the University of Sao
Paulo, comprising teachers, students and sta, took
part in this study and were selected because of their
interest and coconut an-consuming habits. Approximately 40 g of an was presented in white plastic
cups, and the panel was asked to evaluate the threedigit coded samples. A balanced incomplete block
design was applied in this study (Hinkelmann &
Kempthorne, 1994). By this procedure, four samples
(t = 4) were divided into six blocks containing two
samples (k = 2), and each pair (k) was presented once
in both positions (AB and BA). Each block was
repeated eight times, resulting in 24 replicates for each
sample. Panellists used water to clean their palates
between samples and were also instructed to report
any observations of sensory characteristics for the
coconut an samples (e.g. acid or bitter avour, pasty

or spongy texture, yellowish appearance). Analyses
were carried out in individual booths between 3 and
4 p.m.
Table 2 Viability of Lactobacillus paracasei subsp. paracasei in

coconut ans T3 and T4 during storage at 5 C (see Table 1 for
description of trials T3 and T4)
Population of L. paracasei (log CFU g)1)
Trials
Initially, all variables had their normality and variance

homogeneity tested by dispersion graphs and Hartley,
Cochran and Barlett tests. Data were expressed as
mean sd. Dierences between the four treatments
during the shelf life period were evaluated by repeated
measures analysis of variance and Tukey Honestly
Signicantly Dierent (HSD) test. Factorial anova
(variance analysis) was applied to the 22 factorial design
in order to evaluate the interaction eect at the end of
shelf life. A P-value < 0.05 was considered signicant.
All statistical analyses were carried out using the
statistical package Statistica (version 6.1, Statsoft,
Tulsa, USA).
T3
Storage
(days)
Mean*
1
7
14
21
28
6.60Aa
7.15Bb
7.37Bc
8.37Bd
8.64Be
T4
(0.09)
(0.10)
(0.05)
(0.05)
(0.08)
Range**
Mean*
6.496.70
7.047.27
7.307.42
8.308.43
8.518.72
6.42Aa
6.45Aa
6.53Aa
6.87Ab
7.33Ac
Range**
(0.14)
(0.08)
(0.12)
(0.03)
(0.21)
6.266.54
6.326.54
6.436.76
6.816.91
7.037.56
*Mean values (with standard deviation in parentheses).

**Minimummaximum counts obtained for all samples analysed.
a,b,c,d
For each trial, within a column, different superscripts denote
significant differences (P < 0.05) between different days of storage.
A,B,C
Within a row, different superscript capital letters denote significant
differences (P < 0.05) between different trials.
Results
Physico-chemical properties of coconut flans during storage
Mean composition on dry matter of coconut ans T1,

T2, T3 and T4, on day 1 of storage (nal product), was
15.22% (w w) fat, 11.47% (w w) protein, 8.95% (w w)
ash and 64.36% (w w) carbohydrate. Figure 1 shows
the changes in the coconut ans T1, T2, T3 and T4 pH
mean values during storage for up to 28 days. Coconut
an T3 gave signicantly lower pH values during
storage (P < 0.05) when compared with T1, T2 and
T4. On the other hand, probiotic ans T4 gave
signicantly higher pH values up to the seventh day of
storage, compared with T1, T2 and T3, whereas T2 gave
signicantly higher pH values after 14 and 21 days of
storage (P < 0.05). The moisture of ans T1, T2, T3
and T4 showed no signicant dierence (P > 0.05)
between trials and during the shelf life period. Mean
7.0
T1
T2
T3
T4
6.8
6.6
pH
6.4
6.2
6.0
5.8
5.6
1
14
21
28
Days of storage
Figure 1 Mean pH values obtained for the different coconut an

trials studied during storage at 5 C (see Table 1 for description of
trials T1, T2, T3 and T4).
moisture values were always between 71.24% and

71.67%.
Viability of Lactobacillus paracasei and of Bidobacterium
lactis during coconut flan storage
The viability of L. paracasei during the storage of

coconut ans T3 and T4, and the minimum and
maximum populations obtained for all samples analysed
are presented in Table 2. Both ans T3 and T4 presented
mean average population around 6.5 log CFU g)1
during the rst day and an increase during the entire
storage, and populations were signicantly higher for T3
after 7 days of storage (P < 0.05). A signicant increase
(P < 0.05) in L. paracasei populations during all the
weeks of storage evaluated was observed for coconut
an T3, whereas for T4 the increase in populations was
signicant only after 21 days of storage (Table 2).
Table 3 shows the viability of B. lactis in coconut an
during storage of the product for up to 28 days for ans
T2 and T4, and the minimum and maximum populations obtained for all samples analysed. Coconut an T2
presented slightly but signicantly higher probiotic
populations after 1, 21 and 28 days of storage, and a
signicant decrease (P < 0.05) only after 28 days of
storage (Table 3). On the other hand, in coconut an
T4, for both types of agar employed, the B. lactis
population increased signicantly (P < 0.05) only after
7 days of storage, decreasing signicantly after 21 days
of storage (Table 3). Both agars employed reected a
tendency that the probiotic cells had to increase at the
beginning of storage of T4 and to decrease by the second
half of the storage period, and might be chosen for
counting Bidobacteria in the presence of Lactobacillus.
1563
1564
Table 3 Viability of Bidobacterium lactis
Population of B. lactis (log CFU g)1)
(LP-MRS agar) in coconut an trials T2 and

T4 during storage at 5 C (see Table 1 for
description of trials T2 and T4)
Trials
T2
Storage
(days)
1
7
14
21
28
T4
Mean*
7.46Ba
7.46Ba
7.41Aab
7.41Cab
7.36Bb
(0.06)
(0.06)
(0.03)
(0.09)
(0.05)
Range**
7.377.54
7.387.53
7.367.45
7.287.52
7.307.41
T4***
Mean*
7.23Aa
7.41Bb
7.41Ab
7.27Ba
7.26Aa
(0.03)
(0.04)
(0.07)
(0.06)
(0.09)
Range**
7.187.27
7.377.47
7.327.48
7.207.32
7.137.30
Mean*
7.21Aa
7.34Ab
7.36Ab
7.15Aa
7.17Aa
(0.10)
(0.03)
(0.08)
(0.02)
(0.08)
Range**
7.077.36
7.307.39
7.267.46
7.107.17
7.067.27
*Mean values (with standard deviation in parenthesis).

** Minimummaximum counts obtained for all samples analysed.
***Population determined using modified de Mann Rogosa Sharpe (MRS) medium.
a,b
For each trial, within a column, different superscripts denote significant differences (P < 0.05)
between different days of storage.
A,B,C
Within a row, for subsequent columns, different superscript capital letters denote significant
differences (P < 0.05) between different trials or between different media for the same trial.
reduction of the scores when both micro organisms

were present in the coconut an formulation (Fig. 2).
Discussion
Figure 2 Acceptability scores of coconut ans measured by sensory

analysis during the shelf life period (7, 14 and 21 days) based on
hedonic scale where one represented dislike intensely and nine
represented like greatly.
Sensory evaluation
Acceptability scores obtained from sensory analysis of

coconut ans are shown in Fig. 2. No signicant
dierences were observed between groups (P = 0.058)
during the shelf life period of 21 days (P = 0.243).
However, a tendency towards better scores was observed
for probiotic coconut ans compared with the control
product. The interaction eect of the B. lactis and
L. paracasei combination in the acceptability of ans at
the end of the shelf life time was also not signicant
(P = 0.081), but there was a tendency towards a
Probiotic foods have been popular with consumers

(Heenan et al., 2004). In order to obtain the benecial
eects of probiotic foods, a minimum probiotic therapeutic daily dose of 8 up to 9 log CFU has been
proposed, which corresponds to a 100 g daily intake of a
food product containing 6 up to 7 log CFU g)1 (Lee &
Salminen, 1995; Blanchette et al., 1996; Hoier et al.,
1999).
Coconut an seems to be a promising vehicle for the
potentially probiotic strains tested, when compared with
the probiotic products described in the literature, as a
rather high population (always above the range of
recommended levels of between 6 and 7 log CFU g)1)
of the probiotic strains during the entire shelf life of the
product was observed. Slightly higher B. lactis populations were achieved when the probiotic was added
individually (T2). When B. lactis was added in coculture with L. paracasei (T4) during coconut an
production, B. lactis populations remained almost the
same as when an isolated culture was added (T2).
Moreover, L. paracasei showed higher growth when
added individually during coconut an production, and
L. paracasei populations obtained for an T3 were
signicantly higher than for T4 already after 7 days of
storage, for which viability between 8.30 and 8.72
log CFU g)1 was achieved after 21 days of storage.
Therefore, a favourable interaction or protocooperation
between the B. lactis and the L. paracasei strains probably did not occur in coconut an T4.
Dierent results were reported by Vinderola et al.

(2000) on interaction between probiotic strains and with
starter cultures in Fresco cheese and survival of probiotics during the shelf life of the product. The authors
added probiotic concentrated cultures of Bidobacteria,
L. paracasei and Lactobacillus acidophilus simultaneously with the starters Streptococcus thermophilus
and Lactococcus lactis to the product and observed that
the viability of Bidobacteria was increased by the
presence of other probiotic bacteria, resulting in the
highest populations of all bacteria evaluated. Nevertheless, viability of Bidobacteria decreased up to 1 log
cycle throughout the 60-day storage period, although
the micro organism maintained populations of about
106 CFU g)1.
Buriti et al. (2005a) tested the same L. paracasei
subsp. paracasei (LBC82) strain employed in the present
study in Brazilian Minas fresh cheese. The authors
reported that the L. paracasei populations increased
from 6.61 log CFU g)1 on day 1 of storage at 5 C up to
8.22 log CFU g)1 on day 21 of storage when strain
LBC82 was added individually. When added in coculture with type O lactic culture, the L. paracasei
population increased from 6.79 log CFU g)1 on day 1 of
storage up to 8.44 log CFU g)1 on day 21. These
ndings were similar to results obtained in the present
study with coconut an with respect to the LBC82
employed as a single culture. Nevertheless, the authors
suggested a possible protocooperation between the
LBC82 strain and the type O lactic culture, an interaction which was not observed in the present study
between the same LBC82 strain and the Bidobacterium
strain employed.
Similar to what was observed in this study with a nonfermented product, Heenan et al. (2004) observed that
all cultures survived during the 6 months of storage trial
in populations of 107 CFU g)1 or higher, in a study with
a non-fermented frozen vegetarian dessert to which
L. paracasei subsp. paracasei, B. lactis and Lactobacillus
rhamnosus were added. Also studying a non-fermented
dairy product (pH between 5 and 6.48), Aragon-Alegro
et al. (2007) developed a potentially probiotic and
synbiotic chocolate mousse. As in the present study, the
strain was used for the supplementation of mousse was
the L. paracasei subsp. paracasei LBC82 strain. The
authors reported that the viability of the probiotic was
maintained above 7 log CFU g)1 over 28 days and that
the prebiotic ingredient inulin did not interfere in its
viability.
Mart n-Diana et al. (2003) described a study with
fermented goats milk containing Bidobacterium Bb-12,
L. acidophilus La-5 and S. thermophilus ST-20Y. The
authors observed that Bidobacterium growth and
viability were greatly enhanced by whey protein concentrate supplementation (WPC) and maintained levels
always above 107 CFU g)1. When WPC was not added,
there was no increase in bidobacterial populations,

which remained at the initial level of inoculum
(3 106 CFU g)1). In the present study, there was no
need for the supplementation of any additional protein.
According to Kos et al. (2000), it is better to introduce
the probiotic strain in a buered system such as milk or
milk protein-based foodstus. The authors observed
that milk, WPC and mucin may function as buering
agents and digestive protease activity inhibitors in vivo.
Besides, the addition of WPC to goats milk enhanced
the yoghurt textural characteristics (Herrero & Requena
2006). The pH values obtained for coconut an in the
present study were relatively higher, near to neutral,
possibly leading to a higher B. lactis and L. paracasei
growth, as well as increasing the chance of functioning
as a buering system in vivo.
Behaviour of Bidobacterium strains is strain-dependent, and the B. lactis BL-04 strain tested in the
present study turned out to be rather resistant to the
conditions employed. The same strain (previously
classied as B. longum) was also tested in a more
acidic product (pH between 4.3 and 4.8) petit-suisse
cheese, and showed good viability, presenting a population always above 7 log CFU g)1 (Maruyama et al.,
2006). In fact, Lankaputra & Shah (1995) studied the
survival of nine strains of Bidobacterium spp. in acidic
conditions (pH between 1.5 and 3.0) and concluded
that B. longum and Bidobacterium pseudolongum
survived better in acidic conditions than B. bidum.
More recently, Reilly & Gilliland (1999) evaluated four
strains of B. longum survival in terms of pH during
growth and found that one of the strains, B. longum
S9, was more stable than the others, regardless of pH
during growth.
Overall, most bidobacteria strains are sensitive to pH
values below 4.6. Therefore, for practical application,
the nal product pH value must be maintained above
4.6 to prevent the reduction of Bidobacteria populations (Modler et al., 1990; Laroia & Martin, 1991),
which could be conrmed in the present study, where the
high pH values (close to neutral) were favourable for the
survival of the B. lactis strain studied. Indeed, Heenan
et al. (2004) reported that the neutral pH of frozen soy
dessert was conducive to probiotic survival, as some
probiotic organisms are susceptible to inactivation when
stored in acidic conditions. Moreover, high survival of
Lactobacillus johnsonii La1 and of Lactobacillus rhamnosus GG in retail-manufactured ice creams (pH values
above 6.3) were also reported (Alamprese et al., 2002,
2005). It is possible that the continued pH decrease
during storage in coconut ans T3 and T4 in the present
study can be attributed to the presence of L. paracasei.
In fact, Kourkoutas et al. (2005) reported that the nal
pH of quince pieces with immobilised L. casei cells was
also lower after storage, as they supported L. casei
growth.
1565
1566
As for the sensory evaluation of coconut an in the

present study, most panellists commented on their low
amount of sucrose. In fact, prune syrup was not added
to the dessert as it usually is, since it might hide their
sensorial features. Although a tendency towards better
scores was observed for probiotic coconut ans, compared with the control product, and of reduction of the
scores when both micro organisms were present in the
coconut an formulation, no signicant dierences were
observed between the groups. Therefore, no evident
dierences resulting from the addition of single or mixed
cultures of bidobacteria and L. paracasei in the sensory
characteristics of coconut ans were observed. Similarly,
Aragon-Alegro et al. (2007) reported that the addition
of L. paracasei subsp. paracasei strain LBC 82 did not
interfere in the chocolate mousse sensory preference of
consumers. On the other hand, Alamprese et al. (2005)
observed that the triangle method highlighted a signicant dierence (P < 0.01) between ice-cream with
Lactobacillus rhamnosus GG cells and a control icecream produced without micro organisms.
Our results for coconut an agree with those observed
by Corbo et al. (2001) for Canestrato Pugliese cheeses.
The authors reported that the Canestrato Pugliese
cheese evaluation by sensory panellists after 56 days of
ripening for appearance, colour, mechanical characteristics, smell, taste and texture, indicated that all the
cheeses could be characterised as the traditional Canestrato Pugliese cheese, with no sensory dierences
evident from the addition of single or mixed bidobacteria cultures. In addition, Gardiner et al. (1998) and
Stanton et al. (1998) reported that Cheddar cheese with
high levels of the Lactobacillus adjuncts (L. paracasei,
added in co-culture with Lactobacillus salivarius and
L. paracasei, respectively) exhibited avour, texture and
composition comparable with those of the control
cheeses, indicating that the addition of lactobacilli had
no adverse eect on sensory criteria, and did not alter its
composition, as observed in the present study with
coconut an. The presence of Bidobacteria in cheeses
resulted in higher acetic and lactic acid concentrations,
but the minor dierences in sensory attributes of these
dairy products indicated that the Bidobacteria did not
exhibit extensive metabolic activity (Vinderola et al.,
2000). However, Jayamanne & Adams (2004) reported
that the incorporation of Bidobacteria in bualo curd
resulted in the production of total volatile fatty acids,
diacetyl and acetylmethyl carbinol, thereby improving
the sensory properties of the product.
Davidson et al. (2000) observed that fermented frozen
yoghurt supplemented with probiotic bacteria, B. longum and L. acidophilus, had little eect on frozen yoghurt
avour or compositional characteristics. Moreover,
frozen storage of the product had little or no eect on
culture survival, and bacterial cultures remained at
levels sucient to oer the suggested therapeutic eects.
On the other hand, Greek set type yoghurt produced

with L. paracasei subsp. tolerans exhibited the best
sensory properties, with a rich traditional smooth taste,
and the yoghurt produced with this strain as an adjunct
presented good physico-chemical properties, besides
reaching 7 log CFU g)1 or higher populations (Maragkoudakis et al., 2006). Kailasapathy (2006) also reported that yoghurts incorporating free and encapsulated
probiotic cultures did not substantially alter the overall
sensory characteristics of yoghurts.
Conclusions
Sensory changes during the shelf life of coconut an

have not been previously reported. Our results indicated
that all products obtained had a very acceptable
performance for the sensory panel. Moreover, viability
of both L. paracasei and B. lactis in coconut ans were
always above the recommended levels of 106
107 CFU g)1 during refrigerated storage, thus satisfying
this criterion established for a probiotic food. The
addition of L. paracasei and B. lactis to coconut an
therefore resulted in a product with great potential as a
functional food, and with excellent sensory features.
Nevertheless, further studies should be conducted in
order to verify the resistance of the potentially probiotic
strains added to coconut an towards gastric acidity and
bile salts.
Acknowledgments
The authors wish to thank the Conselho Nacional de

Desenvolvimento Cient co e Tecnologico (CNPq), the
Coordenacao de Aperfeicoamento de Pessoal de N vel
Superior (CAPES) and the Fundacao de Amparo a`
Pesquisa do Estado de Sao Paulo (FAPESP), for
nancial support and fellowships. The authors also wish
to thank Danisco for providing the cultures employed
and Prof. Dr Adalberto Pessoa Junior for his useful
suggestions.
References
Akalin, A.S., Fenderya, S. & Akbulut, N. (2004). Viability and activity
of bidobacteria in yoghurt containing fructooligosaccharide during
refrigerated storage. International Journal of Food Science and
Alamprese, C., Foschino, R., Rossi, M., Pompei, C. & Savani, L.
(2002). Survival of Lactobacillus johnsonii La1 and inuence of its
addition in retail-manufactured ice cream produced with dierent
sugar and fat concentrations. International Dairy Journal, 12, 201
208.
Alamprese, C., Foschino, R., Rossi, M., Pompei, C. & Corti, S. (2005).
Eects of Lactobacillus rhamnosus GG addition in ice cream.
International Journal of Dairy Technology, 58, 200206.
Aragon-Alegro, L.C., Alegro, J.H.A., Cardarelli, H.R., Chiu, M.C. &
Saad, S.M.I. (2007). Potentially probiotic and synbiotic chocolate
mousse. LWT Food Science and Technology, 40, 669675.
Association of Ocial Analytical Chemists (2000). Ofcial Methods of

Analysis of AOAC International, 17th edn. Gaithersburg: AOAC.
Blanchette, L., Roy, D., Belanger, G. & Gauthier, S.F. (1996).
Production of cottage cheese using dressing fermented by bidobacteria. Journal of Dairy Science, 79, 815.
Boylston, T.D., Vinderola, C.G., Ghoddusi, H.B. & Reinheimer, J.A.
(2004). Incorporation of bidobacteria into cheeses: challenges and
rewards. International Dairy Journal, 14, 375387.
Buriti, F.C.A., Rocha, J.S., Assis, E.G. & Saad, S.M.I. (2005a).
Probiotic potential of Minas fresh cheese prepared with the addition
of Lactobacillus paracasei. LWT Food Science and Technology, 38,
173180.
Buriti, F.C.A., Rocha, J.S. & Saad, S.M.I. (2005b). Incorporation of
Lactobacillus acidophilus in Minas fresh cheese and its implications
for textural and sensorial properties during storage. International
Castro, I.A., Tirapegui, J., Silva, R.S.S.F. & Cutrim, A.J.S. (2004).
Sensory evaluation of a milk formulation supplemented with n3
polyunsaturated fatty acids and soluble bres. Food Chemistry, 85,
503512.
Charalampopoulos, D., Wang, R., Pandiella, S.S. & Webb, C.
(2002). Application of cereal and cereal components in functional
foods: a review. International Journal of Food Microbiology, 79,
131141.
Corbo, M.R., Albenzio, M., De Angelis, M., Sevi, A. & Gobbetti, M.
(2001). Microbiological and biochemical properties of Canestrato
Pugliese hard cheese supplemented with bidobacteria. Journal of
Dairy Science, 84, 551561.
Davidson, R.H., Duncan, S.E., Hackney, C.R., Eigel, W.N. & Boling,
J.W. (2000). Probiotic culture surival and implications in fermented
frozen yoghurt characteristics. Journal of Dairy Science, 83, 666
673.
FAO WHO (2001). Evaluation of Health and Nutritional Properties of
Probiotics in Food Including Powder Milk with Live Lactic Acid
Bacteria. Cordoba, Argentina: Report of a Joint FAO WHO Expert
Consultation, Available at: ftp: ftp.fao.org es esn food probioreport_en.pdf (Accessed 2nd October 2005).
Gardiner, G., Ross, R.P., Collins, J.K., Fitzgerald, G. & Stanton, C.
(1998). Development of a probiotic Cheddar cheese containing
human-derived Lactobacillus paracasei strains. Applied and Environmental Microbiology, 64, 21922199.
Gibson, G.R., Probert, H.M., Van Loo, J., Rastall, R.A. &
Roberfroid, M.B. (2004). Dietary modulation of the human colonic
microbiota: updating the concept of prebiotics. Nutrition Research
Review, 17, 259275.
Halsted, C.H. (2003). Dietary supplements and functional foods: 2
sides of a coin? American Journal of Clinical Nutrition, 77(Suppl),
1001S1007S.
Hatting, A.L. & Viljoen, B.C. (2001). Yogurt as probiotic carrier food.
International Dairy Journal, 11, 117.
Heenan, C.N., Adams, M.C., Hosken, R.W. & Fleet, G.H. (2004).
Survival and sensory acceptability of probiotic microorganisms in a
nonfermented frozen vegetarian dessert. LWT Food Science and
Herrero, A.M. & Requena, T. (2006). The eect of supplementing
goats milk with whey protein concentrate on textural properties of
settype yoghurt. International Journal of Food Science and Technology, 41, 8792.
Hersleth, M., Ueland, O., Allain, H. & Ns, T. (2005). Consumer
acceptance of cheese, inuence of dierent testing conditions. Food
Quality and Preference, 16, 103110.
Hinkelmann, K. & Kempthorne, O. (1994). Design and Analysis of
Experiments, vol. 1. Pp.. 290296. New York, NY: John Wiley &
Sons.
Hoier, E., Janzen, T., Heriksen, C.M., Rattray, F., Brockmann, E. &
Johansen, E. (1999). The production, application and action of lactic
cheese starter cultures. In: Technology of Cheesemaking (edited by
B.A. Law). Pp. 99131. Boca Raton, FL: CRC Press.
Jayamanne, V.S. & Adams, M.R. (2004). Survival of probiotic

bidobacteria in bualo curd and their eect on sensory properties. International Journal of Food Science and Technology, 39,
719725.
Kailasapathy, K. (2006). Survival of free and encapsulate probiotic
bacteria and their eect on the sensory properties of yoghurt. LWT
Kalantzopoulos, G. (1997). Fermented products with probiotic qualities. Anaerobe, 3, 185190.
Kos, B., Suskovic, J., Goreta, J. & Matosic, S. (2000). Eect of
protectors on the viability of Lactobacillus acidophilus M92 in
simulated gastrointestinal conditions. Food Technology and Biotechnology, 38, 121127.
Kourkoutas, Y., Xolias, V., Kallis, M., Bezirtzoglou, E. & Kanellaki,
M. (2005). Lactobacillus casei cell immobilization on fruit pieces for
probiotic additive, fermented milk and lactic acid production.
Process Biochemistry, 40, 411416.
Lankaputra, W.E.V. & Shah, N.P. (1995). Survival of Lactobacillus
acidophilus and Bidobacterium spp. in the presence of acid and bile
salts. Cultured Dairy Products Journal, 30, 113118.
Laroia, S. & Martin, J.H. (1991). Eect of pH on survival of
Bidobacterium bidum and Lactobacillus acidophilus in frozen
fermented dairy desserts. Cultured Dairy Products Journal, 2, 1321.
Lee, Y.K. & Salminen, S. (1995). The coming age of probiotics. Trends
in Food Science and Technology, 6, 241245.
Maragkoudakis, P.A., Miaris, C., Rojez, P., Manalis, N., Magkanari,
F., Kalantzopoulos, G. & Tsakalidou, E. (2006). Production of
traditional Greek yoghurt using Lactobacillus strains with probiotic
potential as starter adjuncts. International Dairy Journal, 16, 5260.
Mart n-Diana, A.B., Janer, C., Pelaez, C. & Requena, T. (2003).
Development of a fermented goats milk containing probiotic
bacteria. International Dairy Journal, 13, 827833.
Maruyama, L.Y., Cardarelli, H.R., Buriti, F.C.A. & Saad, S.M.I.
(2006). Instrumental texture of probiotic petit-suisse cheese:
inuence of dierent combinations of gums. Ciencia e Tecnologia
de Alimentos, 26, 386393.
Masco, L., Huys, G., De Brandt, E., Temmerman, R. & Swings, J.
(2005). Culture-dependent and culture-independent qualitative
analysis of probiotic products claimed to contain bidobacteria.
International Journal of Food Microbiology, 102, 221230.
Mc Brearty, S., Ross, R.P., Fitzgerald, G.F., Collins, J.K., Wallace,
J.M. & Stanton, C. (2001). Inuence of two commercially available
bidobacteria cultures on Cheddar cheese quality. International
Minelli, E.B., Benini, A., Marzotto, M., et al. (2004). Assessment of
novel probiotic Lactobacillus casei strains for the production of
functional dairy foods. International Dairy Journal, 14, 723736.
Modler, H.W., McKellar, R.C. & Yaguchi, M. (1990). Biobacteria
and bidogenic factors. Canadian Institute of Food Science and
Technology Journal, 23, 2941.
Reilly, S.S. & Gilliland, S.E. (1999). Bidobacterium longum survival
during frozen and refrigerated storage as related to pH during
growth. Journal of Food Science, 64, 714718.
Sanders, M.E. (2003). Probiotics: considerations for human health.
Nutrition Reviews, 61, 9199.
Schirin, E.J. & Blum, S. (2001). Food processing: probiotic
microorganisms for benecial foods. Current Opinion in Biotechnology, 12, 499502.
Schillinger, U. (1999). Isolation and identication of lactobacilli from
novel-type probiotic and mild yoghurts and their stability during
refrigerated storage. International Journal of Food Microbiology, 47,
7987.
Stanton, C., Ross, R.P., Fitzgerald, G.F. & Van Sinderen, D. (2005).
Fermented functional foods based on probiotics and their biogenic
metabolites. Current Opinion in Biotechnology, 16, 198203.
Stanton, C., Gardiner, G., Lynch, P.B., Collins, J.K., Fitzgerald, G. &
Ross, R.P. (1998). Probiotic cheese. International Dairy Journal, 8,
491496.
1567
1568
Talwalkar, A., Miller, C.W., Kailasapathy, K. & Nguyen, M.H.

(2004). Eect of packaging materials and dissolved oxygen on the
survival of probiotic bacteria in yoghurt. International Journal of
Vinderola, C.G. & Reinheimer, J.A. (1999). ulture media for the
enumeration of Bidobacterium bidum and Lactobacillus acidophilus in the presence of yoghurt bacteria. International Dairy Journal,
9, 497505.
Vinderola, C.G., Prosello, W., Ghiberto, D. & Reinheimer, J.A.

(2000). Viability of probiotic (Bidobacterium, Lactobacillus
acidophilus and Lactobacillus casei) and nonprobiotic microora
in Argentinean fresco cheese. Journal of Dairy Science, 83, 1905
1911.
Ziemer, C.J. & Gibson, G.R. (1998). An overview of probiotics,
prebiotics and synbiotics in the functional food concept: perspectives
and future strategies. International Dairy Journal, 8, 473479.
Original article
Aromatic profiles of spray-dried encapsulated orange flavours:
influence of matrix composition on the aroma retention evaluated
by sensory analysis and electronic nose techniques
M. V. Galmarini,1,2* M. C. Zamora,1,2 R. Baby,1,3 J. Chirife2 & V. Mesina4
1 Consejo Nacional de Investigaciones Cient cas y Tecnicas (CONICET), Av. Rivadavia 1917 (C1033AAJ), Buenos Aires, Argentina
2 Facultad de Ciencias Agrarias, Universidad Catolica Argentina (UCA), Cap. Gral. Ramon Freire 183 (C1426AVC), Buenos Aires, Argentina
3 Centro de Investigaciones en Solidos (CINSO), Juan B. de Lasalle 4397 (B1603ALO), Villa Martelli, Argentina
4 Instituto de Investigaciones Cient cas y Tecnicas de las Fuerzas Armadas (CITEFA), Juan B. de Lasalle 4397 (B1603ALO), Villa Martelli,
Argentina
(Received 28 September 2006; Accepted in revised form 22 March 2007)
Abstract
Spray-dried orange avour encapsulated in dierent amorphous matrices composed of maltodextrin (MD)
and dierent combinations with; sucrose, trehalose, lactose, modied starch and gum arabic (ga); were
evaluated by sensory analysis and electronic nose (e-nose). With both techniques the avours encapsulated in
MD-sucrose and MD-lactosesucrose were perceived as similar. However, the e-nose did not detect any
dierences among the other matrices (MD-trehalose, MD, MD-sucrose at a dierent concentration and
MD-ga). On the contrary, sensory analysis was able to group MD-trehalose and MD describing them by:
woody, marmalade, syrup, citrus terpenes, and Vitamin C; MD-sucrose at 40% and 10% concentration and
MD-lactosesucrose were grouped and represented by powder juice, tangerine and pungency, while MD-ga
was dierentiated from the rest by the attributes peely, plastic, solvent and green. In this way, it was shown
that matrix composition determines the aromatic prole of spray-dried encapsulated orange avours.
Keywords
Electronic nose, orange encapsulated avours, sensory prole, spray-dried, sucrose, trehalose.
Introduction
Spray-drying is the most common technique to produce

avour powders from food avour emulsion; recipes for
spray-dried avours contain, in addition to the liquid
avour, carrier materials, such as maltodextrin (MD),
gum arabic and modied starch. Ingredients are mixed,
emulsied homogenized and spray-dried; water content
is reduced to below 5% and the avour is encapsulated
in an amorphous glassy carbohydrate matrix.
The requirements for an ideal spray-drying carrier
include a high degree of solubility, limited viscosity at
the 3545% solution solids range, emulsifying characteristics, good drying properties, non hygroscopic character, bland taste, non reactivity, and low cost. For these
reasons MDs are commonly used for this type of spraydrying applications. Also, the high glass transition
temperature (Tg) of low DE MDs provides good
product stability to the dried powder (Beristain et al.,
2002; Bhandari & Hartel, 2005).
*Correspondent: Tel Fax (5411) 4552-2711;
e-mail: mgalmarini@gmail.com
doi:10.1111/j.1365-2621.2007.01592.x
It is well known that, among other factors, the type of

carrier governs avour retention during the spraydrying process (Menting et al. 1970; Thijssen, 1971);
and for this reason disaccharides, such as sucrose or
lactose are sometimes included with MD in commercial
formulations to improve retention characteristics.
However, the low Tg of sucrose (Roos, 1995) may
adversely aect stability. Food powders containing
amorphous carbohydrates can undergo several physical
changes which lead to the deterioration of quality (Roos
& Karel, 1991). Recent work carried out on strawberry
and orange spray-dried powder avours showed that the
presence of sucrose in the carrier formulation (MD and
sucrose) aected storage stability in a negative way
(Busso Casati et al., 2007), because it reduces the glass
transition temperature dramatically when 40% of
sucrose is incorporated in the dry carbohydrate matrix.
Trehalose is a naturally occurring non-reducing disaccharide which consists of two glucose molecules linked in
a 1,1-position by a a-glycosidic bond. In the last years the
use of trehalose as a functional food additive has been
approved in many countries. The glass transition temperature for trehalose is much higher than that of sucrose
1569
1570
Aromatic profiles of encapsulated flavours M. V. Galmarini et al.
(115 C as opposed to 77 C for sucrose; Komes et al.,

2003) which would contribute to the physical stability of
spray-dried matrices containing trehalose instead of
sucrose. It is well known that during spray-drying
encapsulation of avours partial loss of volatile compounds usually occurs leading to an alteration of the
total volatile composition and or a variation of the ratio
of the dierent compounds in the spray-dried product.
Komes et al. (2003, 2005) observed that the addition of
trehalose to dehydrated strawberry and apricot purees
resulted in the lowest loss of total aroma as well as of
individual fruit volatiles when compared with sucrose.
In present work aromatic proles of orange oil
encapsulated in dierent spray-dried carbohydrate matrices are studied. Orange oil aroma is a complex mixture
of various chemical components, primarily (+)-limonene (>90%) responsible for the base sensory character
of the citrus oils. The aroma is determined by the
aldehydes, mainly octanal and decanal, and both citrals
and esters. The sesquiterpenes aldehydes, a and b
sinensal contribute particularly to the specic sweet
orange aroma (Wright, 1995). Even though odour
mixtures bear a strong resemblance in their character
to the quality characteristics of the individual components, no single compound possess the odour quality
characteristic of the blend (Lawless and Heymann,
1998). In addition, the retention of aroma compounds
during spray-drying encapsulation is expected to depend
on the interactions of aroma compounds and encapsulating components of the matrix. This results in changing the volatile prole and the sensory perception of
aroma upon reconstitution of the avour powder.
Unlike the traditional instruments (e.g. gas chromatography) which discriminate the dierent avour components, electronic noses (e-noses) measure the aroma
components as a whole, thus resembling the human
nose. However, it must be noted that the e-noses do not
measure total odour intensity, but rather give an
undened measure of the total volatiles, being them
aromatic or not. Some semi-specic metal oxides sensors
do exist, but the ability to, for example, detect the
character impact component of some aroma is still far
away. In this way, it is important to compare the study
with sensory analysis which gives a specic perception of
a particular avour while allowing a descriptive and a
quantitative evaluation of the dierent aroma com-
Sample
Maltodextrin
Modified
starch
1
2
3
4
5
6
78
72
40
50
64
40
4
6
7
4
pounds. This comparative and complementary analysis

was used by Baby et al. (1999) to determine odour
intensities of dierent dilutions of powder orange juice
using a sensory panel and an e-nose as measuring
instruments, observing that both intensities could t to
similar mathematical functions.
The aim of this work was to compare the aromatic
prole of spray-dried orange avours encapsulated in
dierent carbohydrate matrices by means of sensory
analysis and an e-nose. In addition, the eciency of
aroma retention of sucrose and trehalose (in a mixture
with MD) were compared in order to explore the
potential of this disaccharide as a carrier for producing
spray-dried avours.
Samples
Five commercial spray-dried encapsulated orange avours (samples 15) (5% water content) and one sample
specially prepared using trehalose were manufactured in
a avour company located in Buenos Aires, Argentina.
Table 1 shows the composition of the spray-dried
avours. Maltodextrin DE 12 (MD 12) accounts for
most of the amorphous dried matrix in samples 1 and 2;
while in samples 3, 4, 5 and 6 it has been partially
replaced by one or the other of the following disaccharides: sucrose, lactose and trehalose. Samples 1, 3, 4, 5
and 6 also have a small proportion of modied starch in
their composition while sample 2 has only MD and gum
arabic. The sample containing trehalose (number 6) was
specially prepared for this study because this disaccharide is not locally commercially used for this purpose yet.
Processing conditions during avour oil emulsication
(which can aect emulsion size) were similar for all
samples; the same for operating conditions of the
atomizer during spray-drying, to minimize eects on
particle size distribution of the spray-dried powder.
Sensory Analysis
Panel training
Twelve paid assessors (12 females, 1923-years old)

students of Facultad de Ciencias Agrarias, Ponticia
Universidad Catolica Argentina were trained in descripTable 1 Composition (% dry basis) of
spray-dried avours
Sucrose
Gum
arabic
Lactose
Trehalose
Aroma
40
5
10
21
40
16
20
13
18
18
13
tive analysis of orange avours (8 h). During training

period, judges performed the following tasks: (1) odour
identication using standard solutions (Table 2); (2)
attribute generation of the dierent orange samples with
the aid of standards (Table 2); (3) matching of aromas;
and (4) use of unstructured scales.
accounted for the dierences were listed and their

frequency of mention was analysed in order to obtain
a control sample for the descriptive analysis. This
control sample was the one with the highest frequency
of mention for the given attribute (Table 2).
Descriptive analysis was performed by a combination
of the quantitative descriptive analysis (QDA) method
(Stone & Sidel, 1993) and the dierence-from-control
test (2) (Meilgaard et al., 1991). The aid of standards
was provided in order to quantify the 14 attributes:
freshly squeezed, juice powder, citrus terpenes, vitamin
C, tangerine, candy, solvent, plastic, peely, marmalade,
woody, pungency, green, syrup. Six samples and a
control which was specic for each attribute were
presented in each session (seven sessions; 2 h long each,
approximately); assessors did not know that the control
sample was the same as one of the coded samples. For
every descriptor judges were asked to sni the headspace
generated in the control sample and then quantify the
given descriptor for the six samples by comparing the
aroma intensity with the control; samples were evaluated in duplicate. Evaluation was performed only by
Aroma proling
The experiment was divided in two phases: (1) Triangle

Test (ASTM, 1977) was performed to compare orange
avours and develop information about the characteristics of the samples. During two testing sessions (2 h
long each), assessors were required to pick the sample
which they believed was dierent and describe the
attributes responsible for that dierence. The powder
avour solutions (5% in distilled water) were placed
(15 mL) in sealed glass asks (35 mL capacity, 3 cm
diameter opening) and presented to assessors identied
with random three digit codes. Testing took place in
individual booths kept at a temperature of 22 2 C
and illuminated with red light in order to mask
dierences due to colour intensity. Descriptors which
Table 2 Denition, attribute and standard recipe used by the trained panel to describe orange avours
Attribute
Standard recipe
Definition
Control sample
Freshly squeezed
Filter paper soaked in pure orange essence oil

(Fresh orange, Firmenich, Argentina), placed
in a sealed glass flask
50 g of juice powder (orange, Tang, Argentina)
reconstituted in 100 ml of distilled water
Aroma evocative of natural, fresh,

recently prepared orange juice
Related to the aroma of orange

juice obtained from reconstituted
juice powder
Aroma associated to strong,
piercing, citric smell
Juice powder
Citrus terpenes
Vitamin C
Tangerine
Candy
Solvent
Plastic
Peely
Filter paper soaked in Gamma terpinene essence oil;

(Gamma terpinene, IFF, Argentina) placed in a
sealed glass flask
Orange flavoured Vitamin C (Redoxon, Argentina)
5 g of juice powder (Tangerine, Clight, Argentina)
Humidified orange flavoured acid candy (Sugus, Argentina)
placed in a sealed glass flask
10 g of juice powder (Tangerine, Clight, Argentina)
Fresh orange juice placed in a plastic recipient, heated for
2 minutes in a microwave oven and stored at room temperature.
Filter paper soaked in pure orange essence oil (Orange peel,
Ledesma, Argentina), placed in a sealed glass flask
Related to the aroma of medicine

and vitamin complex.
Aromatics related to tangerine juice
Characteristic odour of orange candy
Aroma associated to solvent and paint
Aroma associated to plastic containers.
Aromatics associated to orange peel
No control
could be
established
No control
could be
established
6
Marmalade
Orange juice, peel and sugar heated at 85C for 60 min
Aromatics associated to marmalade
Woody
Suggestive of the odour of tree bark
Pungency
Filter paper soaked in orange pure essence oil (Valencia Orange,

Ledesma, Argentina), placed in a sealed glass flask
Orange flavoured, effervescent antacid salts (Uvasal, Argentina)
Green
Syrup
No standard was used

No standard was used
Tingling sensation on the surface

of the nose mucosa
Characteristic odour of unripe orange.
Sweet
4
2
6
1571
1572
sning the headspace in order to be able to compare the

results with those obtained with the e-nose.
E-nose analysis
The E-nose used MOSES II (MOdular SEnsor System),

Tubingen, Germany, was equipped with eight tin
dioxide (SnO2) sensors and eight quartz microbalance
(QMB) sensors. Dierent doping of tin oxide sensors as
well as dierent polymeric coatings of QMB sensors
confer them higher sensitivity and special selectivity for
certain gases. Surface resistivity of SnO2 sensors changes
according to the oxidant or reducing properties of the
adsorbed gases. The quartz crystals of QMB are
operated in an oscillator circuit and in the absence of
analyte gas they oscillate with their usual frequency (12
14 MHz). If the polymer coating adsorbs selectively an
analyte gas (adsorption is temporary and reversible) the
crystal mass increases, thus reducing the frequency of
vibration. In consequence, the frequency change results
proportional to the analyte concentration (Mitrovics
et al., 1997).
In this experience, only the oxide sensors were used as
they allowed a better discrimination among samples.
Solutions at 5% in water of the six samples were
analysed. For each sample three vials containing 3 mL
of the solution were prepared, incubated at 40 C for
15 min with a 20 min interval between samples and
placed in a headspace sampler Dani HSS 86.50. Volatiles were carried to the sensors by a stream of synthetic
air (ow rate 30 mL min)1).
Each analysed specimen produced, consequently, a
plot of 16 curves (being each one the output signal of
each sensor vs. time). It is necessary to reduce the large
quantity of data so as to describe each sample with a
minimal number of parameters. The Principal Component Analysis (PCA) algorithm makes use of the
advantage that sensors are relatively non-specic and
that it can combine the signals of all the sensors in a
unique signal. Employing this method, similar odours
tend to be grouped in clusters and the result is a
bidimensional plot (axes are the components which
contribute most to the odours expressed in %).
The binomial distribution was used to calculate the

signicant level for the triangle test, based on a number
of correct answers. Analysis of variance (anova) was
carried out to assess the attributes which are signicantly dierent among samples using the General Linear
Model command in spss v. 13.0. The variability of each
descriptor was studied using a model where assessor was
considered a random factor and sample and replication
xed factors. Multiple means comparisons were carried
out by StudentNeumanKeuls (SNK) test at P < 0.05.
PCA was conducted to examine the relationship among

attributes and samples, correlation matrix was used and
the minimum eigenvalue was set at 1. Clusters were
performed by K-Means command in spss v. 13.0. E-nose
used PCA as the statistical method to process the data
obtained by MOSES II.
Results
Triangle test
Table 3 shows the results for the triangle test for all
samples. As it can be seen all samples resulted signicantly dierent except for sample 4 respect to sample 5.
These two samples had sucrose in low concentrations (5
and 10% respectively) and the same amount of aroma
(18%).
Analysis of variance
anova of mixed model for attribute scores are summarized in Table 4. Sources of variation were samples,
assessors and sample assessor interaction. Replications and assessor replication and sample replication interactions were non-signicant supporting the
interpretations that the use of attributes was consistent.
Eect of assessors was signicant for only one
attribute (vitamin C, P < 0.05) and sample assessor
interaction was signicant for two attributes (freshly
squeezed, P < 0.01 and vitamin C, P < 0.05) indicating that not all the judges evaluated all the samples in
the same fashion for those descriptors. Other authors
observed that this can happen when samples are very
Table 3 Discrimination between samples: triangle test

Samples
compared
Correct
answers
Total
answers
Significance
level (%)
12
13
14
15
16
23
24
25
26
34
35
36
45
46
56
18
19
18
19
19
20
21
20
18
15
16
15
9
15
15
24
24
24
24
24
24
24
24
24
24
24
24
24
24
24
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
1.0
0.1
1.0
n.s.
1.0
1.0
n.s.: no significant difference. Samples 1 through 6 are described in

Table 1.
Table 4 F-values of samples for aroma attributes

Anova F-value
Attribute
degrees of freedom
Freshly squeezed
Powder juice
Citrus terpenes
Vitamin C
Tangerine
Candy
Solvent
Plastic
Peely
Marmalade
Woody
Pungency
Green
Syrup
Replicate
(R)
1
0.54
0.11
0.57
1.28
1.73
0.08
2.51
0.92
1.79
3.03
0.49
0.44
0.68
0.06
Sample
(S)
5
17.99***
21.37***
25.21***
6.77***
3.76**
10.26***
6.78***
9.11***
3.32**
2.69*
9.77***
4.87***
4.98***
10.00***
Assessor
(A)
11
0.51
1.34
0.95
2.99*
2.15
1.13
0.55
2.37
1.58
0.95
0.52
0.98
2.73
1.42
AS
55
1.96**
1.30
1.59
1.90*
1.28
0.81
0.98
0.89
1.28
1.44
0.98
1.36
0.83
1.22
*P < 0.05; ** P < 0.01; *** P < 0.001.
similar in their sensory properties so assessors cannot

dierentiate easily among them (Tang et al., 1999;
Zamora & Guirao, 2002). In order to verify this
observation, we examined the samples which were
dierent from each other for these attributes. anovas
were calculated for the samples whose means had the
biggest dierence and which were found signicantly
dierent by SNK test (Table 5). No signicant interactions were found for freshly squeezed (samples selected 1
and 4) F1,11 = 1.001 and vitamin C (samples selected 1
and 5) F1,11 = 1.969. For this last attribute no signicant dierence was found among assessors for the most
dierent samples (1 and 5) F1,11 = 0.886. This data
shows that for these two attributes there was an
agreement among assessors. This being so, interactions

(Table 4) could be attributed to the similarity among
samples rather than to dierent criteria used by judges.
The means of attributes which showed signicant
dierences among samples are presented in Table 5.
Sample 1 presented the higher mean scores for the
attributes: citrus terpenes (75.3), vitamin C (65.9) and
marmalade (46.4); sample 2 had the higher mean for
plastic (66.9); and sample 6 had the maximum mean
value for woody (61.4) and syrup (72.2). As regards
samples 3, 4 and 5, even though neither of them
presented a maximum nor a minimum value for any
attribute, all of them shared the maximum value for
freshly squeezed and candy while 4 and 5 also shared the
maximum for powder juice and pungency. These results
are in accordance with triangle tests (Table 3).
Aroma Profiling
PCA explains 93.5% of the variance among samples

with the rst three components; the biplot of Principal
Component 1 (PC1) vs. Principal Component 2 (PC2) is
shown in Fig. 1. The most important attributes which
compose PC1 are juice powder, freshly squeezed,
pungency and candy opposite to woody, syrup and
vitamin C; and PC2 is dened by plastic, solvent and
green.
Three dierent clusters were formed; samples 3, 4 and
5 (MD-sucrose (40%), MD-sucrose (5%)-lactose (21%),
MD-sucrose (10%) were grouped and described by the
attributes juice powder, freshly squeezed, candy, tangerine and pungency. Samples 6 and 1 (MD-trehalose
and MD) had mainly the aromas woody, marmalade,
syrup, citrus terpenes, and vitamin C. Finally, sample 2
(MD-gum arabic) was characterised by solvent, plastic,
peely and green. Samples 4 and 5 are grouped one more
Table 5 Mean sensory scores of the attributes quantied in each of the samples
Attribute
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Sample 6
Freshly squeezed
Powder juice
Citrus terpenes
Vitamin C
Tangerine
Candy
Solvent
Plastic
Peely
Marmalade
Woody
Pungency
Green
Syrup
27.6
19.2
75.3
65.9
31.8
32.6
31.0
29.2
44.6
46.4
43.4
26.8
31.2
33.9
30.4
21.9
35.3
46.8
33.3
28.7
58.5
66.9
57.4
28.5
39.3
25.9
53.5
44.4
64.6 b
42.5 b
23.0a
38.9 a
31.7 ab
50.8 b
37.5 b
34.2 ab
32.5 a
23.3 a
30.6 a
39.7 ab
39.9 abc
36.7 ab
68.9
63.8
33.2
31.7
39.1
55.6
58.6
48.4
42.7
32.7
26.9
51.4
42.4
28.6
64.3
68.4
27.5
31.7
45.5
56.5
39.5
40.4
28.0
21.7
30.2
56.1
49.2
34.1
31.4
30.0
28.2
38.0
20.8
21.0
22.0
23.8
41.5
31.4
61.4
39.2
27.0
72.2
a
a
b
b
ab
a
ab
ab
ab
b
b
a
ab
ab
a
a
a
a
b
a
c
d
b
a
ab
a
c
b
b
c
a
a
bc
b
c
c
ab
a
a
bc
bc
a
b
c
a
a
c
b
b
bc
a
a
ab
c
c
ab
a
a
a
a
a
a
a
a
ab
a
c
ab
a
c
Different letters after medias in every row indicate samples which differed for that attribute, P < 0.05, SNK test.
1573
Plastic
Solvent
0.8
Green
Component 2 (21.84%)
0.6
Peely
2
2
0.4
Vitamin C
0.2
Tangerine
4 4
Citrus terpenes
11
0.2
Candy
55
Freshly squezeed
Woody Marmalade
6
6 Syrup
0.4
Juice powder
Pungency
0.6
0.8
1
1.2
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
1.2
Figure 1 Principal Component Analysis of

sensory data.
Component 1 (56.55%)
time as shown in Table 5 and in the triangle test

(Table 3).
samples contained sucrose in low concentrations (5 and

10%, respectively) and the same amount of encapsulated aroma (18%) in their composition; in addition
sample 4 contained lactose in its matrix (21%) but,
apparently, lactose would not be contributing to the
dierential volatile retention during spray-drying process. The e-nose could not detect any dierences among
samples 1, 2, 3 and 6. On the other hand, according to
the sensory analysis, samples 4 and 5 were grouped
together with sample 3 (Fig.1). These three samples were
described by the same attributes but diered in the mean
intensity of the descriptors mentioned, being sample 3
the one with the lowest intensity for powder juice,
pungency, tangerine, solvent and plastic. Sample 3 also
had sucrose in its composition, but in a higher concentration (40%) and had a smaller aroma fraction (13%).
E-nose
PCA performed with data obtained by the E-nose

explains 99.2% of the variance among samples with the
rst component (Fig.2). Only two clusters can be
observed: samples 4 and 5 are grouped together while
the rest of the samples (1, 2, 3 and 6) cannot be
discriminated, forming a whole dierent group.
Discussion
Samples 4 and 5 are perceived as similar using either

sensory analysis or e-nose evaluation. Both powder
1
0.8
0.6
0.4
3
0.2
PC2 0.4%
1574
5
1
3
3
0.2
6
2 2 6
0.4
4
5
4
4
0.6
0.8
1
5
1
PC1 99.2%
Figure 2 Principal Component Analysis of

e-nose data.
The human nose was able to discriminate samples 1

and 6 from the rest (Fig. 1), describing them by the same
attributes; however, the e-nose did not nd signicant
dierences. These two samples (1 and 6) diered in
matrix composition as well as in aroma proportion
(sample 1: 72% MD, no disaccharide, 16% aroma;
sample 6: 40% MD, 40% trehalose, 13% aroma).
Finally, sample 2 was found dierent from all the rest
(Fig. 1); this sample contained the most aroma (20%)
and gum arabic instead of a disaccharide.
A dierence between the human and e-nose is the
pungency sensation measured only by sensory analysis.
This sensation is mediated not by smell bres, but by
other chemosensitive bres as trigeminal nerve branches. Some compounds which are perceived as being
purely olfactory (e.g. butyl acetate, a fruity odour) can
elicit activity in the trigeminal nerve without creating
sensations of burning or stinging (Cain, 1974). Electrophysiological and psychophysical evidence indicate that
odours at concentrations lower than those generally
considered to be non-irritating can stimulate both
olfactory and trigeminal chemo receptors, and this
stimulation can contribute to the perceived odour
intensity (Maruniak, 1988; Delwiche, 2004). As pungency is only detected by human nose this perception
could contribute to dierentiate both measurements
(human and e-nose).
As regards substitution of sucrose by trehalose in the
matrix composition, it was noted that samples 3 and 6
(MD-sucrose, MD-trehalose) had dierent aroma proles suggesting that dierent volatiles are retained during
spray-drying according to the sugar used in the encapsulating matrix (sucrose or trehalose). Komes et al.
(2005) studied the inuence of the addition of trehalose
or sucrose on the retention of volatiles in dehydrated
apricot puree (foam-mat drying and freeze-drying) and
found that the best retention of aroma compounds was
obtained when trehalose was added. It is to be noted,
however, that in present work we used spray-drying and
also the volatile composition of orange avour is likely to
be quite dierent from that of apricot avour.
Conclusions
Sensory analysis proved to be a more sensitive tool than

the e-nose for grouping and describing the orange
encapsulated avours. Although both methods produced coherent results, by sensory analysis samples
were discriminated more sharply. In addition, sensory
analysis has the advantage that it can name the
dierence, this cannot be performed with the e-nose.
As expected, the composition of the matrix likely
inuenced the type and the amount of volatiles retained,
as suggested by the PCA analysis. In this way, by
changing the matrix composition used for aroma
encapsulation dierent aromatic proles can be
obtained, which will give as a result dierent products.

For example, if attributes such as freshly squeezed,
pungency, tangerine, candy and juice powder are desired
in the nal aromatic prole of an orange avoured
product, then the matrix used for encapsulation should
include sucrose. On the other hand, if woody or
marmalade notes are desired in the nal product,
trehalose should be used in the encapsulating matrix.
Citrus notes are obtained by using a MD and modied
starch matrix while a matrix with MD and gum arabic
would be recommended for a product in which peely
and green attributes are desired.
Acknowledgments
This work was funded by FonCyT (SeCyT) of Argentina (Project PICT No 931951). The authors would like
to thank the sensory panel for their collaboration in this
work and Cargill incorporated (Wayzata, MN, USA)
for donating the trehalose used in this study.
References
ASTM 1977. Manual on Sensory Testing and Methods, STP 434, pp.
3940. Philadelphia, PA: American Society for Testing and Materials.
Baby, R., Cabezas, M., Calvino, A., Tomasi, O. & Walsoe de Reca,
N.E. 1999. Quantitative evaluation of aromatic additives employing
an electronic nose with sensors and a sensorial panel. In: 6th
International Symposium Olfaction and Electronic Nose (ISOEN99)
(edited by M. Frank & U. Weimar). Pp. 305308, Tubingen,
Germany. Tubingen: Institute of Physical Chemistry, University of
Tubingen.
Bhandari, B.R. & Hartel, R.W. 2005. Phase transitions during food
powder production and powder stability. In: Encapsulated and Food
Powder (edited by C. Onwulata & R.P. Konstance). Pp. 261292.
New York, NY: Marcel Dekker.
Beristain, C.I., Azuara, E. & Vernon-Carter, E.J. (2002). Eect of
water activity on the stability to oxidation of spray dried
encapsulated orange peel oil using mesquite gum as wall material.
Busso Casati, C., Schebor, C., Zamora, M.C. & Chirife, J. (2007).
Glass transition temperatures and some physical and sensory
changes in stored spray-dried encapsulated avours. LWT - Food
Science and Technology, (in press).
Cain, W.S (1974). Contribution of the trigeminal nerve to perceived
odour magnitude. Annals of the New York Academy of Sciences, 237,
2834.
Delwiche, J. (2004). The impact of perceptual interactions on perceived
avour. Food Quality and Preference, 15, 137146.
Komes, D., Lovric, T., Kovacevic Ganic, K. & Gracin, L. (2003).
Study of trehalose addition on aroma retention in dehydrated
strawberry puree. Food Technology. Biotechnology, 41, 2.
Komes, D., Lovric, T., Kovacevic Ganic, K., Gajdos Kljusturic, J. &
Banovic, M. (2005). Trehalose improves avour retention in
dehydrated apricot puree. International Journal of Food Science
Lawless, H.T. & Heyman, H. (1998). Sensory Evaluation of Food. Pp.
5860. New York, NY: Chapman & Hall.
Maruniak, J. A. (1988). The sense of smell. In: Sensory Analysis of
Foods (edited by J.R. Piggot). Pp. 2568. New York, NY: Elsevier.
Meilgaard, M., Civille, G. & Carr, B. (1991). Sensory Evaluation
Techniques. 2nd edn. Pp. 88. Boca Raton, FL: CRC Press.
1575
1576
Menting, L.C., Hoogstad, B. & Thiejssen, H.A.C. (1970). Aroma

retention during the drying of liquid foods. Journal of Food
Mitrovics, J., Hulmer, H., Noetzel, G., Weimar, U. & Gopel, W.
(1997). Design of a Hybrid Modular Sensor System for Gas and
Odour Analysis. In: Proceedings Transducers (edited by IEEE
Publications) Pp. 14. Chicago, IL: IEEE Publications.
Roos, Y. (1995). Phase Transitions in Food. New York, NY: Academic
Press.
Roos, Y. & Karel, M. (1991). Phase transition of mixtures of
amorphous polysaccharides and sugars. Biotechnology Progress, 7,
4953.
Stone, H. & Sidel, J.L. (1993). Sensory Evaluation Practices. 2nd edn.
San Diego, CA: Elsevier Academic Press.
Tang, C., Hsieh, F., Heymann, H. & Hu, H.E. (1999). Analyzing and
correlating instrumental and sensory data: a multivariate study of
physical properties of cooked wheat noodles. Journal of Food
Quality, 22, 193211.
Thijseen, H.A.C. (1971). lavour retention in drying preconcentrated
food liquids. J. Appl. Chem. Biotechnol., 21, 372377.
Wright, J. (1995). Essential Oils. In: Food Flavourings Second Edition
(edited by P.A. Ashurst). Pp. 54, 55310. Glasgow: Blackie Academic
and Professional.
Zamora, M.C. & Guirao, M. (2002). nalyzing the contribution of
orally perceived attributes to the avour of wine. Food Quality and
Preference, 13, 5.
Original article
Probiotic foods: consumer perception and attitudes
Julia V. Viana, Adriano G. da Cruz,* Sidney S. Zoellner, Ramon Silva & Aline L. D. Batista
Curso de Farmacia, Universidade Estacio de Sa, campus Reboucas, RJ, Rua do Bispo, 83, Rio Comprido, Rio de Janeiro, Brazil
(Received 17 August 2006; Accepted in revised form 10 April 2007)
Summary
The objective of this study was to evaluate the perception and the attitudes towards probiotic foods of the
population in the city of Rio de Janeiro, Brazil. Four hundred and twenty (100.0%) people were interviewed
in small-, medium- and large-sized supermarkets located in various parts of the city of Rio de Janeiro. One
hundred and twenty-two (29.05%) people dened probiotic foods correctly. Ninety-one (21.67%) were
unable to mention a single example of a probiotic food. The results of this study indicate the need for an
elementary easy-to-understand educational programme using accessible language in order to x the concepts
related to these products.
Keywords
Consumer, probiotic.
Introduction
The term probiotic refers to live microbial cultures,

which, when administered to men or animals (in the
form of dehydrated cells or fermented products), positively aect the hosts state of health by improving the
properties of the original micro-ora (Margoles &
Garcia, 2003). The prophylactic and therapeutic eects
of these micro-organisms have been reported in various
studies in the following terms: (a) balancing the
intestinal ora; (b) increasing lactose tolerance and
ingestion; (c) reducing cholesterol levels; (d) synthesis of
B complex vitamins; (d) absorption of calcium; (e)
modulating the immunological system (Wang &
Ascheri, 1991; Hyun & Shin, 1998; Gomes & Malcata,
1999; Margoles & Garcia, 2003). Although ocial
statistical data are not yet available, the market for
foods containing probiotic cultures is on the rise in
Brazil, dairy products (bio-yoghurts, fermented milks,
cheeses, lactic beverages) being currently the main
representatives.
Consumer surveys are extremely important to the
food industry, as they allow for the identication of the
level of knowledge about a determined subject and also
for the tracing of strategies to correct and or identify
failings, in order to increase the sales of a food product,
consequently increasing company prots and the level of
consumer awareness. Studies of this type involving
probiotic foods are rare in Brazil, and generally show a
complete lack of knowledge of the subject. Castro et al.
(2002) carried out a survey of consumer habits involving
*Correspondent: E-mail: food@globo.com
doi:10.1111/j.1365-2621.2007.01596.x
dairy products in the city of Vicosa, MG, Brazil, and

reported a great lack of knowledge about probiotics and
their benets by the majority of that population.
Rio de Janeiro is an important Brazilian city, being
the second largest economic centre, losing only to Sao
Paulo, and presenting a potential consumer market for
probiotic foods. Thus, evaluating the knowledge of this
population would be extremely useful for a better
understanding of this type of product, contributing to
a sustained, conscious growth of the consumer market
in other parts of the country. Hence, in this context, the
objective of this study was to evaluate knowledge about
probiotic foods by the population of the city of Rio de
Janeiro, Brazil.
Sampling
A structured questionnaire, consisting of three questions, was used in this study. Each question had three
options for the interviewee to choose his reply, and a
free line for him to make his own reply if he failed to
agree with any of the options provided. The interviews
were carried out in small-, medium- and large-scale
supermarkets located in various regions of the city of
Rio de Janeiro, focussing on the area destined for the
sale of refrigerated dairy products the interviewer
placing himself in this precinct. The direct approach
method was chosen, in which, the interviewee was
invited to participate in the study, the size of the sample
being determined according to Triola (2005). Considering a level of precision of D = 0.05, a condence
1577
1578
Probiotic foods J. V. Viana et al.
interval of 95% and a proportion of 50%, the nal

sample size calculated was n = 385 interviewees. Thus,
420 people were interviewed by applying the structured
questionnaire.
The results were tabulated and graphs were constructed

using the Microsoft Excel 7.0, version 2000 computer
programme. The chi-square analysis was used to
investigate the inuence of the educational level of
population sample about probiotic food knowledge.
The population sample used in this study consisted of

420 (100.00%) people aged between 17 and 80 years
166 (39.52%) being male and 254 (60.47%) female. With
respect to the educational level, 125 (29.76%) had only
completed their basic education, 163 (38.80%) had
completed their high school education and 132 (31.43%)
their university education. This proportionality could be
explained by the greater frequency of women shopping
for food products in supermarkets and by the higher
educational degree of the female sex in relation to the
male sex, as shown by various surveys carried out
previously. In an overall basis, the education level of the
population sample did not inuence the knowledge
about probiotic food (P > 0.05).
Table 1 shows the level of consumer knowledge with
respect to the denition of food probiotics. 143
(34.05%) dened probiotic foods as those that contributed to a reduction in weight; 120 (28.57%) failed to
express any concept; and 44 (10.48%) associated probiotic foods with cures for psychological problems. Only
122 (29.05%) dened probiotic foods correctly, that is,
as foods contributing to the equilibrium of the intestinal
ora. These data are surprising, as they show complete
disinformation by the consumers with respect to this
type of food and its benets, and suggest that the high
level of advertising carried out by food industries has
not managed to impart this information to the population at large, additional alternatives being required. The
participants in this survey showed considerable confusion between probiotic foods and diet and light foods,
foods already consolidated on the market, and with
anti-depressive foods, suggesting the need for an educaTable 1 Denitions of probiotic foods
Answers
Percentage
Food that contributed to a reduction in weight

Food related to the equilibrium of the intestinal flora
Failed to express any concept
Products related to cures for psychological problems
34.05
29.05
28.57
10.48
Table 2 Benets arising from the ingestion of probiotic foods

Answers
Percentage
Reductions in cholesterol levels

Diarrhoea
Headache
Caries
29.52
28.33
12.86
7.86
21.43
tional programme at an elementary level using accessible, easy-to-understand language, aimed at the
denitive xation of the concepts related to these
products. Although the participants showed a reasonably high educational level 70% with high-school or
university-level education the replies were observed to
be independent of this requisite.
Table 2 shows the benets arising from the ingestion
of probiotic foods according to the interviewees. The
main items cited were reductions in cholesterol levels
and diarrhoea cited by 124 (29.52%) and 119 (28.33%)
participants, respectively, followed by reductions in
caries and headaches, with indices of thirty-three
(7.86%) and fty-four (12.86%), respectively. Ninety
(21.43%) participants armed that they knew of no
benet caused by the ingestion of these products.
Although a reduction in diarrhoea is the benet most
emphasised by the manufacturers to encourage the
consumption of probiotic foods, a benet of great
interest to the majority of the population undergoing
considerable discomfort provoked by this condition is
the reduction in the cholesterol levels, which was slightly
more cited. This could be explained by the relatively
high educational level of the participants, as many of
those with a university education also possessed postgraduate latu sensu (specialisation) or strict-sensu (masters or doctorate) degrees, thus being used to reading
scientic texts and also owing to the great number of
articles and texts divulged in the media of the harm
caused by maintaining high cholesterol levels. The
maintenance of a healthy life style, with the practice of
regular physical exercise allied with the consumption of
adequate food, including the consumption of probiotic
foods, could be related to this fact.
A similar explanation could be attributed to the
citation of a reduction in caries, a little divulged benet
restricted to scientic research, which could be related to
the access of the participants to dental services oered
by dentists who keep themselves professionally updated,
and who, therefore, have access to this type of
information and transmit it to their patients.
Although a certain advance was shown by the
interviewees in their responses, ninety (21.42%) failed
to relate the consumption of probiotic foods to any
of the benets mentioned in question 2, registering
replies, such as diverse weaknesses, menopause and skin
Table 3 Examples of probiotic foods

Answers
Percentage
Fermented milk
Yogurt
Cheese
Other foods products (coconut water, soft drinks)
27.62
25.71
13.66
11.66
21.67
irritations, and fty-four (12.86%) associated the consumption of these foods with relief from headaches,
reiterating the need for educational elucidatory
campaigns.
Table 3 shows examples of foods considered to be
probiotic by the interviewees. One hundred and sixteen
(27.62%) interviewees indicated fermented milk and 108
(25.71%) yoghurt, while fty-six (13.33%) and fortynine (11.66%) mentioned cheese and other products
including coconut water, soft drinks, fruits and vegetables respectively. Ninety-one (21.67%) were unable to
mention a single example of a probiotic food. A low
index of correct replies was veried (27.62%), although
some cited acidophilus milk, and a certain amount of
confusion existed among a section of the interviewees
citations of yoghurt and cheese occurring at a rate of
25.71 and 13.33%, respectively. Widely consumed over
the ages, yoghurt is only considered a probiotic product
if it includes cultures of the type Lactobacillus acidophilus and Bidobacterium bidum in its composition,
added as co-cultures during processing, because the
cultures normally used to produce this product, Lactobacillus delbrueckii ssp bulgaris and S. thermophilus,
possess no probiotic characteristics. However, a high
index of replies was already expected for this product, as
it is well known and consumed by a wide range of the
population, independent of socio-economic class.
Although showing considerable potential as vehicles
for probiotic cultures, cheeses also can only carry this
status if they have the aforementioned cultures in their
composition, also added during their preparation. The
citation of products, such as coconut water, fruits and
vegetables, together with those who could think of no
examples of probiotic foods, represent 33.3%
(11.66 + 21.67), reinforcing the aforementioned recommendation of the need to divulge these foods to the
population. It is interesting to note that no participant
cited ker as a probiotic food.
The results of this study show, in a general way, the
confused opinions of the population with respect to
probiotic foods and the benets arising from their
ingestion. Wider divulging of this category of foods by
the food industry is required by way of educative
campaigns in accessible, easy-to-understand language.
This would further increase the potential for growth of
this type of food, as aware consumers present a greater
degree of loyalty and could make regular use of these

products. In addition, surveys involving people who
identied improvements in their health reduction of
diarrhoea, reduction of cholesterol level, among others
after the ingestion of probiotic products, should be
carried out so as to provide data on the perception and
consumption of these products and the type of action of
these products with respect to the population.
These results are similar to those of other studies
carried out throughout the world, suggesting that this
problem is not related to the socio-economic level of the
country.
Babajimopoulos et al. (2004) reported a lack of
knowledge about probiotic foods by 71 and 82% of
men and women, respectively, resident in Grevena,
Greece, although 63.5 and 57% armed that they
consumed these foods, live yoghurt or bio-yoghurt and
ker being the most frequently cited examples. On the
contrary of the present study, the authors reported that
the level of knowledge about probiotic foods was related
to the educational level.
Luckow & Delahunty (2004) identied a preference
for functional orange-juice preparations containing
probiotic cultures, in detriment to orange juices
conventionally produced in Ireland, in a study involving
25% of the consumers. Only 11% reported this food as
their rst option these participants being mostly older.
Gorska-Warsewicz (2003) carried out a survey involving functional foods with consumers in various regions
of Poland, and quantied the results using an acceptability index varying from 0 (total rejection) to 1 (total
acceptance). Probiotics received the lowest acceptability
indices (0.22), the best scores being registered for foods
with improved consistency (0.86), owing to the addition
of special ingredients and enriched foods (0.64).
Bruhn et al. (2002) carried out a study involving
consumers in California, USA, with respect to probiotic
foods. Various responses were obtained with respect to
knowledge about probiotic micro-organisms, such as
Lactobacillus and Bidobacterium, and many consumers
did not feel the need to ingest this type of food to
strengthen their immunological systems. The results also
revealed that foods containing probiotic cultures should
not be sold at a higher price than conventional foods
and should provide information on their labels related
to health aspects and possible eects on diseases, such as
May reduce cholesterol and Prevents diarrhoea.
Connon et al. (2004) carried out studies aimed at
obtaining comparative opinions about various types of
functional foods in two diverse population groups: the
public in general and nutritionists, using a structured
5-point scale. Both groups demonstrated relative knowledge of probiotic foods 2.08 and 2.85, respectively.
The general consumers were shown to possess slightly
better knowledge about the benets of ingesting these
foods than the nutritionists 3.11 and 2.91.
1579
1580
Conclusions
The results of this study represent an attempt to provide

data about consumer opinion of probiotic foods. In an
overall way, confused opinions of the population with
respect to probiotic foods and the benets arising from
their ingestion were reported. Wider divulging of this
category of foods by the food industry is required by
way of educative campaigns in accessible, easy-tounderstand language. Similar studies should be carried
out in various parts of Brazil in order to obtain a true
dimension of the problem and facilitate the planning of
strategies aimed at reverting the situation and promoting the conscious use of this type of food. In addition,
surveys involving people who identied improvements
in their health reduction of diarrhoea, reduction of
cholesterol level, among others after the ingestion of
probiotic products, should be carried out so as to
provide data on the perception and consumption of
these products and the type of action of these products
with respect to the population.
References
Babajimopoulos, M., Fotiadou, E., Alexandridou, E. & Nikolaidou,
A.I. (2004). Consumers knowledge on probiotics and consumption
of these products in the city of Thessaloniki, Greece. In: Proceedings
of the 9th Karlsruhe Nutrition Congress (edited by BFEL, Federal
Research Centre for Nutrition and Food). Pp. 24. Karlsruhe, 1012
October 2004. Karlsruhe: BFEL.
Bruhn, J.C., Cotter, A., Garrett, C., Klenk, M. & Powtll, C. (2002).
Consumer attitudes toward use of probiotic cultures. Journal of
Castro, P.R.S. de, Gugliotta, L.M., Ramos, M. da P. P. & Ribeiro
Junior, J.I. (2002). Habitos de Consumo de Leite e Derivados na
cidade de Vicosa, MG. Revista do Instituto de Laticnios Candido
Tostes, 57, 174177.
Connon, A.M.C., Fletcher, P.L., Cade, J.E., Greenwood, D.C. &
Pearman, A.D. (2004). Dierences in perceptions of functional
foods: UK public vs. nutritionists. Nutrition Bulletin, 29, 118.
Gomes, A.P & Malcata, F.X. (1999). Bidobacterium spp and
Lactobacilus acidophilus: biological, biochemical, technological and
therapeutical properties relevant for use as probiotics. Trends in
Food Science & Technology, 10, 139157.
Gorska-Warsewicz, H. (2003). Novelty products in consumers opinion. Przemysl-Spozywczy, 57, 3436.
Hyun, C. & Shin, H. (1998). Utilization of bovine plasma obtained
from a slaughterhouse for economic production of probiotics.
Journal of Fermentation and Bioengineering, 86, 3437.
Luckow, T. & Delahunty, C. (2004). Consumer acceptance of orange
juice containing functional ingredients. Food Research International,
37, 805814.
Margoles, A. & Garcia, L. (2003). Characterization of a Bidobacterium strain wish acquired resistance to cholate: a preliminary study.
Triola, M.F. (2005). Introducao a Estatstica. 435 pp. Sao Paulo: LTC.
Wang, S.H & Ascheri, J.L.R. (1991). Iogurte de soja. Fermentacao
lactica e avaliacao sensorial. Ciencia e Tecnologia dos Alimentos, 11,
221239.
Original article
Comparison of full-fat and low-fat cheese analogues with or
without pectin gel through microstructure, texture, rheology,
thermal and sensory analysis
He Liu,1,2 Xue Ming Xu1,2* & Shi Dong Guo2
1 The Key Laboratory of Food Science and Safety, Ministry of Education
2 School of Food Science and Technology, Southern Yangtze University, Jiangsu, Wuxi, China
(Received 17 November 2006; Accepted in revised form 29 March 2007)
Summary
The eects of pectin gel and protein base on processed semi-solid cheese analogues were studied through
microstructure, texture, rheology, thermal analysis and sensory evaluation. Scanning electron microscopy
revealed dierences in the microstructure of processed cheese analogues. Samples made with full-fat
contained higher concentrations of fat globules and were denser compared with low-fat cheese analogues
with or without pectin gel. The pectin gel in the products acted as a linkage with other ingredients and made
the products more compact and had less cavity compared with the products without pectin gel added. On
rheological analysis, the full-fat products manifested a more solid-like form. The storage modulus of pectin
gel sample was higher than that without pectin gel. All the samples rheological parameters were depending
on the oscillatory frequency and temperature. In low-fat samples, pectin gel added or not aected the
hardness, gumminess, chewiness and adhesiveness signicantly. The pectin gel addition show positive eect
to the texture prole of the low-fat cheese analogues. Through thermal analysis, the meltability and glass
transition temperature of the processed cheese analogues were measured. The low-fat cheese analogue with
pectin gel addition got the higher texture and mouthfeel scores through sensory evaluation.
Keywords
Cheese analogue, dierential scanning calorimetry, low fat, microstructure, pectin gel, rheology, texture.
Introduction
Cheese analogues or imitation cheeses contain edible oil

or fat emulsied in an aqueous protein phase. Cheese
analogues are usually dened as products made by
blending individual constituents, including the nondairy fats or proteins, to produce a cheese-like product
that can meet specic requirements. They are being used
increasingly because of their cost-eectiveness, attributed to the simplicity in their manufacture and the
replacement of selected milk ingredients by cheaper
vegetable products (Bachmann, 2001).
Over the past decade, the consumption of low-fat
food products has become more than just a trend. In
view of the general consensus that the amount and type
of fat consumed is of importance to the aetiology of
several chronic diseases (e.g. obesity, cardiovascular
diseases, cancer), it is not surprising that consumers
*Correspondent: Fax: +86-510-85879711;
e-mail: xmxu@sytu.edu.cn
more readily adhere to nutritional guidelines concerning

fat consumption. Largely inuenced by health-related
concerns, there has been pressure on the food industry
to reduce the amount of fat, sugar, cholesterol, salt and
certain additives in the diet. Food manufacturers have
responded to consumer demand and there has been
rapid market growth of products with a healthy image.
Low-fat dairy products, such as milk, yoghurt, ice cream
and some cheese products have been available for
several years. In cheese production, the removal or
reduction of fat adversely aects both the avour and
texture (Ehab et al., 2002). Therefore, several strategies
have been proposed to improve the avour and texture
of low-fat cheeses. These strategies can be collected in
three titles (Drake & Swanson, 1995; Mistry, 2001): that
is making-process modications; starter culture selection
and use of adjunct cultures; and use of fat replacers. Fat
replacers are ingredients that are intended to replace
natural fats with the main objective of obtaining a
reduction in the caloric value. They are categorised as
fat substitutes which are fat-based and as fat mimetics
doi:10.1111/j.1365-2621.2007.01616.x
1581
1582
Comparison of full-fat and low-fat cheese analogues H. Liu et al.
which are protein- and carbohydrate-based. Fat mimetics have often been recommended to be used in cheese
products consisting mainly of microparticulated protein- and carbohydrate-based materials (Romeih et al.,
2002).
As the introduction of Siebel & Sylvia (1996), pectin is
a puried carbohydrate product, obtained by aqueous
extraction under mildly acidic conditions of some plant
material usually citrus fruits and apples. Traditionally,
pectin is used as a gelling agent for jams and jellies. The
major parts of all pectin production are consumed by
the fruit processing industry. Other traditional applications are confectionery products, dairy products, fruit
preparations, bakery llings.
New applications of pectin within the food area are
constantly developing, and fat replacement is one of the
latest newcomers. SLENDID, a registered trademark
of Hercules Incorporated, was introduced in 1991
(Siebel & Sylvia, 1996). The SLENDID concept covers
a range of specialty pectins tailor-made for fat replacement. The production of SLENDID takes place on the
premises of a company in Denmark. In 1994, the
company was granted a patent covering a fat-simulating
composition consisting of heat-stable carbohydrate gel
particles, a food product normally containing fat oil
that has been improved by substituting all or a portion
of the fat oil by gel particles, and the process by which
the gel particles are formed. SLENDID may be used in
a wide range of food applications such as spreads,
mayonnaises and salad dressings, processed meats, ice
cream, processed cheeses, soups and sauces, desserts and
bakery products, in which fat may be partly or fully
replaced.
Use of scanning electron microscopy (SEM) techniques to cheeses and gels and evaluation of the product
were successful in showing the microstructure (Sipahioglu et al., 1999; Sanche et al., 2000). Texture properties
of Cheddar cheese samples were determined using
compression and stress relaxation tests carried out on
an Instron Universal Testing machine (Hort and Grys,
2001). It is convenient to employ instrumental texture
analysis in the current accepted form using uniaxial
compression. Literature introduced the texture prole
analysis (TPA) test on cheeses and discussed the
properties of the texture of the cheese samples (Ehab
et al., 2002; Truong et al., 2002; Kealy, 2006). Rheology
is mainly concerned with the relationship between
strain, stress and time. When subjected to external
forces, solids (or truly elastic materials) will deform,
whereas liquids (or truly viscous materials) will ow.
However, contemporary rheology is more interested in
the behaviour of real materials with properties intermediate between those of ideal solids and ideal liquids
(Doraiswamy, 2002). These industrially important materials are called viscoelastic materials, which include
almost all real materials. Without question, cheeses are
viscoelastic materials. The rheology analysis of cheese

samples had been carried out in many literatures
(Paraskevopoulou et al., 2003; Romdhan & Eric,
2003; Govindasamy-Lucey et al., 2005). Thermal analysis by dierential scanning calorimetry (DSC) on
cheese is scarce, but the DSC tests on gel were
introduced in the literatures (Deszczynski et al., 2003;
Normand, et al., 2003; Lazaridou et al., 2004; Ross
et al., 2006).
The objective of this study was to determine the eects
of pectin gel on the physical, composition of low-fat
cheese analogues and also to nd the correlation between
these properties. The rheology properties and thermal
properties of the samples were also studied. The eects of
the fat reduction on these properties were also determined. At the same time, the possibility of pectin gel as a
fat mimetic addition to cheese analogue was examined.
Materials
The casein and sodium caseinate were supplied by

Linxia Huaxia Dairy Products Co. Ltd, China. The
citrus low-methoxylated pectin gel was prepared by
mixing the pectin with water and interacted with calcium
ion to form a weak-gel Pectin was from Jaingxisheng
Shangrao Fuda Pectin Co. Ltd, China. Other materials
used for manufacture of the processed cheese anologues
were (a) emulsifying salt and sodium chloride prepared
in laboratory and related material were from China
Medicine (Group) Shanghai Chemical Reagent Corporation (b) nisin from Tianyu Group., China. (c)
cheddar Paste 565-1 and (d) butter avour from
Chr.Hansen., Denmark. (e) colour from Wuhan Stars
Modern Bio-engineering Co. Ltd, China.
Production of protein bases
The production of protein base and the processed cheese

was as introduced by Muir et al. (1999). The emulsifying
salts were dissolved in suitable quantity of water and
poured into the glass beaker which was placed in a water
bath. The temperature was raised to 5060 C, then a
measured quantity casein or sodium caseinate was
added to make the protein base using rst a low-speed
mixer, then a high-speed mixer until the lubricous cream
was formed. After overnight storage for 1416 h at
4 C, the protein bases were used in the production of
processed cheese analogues.
Processed cheese analogue production
After overnight storage, the protein bases had formed

gels. These gels were placed in a processing kettle (A.
Stephan. U. Sohne GmbH, Germany); 2 kg capacity
and blended the other ingredients according to the

recipes shown in Table 1. The ingredients were rst
mixed for 1 min at low speed, following application of
vacuum, the mixture was then heated to 70 C with
direct steam injection. The vacuum was then switched
o and heating continued to 90 C followed by mixing
at high speed for 2 min. The hot melted cheese was
packaged in rigid plastic cups and heat sealed with
aluminium foil. All samples were cooled and stored for
2 months at 4 C. The complete experiment was replicated twice more on separate occasions.
Chemical analysis
The amounts of moisture, and ash in the cheese samples

were measured by AOAC Ocial Method 926.08 (1995)
and AOAC Ocial Method 935.42 (1995), respectively.
Total protein and total fat content of the cheeses were
determined using the Kjeldahl, and the modied Mojonnier method, respectively (Marshall, 1992). The
protein content of cheeses was calculated by multiplying
the total nitrogen content by 6.38.
Microscopic analysis
Scanning electron microscopy is a valuable technique in

dairy research because it provides information on
microstructure of dairy products which can be related
to physical properties. Small cubes of the cheese
analogues were xed with 2.5% (v v) glutaraldehyde
in water for 1 h and rinsed three times with phosphate
buer. After that, the samples were then put in 0.2%
(w v) OsO4 left overnight, rinsed three times with
phosphate buer and dehydrated in a graded ethanol
series [(507090100)% (v v); 20 min per step] and
placed in 100% (v v) ethanol for an overnight. The
samples were critical point dried through CO2. They
were then fractured and coated with Au by diode sputter
coating. Micrographs were made with a QUANTA-200
(FEI) at an acceleration voltage of 10.0 kV.
Table 1 Experimental recipes (%) for the production of processed
cheese analogues
Ingredient
Ff
Lf
Lfc
Casein Sodium caseinate

Butter
Pecin gel
Emulsifying salt
Cheddar cheese flavour
Butter flavour
Nisin
Sodium chloride
Water
15
20
0
2
1
1
0.01
1.5
60
15
10
10
2
1
1
0.01
1.5
60
15
10
0
2
1
1
0.01
1.5
70
Ff, full-fat cheese analogue; Lf, low-fat with fat mimetics cheese
analogue; Lfc, low-fat control cheese analogue.
Textural analysis
Texture prole analysis parameters were determined by

using a texture analyser TA-XT2i (Stable Micro System,
Ltd, UK). A at plate probe (P 0.5-Delrin cylinder
probe) with 0.5 inches of diameter was attached to
moving crosshead. Samples were not moved from the
cup and it was ensured that the height of the samples
were identical by cut at least 1 cm away from cheese
analogues surface. They were left at 25 C for about
30 min until they reached the denite temperature. The
central temperature of a control specimen was measured
by a thermocouple. The operating conditions were:
selecting TPA as test mode and option. Pretest speed
was 2.0 mm s)1, test speed was 1.0 mm s)1. Postspeed
was 5.0 mm s)1. Two bite time interval was 5.00 s. Trig
type was auto. Trig force was 20 g. Acquisition rate was
200.00 pps, 20% of compression ratio from the initial
height of the sample in two bites. The texture prole
parameters were determined using the TPA curve, an
example, given in Fig. 1: the compressive force(g)
recorded at maximum compression during the rst bite
as a measure of cheese hardness (Katsiari et al., 2002);
the distance of the detected height of the product on the
second compression divided by the original compression
distance (Length 2 Length 1) as a measure of springiness; The negative force area (A3, cm2) during the rst
bite as a measure of adhesiveness (Antoniou et al., 2000);
The ratio of positive area during the second compression
to the positive area during the rst compression (A2 A1)
as a measure of cheese cohesiveness; the product of
hardness cohesiveness as a measure of gumminess; the
product of gumminess springiness as a measure of
chewiness (Katsiari et al., 2002). Texture values were the
mean of three replicates tested each sampling time.
Rheological analysis
Samples were put on the bottom plate of the rheometer

(TA Instruments AR-1000, UK) which was equipped
with a 40-mm, plateplate measuring system and a
1000 lm spacing. To prevent evaporation and protect
against dehydration during test of the samples, lowviscosity silicone oil was applied to the exposed surfaces
of the samples. Preliminary experiments were carried
out to determine the linear viscoelastic regions at which
the frequency sweep of the samples was obtained.
Frequency sweep and temperature sweep were performed and measured values obtained included G
(elastic modulus), G (loss modulus) and tan d (loss
tangent = G G). Cheese analogue samples were subjected to heating and cooling proles: A strain of 5%
was applied at a frequency of 1 Hz to the samples, the
heating and cooling rate used was 2 C min)1 at 20
60 C and the tan d values were obtained (Gunasekaran
and Mehmet Ak, 2003).
1583
1584
Figure 1 The typical TPA curve of full-fat

cheese analogue (springiness = Length
2 Length 1; adhesivness = A3;
cohesiveness = A2 A1; gumminess = hardness cohesiveness; chewiness = gumminess springiness).
Thermal analysis
Dierential scanning calorimetry analysis was performed with a DSC-7 calorimeter (PerkinElmer, Norwalk, CT, E. U. A.) An empty pan was used as a
reference. The samples (about 10 mg), previously weighted in aluminum pans, were analysed according to the
following two independent program: (a) Heating from
2580 C at a rate of 5 C min and measure the
meltabilities of samples. Temperature for the dierent
transitions (i.e. the onset temperature, To; ending
temperature, Te) were determined using the rst derivative of the heat capacity calculated with the DSC
program library and by comparison with the baseline.
On the contrary, enthalpy (Delta H, J g)1) for solid
liquid transition was estimated integrating the corresponding endothermic peak. (b) Heating from )25 to
20 C at a rate of 10 C min)1, then cooling from 20 C
to )25 C at a rate of 10 C min)1, then a 10 C min)1
heating rate scanning from )25 C to )14 C and
holding for 120 min, followed by cooling to )25 C at
2 C min)1 before scanning from )25 C to 10 C at
2 C min)1. On basis of the last step, glass transition
region might be detected and the Tg were determined by
using the DSC programs.
Sensory evaluation
A ten member sensory panel evaluated one cheese of

each type at the end of the ripening period. The cheeses
were evaluated for aroma (creamy, milky, acid), colour

(neutral, white), texture (presence of holes and eyes,
elasticity, stickiness) and mouthfeel (creamy, hard,
crumbly). A ve-point hedonic scale (5 = very good;
1 = very poor) was used for aroma, colour, and tenpoint scale (10 = very good; 1 = very poor) for texture
and mouthfeel.
A two-way analysis of variance for the data for chemical

analysis, textural, sensory analysis (the factors being the
pectin gel addition and the protein base) were carried out
to determine the signicance of the individual dierences. Signicant means were compared using F-test and
standard deviations for mean values of chemical analysis
were also calculated. Simple correlations were performed
between the textural, rheology, thermal properties and
microstructure of the cheese analogue samples. All the
statistical analysis was conducted using the Matlab
(Version 6.5) commercial statistical package.
Composition of cheeses
The composition of full-fat, low-fat with pectin gel and

low-fat control cheese analogues were given in Table 2.
The moisture of low-fat control cheese analogue
and low-fat cheese analogue with pectin gel were
Table 2 Percentage chemical composition of cheese analogue samples

(mean SD)
Base
Fat
type
Sodium
FfSC
caseinate LfSC
LfcSC
Casein
FfC
LfC
LfcC
SSE
Base
Fat
Pectin
Moisture
Ash
Fat
Protein
Mean SD
Mean SD
Mean SD
Mean SD
0.608
0.705
0.704
0.597
0.687
0.704
***
*
4.707
4.593
4.147
3.573
3.530
3.633
***
*
0.008
0.004
0.005
0.007
0.012
0.002
***
0.242 17.257 0.086 17.213 0.445

0.047 8.173 0.189 18.260 0.535
0.166 7.837 0.146 17.543 0.186
0.060 18.033 0.110 18.217 0.398
0.115 9.690 0.155 17.470 0.780
0.114 7.817 0.172 18.073 0.764
FfSC, full-fat cheese analogue with sodium caseinate as protein base;

LfSC, low-fat cheese analogue with fat mimetic when the protein base is
sodium caseinate; LfcSC, low-fat cheese analogue control when the
protein base is sodium caseinate; FfC, full-fat cheese analogue with
casein as protein base; LfC, low-fat cheese analogue with fat mimetic
when the protein base is casein; LfcC, low-fat cheese analogue control
when the protein base is casein. SSE, statistical significance of effect (Ftest) from ANOVA; , not significant.
*P < 0.05; **P < 0.01; ***P < 0.001.
signicantly higher while their fat contents were lower

than those of full-fat cheese, as shown in Table 1. The
water added to the recipe was higher than the water
added to the full-fat cheese analogues and fat was
reverse. The use of dierent protein base aected the
values of ash, and protein content. The means of ash content in sample with sodium caseinate as protein base
were higher than that in sample with casein as protein
base. The Na+ might contribute to this result. The protein
contents in all samples were not dierent signicantly.
Microscopic analysis
Figure 2 showed that the type of protein base and water

content aected the microstructure of the cheeses. Three
distinct structures: protein aggregates (P), fat globules
(F) and cavities(C) (formed after water or air in the
cheese was eliminated when the prepared cheese analogues were to run the SEM test) were observed in SEM
micrographs of all cheese analogues (Fig. 2). Uniform
protein aggregate networks were observed in full-fat
cheese analogue samples as shown in Fig. 2a, d. The
microstructure of full-fat samples with sodium caseinate
as protein base was dierent from the structure of fullfat ones with casein as the protein base. The fat globules
in former were smaller than those in the latter and this
might be due to the emulsifying ability of sodium
caseinate was better than casein. Microstructures of lowfat samples with or without pectin gel addition and
sodium caseinate as the protein base looked similar to
those with casein as the protein base of the protein
aggregates. Because of the signicant dierences in
water content in the full-fat, low fat with pectin gel

addition, and low fat without pectin gel addition cheese
analogue (Table. 2), the full-fat cheese analogue in
which water content were about 10% less than that in
low fat with or without pectin gel addition cheese
analogues, the full-fat samples were more rm than the
low-fat and low-fat control samples and showed less
porous characteristic. The quantities of cavities in lowfat samples with pectin gel addition were less than those
in low-fat samples control. As introduced that a pectin
gel is formed when portions of homogalacturonan are
cross-linked to form a three-dimensional crystalline
network in which water and solutes are trapped
(Willatsa et al., 2006). Thus, the structure of formers
looked denser. The results might be attributed to the
eect of water entrapped in the pectin gel and formed a
more continuous phase. The microstructure aects the
mouthfeel, for the cheese, when the structure was more
compact, the mouthfeel might be harder. Structures of
pectin were not observed in the SEM micrographs.
Pectin was probably embedded in the protein matrix just
as the starch role in low-fat feta cheese (Sipahioglu
et al., 1999).
Textural properties
The mean values of the TPA parameters were given in

Table 3. Hardness was dened as the maximum peak
force during the rst compression cycle (rst bite) and
had often been substituted by the term rmness.
Adhesiveness was dened as the negative force area
for the rst bite and represented the work required to
overcome the attractive forces between the surface of a
food and the surface of other materials with which the
food came into contact, i.e. the total force necessary to
pull the compression plunger away from the sample. For
materials with a high adhesiveness and low cohesiveness,
when tested, part of the sample was likely to adhere to
the probe on the upward stroke. Springiness (originally
called elasticity) was related to the height that the food
recovered during the time that elapsed between the end
of the rst bite and the start of the second bite.
Chewiness meant the amount of the energy required to
masticate a solid food. This was the characteristic most
dicult to measure precisely, because mastication
involved compressing, shearing, piercing, grinding,
tearing and cutting, along with adequate lubrication
by saliva at body temperatures. Cohesiveness might be
measured as the rate at which the material disintegrated
under mechanical action. Tensile strength was a manifestation of cohesiveness. If adhesiveness was low
compared with cohesiveness, then the probe was likely
to remain clean as the product had the ability to hold
together. Cohesiveness was usually tested in terms of the
secondary parameters brittleness, chewiness and gumminess (from the instruments manual).
1585
1586
(a)
(b)
(c)
(d)
(e)
(f)
The full-fat cheese analogues were signicantly harder

than the low-fat cheese analogues whether with or
without pectin gel addition. This meant that fat globules
reinforce the gel strenth of the cheese analogues.
Although protein matrix content contributed to the
hardness of cheese (Romeih et al., 2002), and the protein
contents were similar in all samples (Table. 2), it was not
a surprising result because water broke up the protein
matrix and plays the role of lubricant to provide
smoothness and a softer texture. It was found that the
higher the water content in the cheeses, the softer the
samples. It was also found that the positive eect of
pectin gel on the hardness of low-fat cheese analogues
which could be attributed to both their ability of holding
water and made the cheese more compact and harder.
As the fat content of cheese analogues decreased, all the
TPA parameters except Springiness and Cohesiveness
decreased. The reason might be that the crystal fat
globules which could contribute to the rm structure of
the cheese analogues were partially substituted by pectin
Figure 2 Scanning electron micrographs of

(a) Full-fat cheese analogue with Sodium
Caseinate as protein base (FfSC), (b) Low-fat
cheese analogue with fat mimetic when the
protein base is Sodium Caseinate (LfSC),
(c) Low-fat cheese analogue control when
the protein base is Sodium Caseinate (LfcSC),
(d) Full-fat cheese analogue with Casein as
protein base (FfC), (e) Low-fat cheese
analogue with fat mimetic when the protein
base is Casein (LfC), (f) Low-fat cheese
analogue control when the protein base is
Casein (LfcC). The letters C, P, F by white
arrow in gures denoted the cavity formed
after water or air in the cheese was eliminated
when the prepared cheese analogues were to
run the SEM test, protein aggregate and fat
globule in the cheeses, respectively.
gel or only water. Compared with cheese analogues with

or without pectin gel addition, it was inspiring that the
TPA parameters of low-fat cheese analogues with pectin
gel addition were more similar to full-fat cheese
analogue. This might due to the pectin gel formed
network which could maintain the rmness of the cheese
analogues. It was reasonable to state that the pectin gel
had the potential to imitate the role of fat in the cheese
analogues but need further research on how to prepare
the pectin gel to better act as fat mimetic in cheese
analogues.
Dierent protein base also showed eects on the
hardness, chewiness and guminess of cheese analogues.
The samples with sodium caseinate as protein base were
signicantly higher than hardness, chewiness and guminess of those with casein as protein base. The results
were supported by the microstructure of the samples
(Fig. 2a, d) in which the protein micelle distributed
dierently. The likely comparison was found between
low-fat cheese analogues with dierent protein base.
Table 3 TPA (texture prole analysis) parameters for different cheese analogues
Hardness(g)
Springiness
Cohesiveness
Gumminess (g)
Chewiness (g)
Adhesiveness
ID
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Mean
SD
FfSC
LfSC
LFcSC
FfC
LfC
LfcC
SSE base
SSE fat
SSE pectin
681.11
172.59
12.22
307.34
105.19
31.11
**
**
1.8738
6.4456
0.2515
23.6136
2.7968
5.9404
0.92
0.90
0.94
0.92
0.96
0.95
0.0354
0.0200
0.0252
0.0473
0
0.0115
0.55
0.55
0.59
0.54
0.59
0.66
0.0071
0.0100
0.0115
0.0153
0.0058
0.0361
375.17
94.68
7.18
166.49
62.14
20.43
**
**
3.2173
2.7292
0.0503
4.4666
0.0264
0.7795
4919.04
1193.00
99.31
2166.84
786.15
261.23
**
**
45.5942
47.6264
5.2622
97.7403
10.6930
14.6051
)566.13
)88.03
)31.01
)331.47
)244.90
)125.53
**
***
24.2654
2.6890
1.3583
5.2124
2.8094
7.7138
FfSC, full-fat cheese analogue with sodium caseinate as protein base; LfSC, low-fat cheese analogue with fat mimetic when the protein base is sodium
caseinate; LfcSC, low-fat cheese analogue control when the protein base is sodium caseinate; FfC, full-fat cheese analogue with Casein as protein base;
LfC, low-fat cheese analogue with fat mimetic when the protein base is casein; LfcC, low-fat cheese analogue control when the protein base is casein.
Significance of effect (F-test) from ANOVA; , not significant.
*P < 0.05; **P < 0.01; ***P < 0.001.
Rheological properties
Rheological characterisation of cheese is important as a

means of determining body and texture characteristics
and to determine how these parameters are aected by
composition, and processing techniques (Konstance &
Holsinger, 1992). Cheese analogues are viscoelastic
materials and their viscoelastic behaviour can be inuenced by changes in their formulation caused by the
incorporation of llers which interact with the casein
matrix in the curd (Lobato et al., 2000). Instrumental
measurements of food texture are based on the foods
rheological properties. Dynamic low amplitude strain
testing oers very rapid result with minimal chemical
and physical change. Small strain dynamic rheological
methods have been used to dene both the elastic and
viscous nature of cheese. Such information is useful to
characterise and dierentiate cheese varieties (Tunick
et al., 1990). Such methods are implemented within the
linear viscoelastic region of the material and, therefore,
are designed to be nondestructive to the basic structure
of the material. Additionally, the elastic and loss moduli
become only a function of time and a function of the
magnitude of the stress or strain applied by performing
tests within the linear viscoelastic region (Tunick, 2000).
In low amplitude oscillatory shear experiments, the
sample was contained between two parallel plates and
underwent sinusoidally oscillating deformation as the
lower plate was xed while the upper plate was
oscillating at a specied frequency and transient
responses were recorded.
Rheology measurements were carried out in order to
investigate cheese heterogeneity and to discriminate
between the six cheese analogues. First, in order to
determine the linear viscoelastic region of the cheese
analogues, strain sweep was performed between 1% and
10%. As shown in Fig. 3a, from 1% to 10% strain the

complex modulus (G*) varied little. We selected 5%
strain as the preference to perform oscillatory rheological analysis.
The changes in G, G and tan d as functions of the
applied frequency for samples were presented in Fig. 3.
Among the investigated parameters, G and G
increased when the frequency increased, whereas tan d
decreased. As for G, the values of full-fat samples were
higher than the low-fat samples with pectin gel addition
and the low-fat control samples were the lowest
(Fig. 3b). Storage modulus (G) represents the solid-like
or elastic character of a viscoelastic material such as
cheeses. As water in cheese acts as a plasticizer, more
water will make the cheese analogues plastic and vice
versa. Lowering the water content increases the intermolecular linkages by concentrating the proteins,
whereas a higher moisture content caused a swelling of
the casein matrix and decreases the molecular interactions and hence caused a lower G value. Without
considering the protein base type, the full-fat cheese
analogues had higher G value. For the low-fat cheese
analogues, the samples with pectin gel added had higher
G value. It appeared that the rmness of cheese
analogues depend on three factors: the amount of water
bound to the cheese analogues interior framework, the
fat content and free water. The much free water the
samples had, the lower the G of the samples. As it is
known that hydrocolloid has the ability to hold water
and pectin gel had the good ability to entrap water. The
pectin gel addition lowered the free water content. So
the result which the samples with pectin gel added had
higher G value when compared with those low-fat
cheese control was unquestionable.
As for G, the curves of full-fat samples were distinct
higher than low-fat ones. The curve of low-fat control
1587
(a)
(b)
30 000
10 000
25 000
9000
8000
20 000
7000
G' (Pa)
|G*| (Pa)
Strain sweep step

6000
5000
4000
LfcC
FfSC
LfSC
FfC
LfC
LfcSC
15 000
10 000
3000
2000
5000
1000
0
0.05000
0
0
20 000
40 000
60 000
80 000
10 000
12 000
(d)
12 000
10 000
8000
2.000
1.750
LfcC
FfSC
LfSC
FfC
LfC
LfcSC
1.500
tan(delta)
(c)
100.0
Frequency (Hz)
% strain
G'' (Pa)
1588
6000
4000
1.250
LfcC
FfSC
LfSC
FfC
LfC
LfcSC
1.000
0.7500
0.5000
2000
0.2500
0
0.05000
100.0
Frequency (Hz)
0
0.05000
100.0
Frequency (Hz)
Figure 3 Oscillatary rheological characterisation of cheese analogue samples at 20 C: (a) linear viscoelastic region at a constant frequency of
1 Hz, (b) changes of G (storage modulus) at different frequency, (c) changes of G (loss modulus) at different frequency, (d) changes of tan d
(loss tangent) at different frequency.
samples with sodium caseinate as protein base situated

the lowest position. (Fig. 3c). The low-fat samples with
pectin gel addition and low-fat control samples show
similar curves.
As for tan d, the curves situation were just reverse to
the G vs. frequency curves of the cheese analogues
(Fig. 3b, d). An increase in loss tangent indicated a
change in the character of the cheese from solid-like to a
viscous- or liquid-like character. When the loss tangent
equalled one, this meant that the systems were on the
crossover of G and G. The tan d values of full-fat and
sodium caseinate based low-fat samples with pectin gel
addition were all below one along with increasing of
frequency. This showed that these samples were more
solid-like than liquid-like. Casein based low-fat sample
with pectin gel addition was liquid-like at initial
frequency and suddenly changed to solid-like when
frequency increased. The samples which were low-fat

cheese analogues without pectin gel addition were solidlike only at higher frequency. Pectin gels played an
important role in the cheese analogues systems. The
cross-linking structure of the gels improved the solidlike system through adsorption by protein (Marozienea
and Kruif, 2000).
Temperature sweep tests were performed and changes
of loss tangent during heating and cooling procedure
were shown in Fig. 4. The value of the loss tangent at
higher temperature was used as index of meltability
(Mounsey and ORiordan, 1999). The value of the loss
tangent at lower temperature may be used as index of
gelation. During the heating procedure, the loss tangent
values of low-fat control and casein-based sample with
pectin addition were all above one which showed that
the samples were liquid-like. This result was consistent
(a)
10.00
9.000
8.000
tan(delta)
7.000
protein base which had no point of crossover of G and

G and was solid-like throughout the cooling period.
This phenomenon suggested that it was a thermal
irreversible gel system and may be due to the properties
of protein. The ndings that the loss tangent of full-fat
cheese analogue with casein as protein base and low-fat
cheese analogue with same protein base and with pectin
gel addition meanwhile decreased during the heating
procedure and increased during the cooling procedure
when the temperature was higher were abnormal. The
reason of the abnormality might be elucidated if a
further study was carried out.
LfC
LfSC
FfC
FfSC
LfcSC
LfcC
6.000
5.000
4.000
3.000
2.000
1.000
Thermal analysis
0
20.1
60.0
Temperature (C)
(b)
8.000
7.000
6.000
LfC
LfSc
FfC
FfSC
LfcSC
LfcC
tan(delta)
5.000
4.000
3.000
2.000
1.000
0
19.9
60.0
Temperature (C)
Figure 4 Oscillatary rheological characterisation of cheese analogue
samples at heating and cooling procedure: (a) heating procedure,
(b) cooling procedure.
with the data shown in the (Fig. 3d). At 25 C, the

sodium caseinate-based samples with pectin gel addition
melted and full-fat cheese analogues melting temperatures were higher than it at approximately 30 C.
During cheese melting, the temperature at crossover
modulus was an indication of the softening point of
cheese, the onset temperature of rapid melt and ow.
Low-fat cheese analogues meltability was better than
full-fat samples because the moisture contents were
higher and the structure of the samples was not so rigid.
With similar fat content, the sample with casein as
protein base melted more easily and this may be due to
the proteins structure and other properties. During the
cooling procedure, all samples showed a gel-like behaviour and the temperature of the gelation was dierent
(Fig. 4b) except for full-fat samples with casein as the
The eects of fat contents or pectin gel on the melting

temperature, melting enthalpy and glass transition
temperature involved were determined by DSC (Fig. 5;
Table 4). The increase of fat content had a positive
inuence on the melting enthalpy (Delta H) and
negative inuence on the onset melting temperature.
Because of the higher fat content, the Ff cheese
analogues needed more energy than the Lf cheese
analogues and manifested poor meltability. The lowfat cheese analogues with pectin gel addition had a lower
Delta H and this might be due to the pectins structure
that changed and released energy at a temperature of
about 40 C or the pectin interacted with fat or protein
and decrease the Delta H. As to the glass transition
temperature (Tg), all samples Tg were approximately
)14.8 C and diered not so much. Overall, the results
of the thermal analysis were not according with the
theory hypothesis that higher moisture might decrease
the Tg of the cheese analogue. This result might be due
to the unsuitable of the measure means or complex
interaction of the ingredient added to prepare the
cheese analogues and this needs to further study and
discussion.
Sensory evaluation
As shown in Table 5, there were no dierences

(p > 0.05) in aroma and colour characteristics between
full-fat and low-fat cheese analogues. The panellists
found the dierences in texture and mouthfeel between
full-fat and low-fat cheese analogues with or without
pectin gel addition. However, the full-fat and low-fat
cheese analogue without pectin gel addition were poorer
in texture and mouthfeel as reected by their lower
score. The lower score in mouthfeel and texture of fullfat samples was likely due to the denser microstructure
which made the sample too hard. In addition, the lower
score in mouthfeel and texture of low-fat samples
without pectin gel addition might be due to the too soft
feeling of the samples resulting from the high moisture
level of the product. On all accounts, based on the
1589
(a)
23.0
22.5
LfC: To = 39.241 Te = 44.556 Delta H = 0.632
22.0
21.5
Heat flow endoup (mVV)
LfcSC: To = 38.069 Te = 46.541 Delta H = 1.249

21.0
LfcC: To = 41.467 Te = 42.215 Delta H = 1.094
20.5
20.0
FfC: To = 37.162 = Te = 45.794 Dekta H = .813
19.5
LfSC: To = 41.766 Te = 46.510 Delta H = 0.925
19.0
FfSC: To = 45.727 Dekta H = 1.455
18.5
18.0
34.51
36
38
40
42
44
46
48
50
52
54
55.5
Temperature (C)
(b) 39.626
39.624
39.622
39.620
LfSC: Tg = -14.794
39.618
Heat flow endoup (mVV)
1590
39.616
LfC:Tg = -14.803
39.614
39.612
FfSC: Tg = -14.790
39.610
Lfc C: Tg = -14.833
39.608
39.606
Lfc SC: Tg = -14.813
39.604
39.602
39.600
14.958
FfC: Tg = -14.793
14.90
14.85
14.80
14.75
14.70
Temperature (C)
14.65
14.60
Table 4 The onset temperature (To), ending temperature (Te), enthalpy

(Delta H, J g)1) of solidliquid transition and glass transition temperature (Tg) of dierent samples measured with dierent scanning
programs by DSC
)1
ID
To (C)
Te (C)
Delta H (J g )
Tg (C)
FfSC
LfSC
LfcSC
FfC
LfC
LfcC
37.833
41.766
38.069
37.162
39.214
41.467
45.727
46.510
46.541
45.794
44.556
42.215
1.455
0.925
1.249
1.813
0.632
1.094
)14.790
)14.794
)14.813
)14.793
)14.803
)14.883

when the protein base is casein.
14.55
14.5
Figure 5 DSC traces of cheese analogues

during the two different programs: (a) the
power varieties of the cheese analogues phase
transition, (b) the measurement of the glass
transition region.
sensory evaluation results, the authors were optimistic

to apply pectin gel to substitute partial fat to prepare
cheese analogues.
Conclusions
Fat reduction in semi-solid cheese analogues and the use

of pectin gel aected low-fat samples in dierent ways.
Fat reduction and moisture augment caused the microstructure of low-fat cheese analogues obviously more
loose. The low-fat samples showed lower hardness,
gumminess, chewiness and adhesiveness and low-fat
samples with pectin gel addition were more similar to
the full-fat cheese analogues. Pectin gel addition showed
that it can decrease the melt enthalpy of the samples
which may be due to the conformation or structural
change of pectin or pectin interacted with fat or protein
to decrease the Delta H. The low-fat cheese analogue
made with pectin gel addition was well received by the
sensory panel. The properties of the low-fat cheese with
Table 5 Sensory characteristics of full-fat and low-fat cheese analogues

Aroma
Base
Fat type Mean SD
Sodium
FfSC
caseinate LfSC
LfcSC
Casein
FfC
LfC
LfcC
SSE
Base
Fat
Pectin
4.65
4.25
4.34
4.57
4.20
4.12
0.28
0.34
0.35
0.37
0.42
0.22
Colour
Texture
Mouthfeel
Mean SD
Mean SD
Mean SD
4.74
4.68
4.57
4.59
4.73
4.63
6.54
8.17
6.37
6.73
8.39
7.81
*
*
6.46
7.68
5.43
6.23
7.89
5.47
*
*
0.24
0.34
0.46
0.60
0.51
0.41
0.76
0.49
0.66
0.65
0.75
0.52
0.74
0.43
0.48
0.39
0.68
0.46

when the protein base is casein.
SSE, statistical significance of effect (F-test) from ANOVA; , not significant.
*P < 0.05.
pectin gel such as texture, microstructure and rheology

were midst full-fat and low-fat without pectin gel
addition samples. The results suggest that pectin gel
may play the role of fat mimetic when added to cheese
analogues.
Acknowledgments
The authors are grateful to Shanghai Bright Dairy &

Food Co., Ltd for the help of preparation of the cheese
analogues.
References
Antoniou, K.D., Petrides, D., Raphaelides, S., Ben Omar, Z. &
Kesteloot, R. (2000). Texture assessment of French cheeses. Journal
of Food Science, 65, 168172.
AOAC Ocial Method 926.08 (1995). Moisture in cheese. In: Ofcial
Methods of Analysis of AOAC International. Vol. II, 16th edn. (edited
by V.A. Arlington) Pp. 58. Gaithersburg, MD: AOAC International.
AOAC Ocial Method 935.42 (1995). Ash of cheese. In: Ofcial
Methods of Analysis of AOAC International. Vol. II, 16th edn. Pp. 59.
Gaithersburg, MD: AOAC International.
Bachmann, H.P. (2001). Cheese analogues: a review. International
Deszczynski, M., Kasapis, S., MacNaughton, W. & Mitchell, J.R.
(2003). Eect of sugars on the mechanical and thermal properties of
agarose gels. Food Hydrocolloids, 17, 793799.
Doraiswamy, D. (2002). The origins of rheology: a short historical
excursion. Rheology Bulletin, 71, 717.
Drake, M.A. & Swanson, B.G. (1995). Reduced and low-fat cheese
technology: a review. Trends in Food Science and Technology, 6, 366
369.
Ehab, A.R., Alexandra, M., Costas, G.B. & Gregory, K.Z. (2002).
Low-fat white-brined cheese made from bovine milk and two
commercial fat mimetics: chemical, physical and sensory attributes.
Govindasamy-Lucey, S., Jaeggi, J.J., Johnson, M.E., Wang, T. &
Lucey, J.A. (2005). Use of cold ultraltered retentates for standardization of milks for pizza cheese: impact on yield and functionality.
Gunasekaran, S. & Mehmet Ak, M. (2003). Cheese Rheology and
Texture. New York: CRC Press.
Hort, J. & Grys, G.L. (2001). Developments in the textural and
rheological properties of UK Cheddar cheese during ripening.
Katsiari, M.C., Voutsinas, L.P. & Kondyli, E. (2002). Improvement of
sensory quality of low-fat Kefalograviera-type cheese with commercial adjunct cultures. International Dairy Journal, 12, 757764.
Kealy, T. (2006). Application of liquid and solid rheological technologies to the textural characterisation of semi-solid foods. Food
Konstance, R.P. & Holsinger, V.H. (1992). Development of rheological test methods for cheese. Journal of Food Technology, 46, 105109.
Lazaridou, A., Biliaderis, C.G., Micha-Screttas, M. & Steele, B.R.
(2004). A comparative study on structurefunction relations of
mixed-linkage (1 3), (1 4) linear -d-glucans. Food Hydrocolloids, 18, 837855.
Lobato, C.C., Aguirre, M.E., Vernon, E.J. & Sanchez, G.J. (2000).
Viscoelastic properties of white fresh cheese lled with sodium
caseinate. Journal of Texture Studies, 31, 379390.
Marozienea, A. & Kruif, C.G. (2000). Interaction of pectin and casein
micelles. Food Hydrocolloids, 14, 391394.
Marshall, R.T. (1992). Standard Methods for the Examination of Dairy
Products. Pp. 571573. Washington, DC: American Public Health
Association.
Mistry, V.V. (2001). Low fat cheese technology. International Dairy
Journal, 11, 413422.
Mounsey, J.S. & ORiordan, E.D. (1999). Empirical and dynamic
rheological data correlation to characterize melt characteristics of
imitation cheese. Journal of Food Science, 64, 701703.
Muir, D.D., Tamime, A.Y., Shenana, M.E. & Dawood, A.H. (1999).
Processed cheese analogues incorporating fat-substitutes 1. Composition, microbiological quality and avour changes during storage at
5C. Lebensmittel Wissenschaft und Technologie, 32, 4149.
Normand, V., Aymard, P., Lootens, D.L., Amicia, E., Plucknett, K.P.
& Frith, W.J. (2003). Eect of sucrose on agarose gels mechanical
behaviour. Carbohydrate Polymers, 54, 8395.
Paraskevopoulou, A., Athanasiadis, I., Blekas, G., Koutinas, A.A.,
Kanellaki, M. & Kiosseoglou, V. (2003). Inuence of polysaccharide
addition on stability of a cheese whey ker-milk mixture. Food
Romdhan, K. & Eric, D. (2003). Dynamic testing rheology and
uorescence spectroscopy investigations of surface to centre dierences in ripened soft cheeses. International Dairy Journal, 13, 973
985.
Romeih, E.A., Michaelidou, A., Biliaderis, C.G. & Zerridis, G.K.
(2002). Low-fat white brined cheese made from bovine milk and two
commercial fat mimetics, physical and sensory attributes. International Dairy Journal, 12, 525540.
Ross, K.A., Pyrak-Nolte, L.J. & Campanell, O.H. (2006). The eect of
mixing conditions on the material properties of an agar gelmicrostructural and macrostructural considerations. Food Hydrocolloids,
20, 7987.
Sanche, C., Zuniga-Lopez, R., Schmitt, C., Despond, S. & Hardy, J.
(2000). Microstructure of acid-induced skim milk-locust bean
gum-xanthan gels. International Dairy Journal, 10, 199212.
Siebel, R. & Sylvia, A.J. (1996). Handbook of Fat Replacers. Pp. 161
174. New York: CRC Press.
Sipahioglu, O., Alvarez, V.B. & Solano-Lopez, C. (1999). Structure,
physico-chemical and sensory properties of feta cheese made with
1591
1592
tapioca starch and lecithin as fat mimetics. International Dairy

Journal, 9, 783789.
Truong, V.D., Daubert, C.R., Drake, M.A. & Baxter, S.R. (2002).
Vane rheometry for textural characterization of Cheddar cheeses:
correlation with other instrumental and sensory measurements.
Lebensmittel Wissenschaft und Technologie, 35, 305314.
Tunick, M.H. (2000). Rheology of dairy foods that gel, stretch, and
fracture. Journal of Dairy Science, 83, 18921898.
Tunick, M.H., Nolan, E.J., Shieh, J.J., Basch, J.J., Thompson,

M.P., Malee, B.E. & Holsinger, V.H. (1990). Cheddar and
Cheshire cheese rheology. Journal of Dairy Science, 73, 1671
1675.
Willatsa, W.G.T., Knox, J.P. & Mikkelsen, J.D. (2006). Pectin: new
insights into an old polymer are starting to gel. Trends in Food
Original article
Consumer attitude and behaviour towards tomatoes after 10 years
of Flandria quality labelling
Wim Verbeke,* Liesbeth Van de Velde, Koen Mondelaers, Bianka Kuhne & Guido Van Huylenbroeck
Department of Agricultural Economics, Ghent University, Coupure links 653, B-9000 Gent, Belgium
(Received 23 August 2006; Accepted in revised from 23 April 2007)
Summary
In recent years, trust in food safety and food quality has decreased as a result of consecutive food crises.
Consequently, numerous quality labels signalling process-related credence characteristics have been
established. One of these labels is the Belgian Flandria quality label for fresh fruit and vegetables. Based
on cross-sectional data collected through a self-administered consumer survey (n = 373), this paper
addresses consumer attitudes, behaviour and perception towards tomatoes in general, and the Flandria
tomato label in particular. Buyers, who constitute 26.8% of the sample, perceive Flandria tomatoes as
superior to other tomatoes because of their guarantee of origin, better taste and stricter production control.
However, they also report the strongest perception of Flandria as an ordinary tomato as compared to nonbuyer segments. Overall, ndings indicate that the Flandria label after being intensively used for 10 years
for a wide range of other fruits and vegetables besides tomatoes has become fairly standard for tomatoes
with little perceived dierentiation apart from its certied production and origin.
Keywords
Consumer attitude, consumer behaviour, quality labelling, tomatoes.
Introduction
This paper focuses on consumer attitude and behaviour

towards quality labelled tomatoes in Belgium. The
analysis is based on cross-sectional consumer data
collected in 2003, i.e. 10 years after the introduction of
a specic quality label Flandria which was initially
introduced for tomatoes, but has been extended gradually to other vegetables as well. This label is dierent
from the organic standard. The main aim of this study
was to investigate whether this 10-year old and widely
used quality label for tomatoes appeals to specic
consumer needs. Before specifying the objectives of the
study, this introduction provides a concise overview of
the role of labels as a token of product quality and as an
information cue available for consumers in their decision-making process.
In recent years, several food safety crises like BSE,
dioxins, and foot and mouth disease have resulted in a
decrease of consumer trust in the performance of the
food chain and an increased need for guarantees of food
safety and food quality. There is a newly awakened
attention among industry, policy makers and scientists
*Correspondent: Fax: +32 9 264 6246;
e-mail: wim.verbeke@UGent.be
doi:10.1111/j.1365-2621.2007.01621.x
for the consumers interest in food production and their

lack of knowledge about it. Consequently, several
initiatives to certify and communicate typical product
attributes, like safety and healthiness issues, have been
developed by a number of dierent actors: the government, food industry, retailers, farmers associations and
primary producers (Salaun and Flores, 2001; Codron
et al., 2005).
To evaluate food products, consumers use several
evaluative criteria which concern the preferred outcomes
from purchase and consumption (e.g. high quality) and
which are determined by their goals or motives for
consumption (e.g. desire for high quality). On the basis
of these criteria, consumers form certain beliefs and
perceptions about the quality of food. The match
between consumers quality beliefs and their desires
forms the basis for consumers evaluative judgements
and preferences. Consequently, product quality can be
described as resulting from a bundle of dierent
characteristics or attributes that determines the products performance (Bredahl, 2003; Grunert, 2005).
Products embody search characteristics if buyers can
inspect their quality before purchase (e.g. price, appearance, etc.) and experience characteristics if quality
is revealed only after purchase and use (e.g. taste,
storage time, etc.). However, search and experience
1593
1594
Consumer attitude towards labelled tomatoes W. Verbeke et al.
characteristics are not the only attributes that determine

the overall perceived and experienced quality of foods.
Quality attributes which are not revealed even after
consuming the product, like animal welfare, safety and
health claims or origin are called credence characteristics
(Nelson, 1970; Darby and Karni, 1973). These credence
attributes mainly focus on the quality of the production
process, and less on the core product itself. Therefore,
quality-of-life issues, such as food ethics, environment
and health cannot be veried upon purchase or consumption. In recent years, these attributes have become
crucial as components of consumer value (Teisl et al.,
1999; Schroder and McEachern, 2004; Bernues et al.,
2003; Miles and Frewer, 2001; Wandel and Bugge,
1997). Consumers face specic diculties to form
quality expectations about the credence qualities of
fresh products like meat, sh, fruit and vegetables.
Quality labels can anticipate these problems, by giving
consumers information about credence characteristics of
the food products at hand (Grunert, 2002; Grunert,
2005; Verbeke, 2005). This has found expression in
many value-based labelling formats that dierentiate
products based on more sustainable, ethical, ecological
and healthy production methods (Teisl et al., 1999;
Nilsson et al., 2004). Food labelling can be seen as one
of the common routes to deliver a message from food
supply chains to the ultimate consumers. This information in the form of labels can contribute to the
completeness and accuracy of a consumers assessment
of search, experience and credence attributes, especially
in the specic case of generic and fresh food products
(Verbeke and Viaene, 1999).
Labelling can perform many dierent functions, like
the identication, description or promotion of food
products (Teague and Anderson, 1995; Bernues et al.,
2003). A food label must contain a minimum amount of
mandatory or legally set information, but a producer or
food chain is free to add any voluntary but correct
information (Przyrembel, 2004), which may contribute
to the dierentiation of food products. The market
mechanism allows this added information to be of extra
value when consumers are insuciently satised with
the minimum mandatory information. Owing to this,
food quality labels can be regarded as having a demand
and supply that interact to determine a market-clearing
price (Caswell and Mojduszka, 1996). Private or public
institutions that certify product quality are very useful in
providing information to buyers and such a certication,
identiable by a label, is a voluntary way used by
producers to signal product quality (Marette et al.,
1999). Therefore, an ecient quality labelling system
stipulates that a product gets the certicate in question
only when it meets a set of specic requirements
(Verbeke and Viaene, 1999).
On the one hand, food labels that carry information
on health and safety issues are considered to be sources
of direct consumer information. As such, labels are a

part of the information set used by consumers in making
product purchasing decisions (Salaun and Flores, 2001;
Verbeke and Ward, 2006). On the other hand, a label
can also serve as an important extrinsic product quality
cue in the consumer decision-making process (Caswell,
1992; Nayga, 1999; Issanchou, 1996). Through oering
a quality guarantee by a label, consumer re-assurance
can be established and the buying decision can favourably be inuenced (Caswell and Padberg, 1992). In this
way, quality labelling can dierentiate products by
increasing product attractiveness or assuring the consumer of a certain level of quality (Bernues et al., 2003;
Caswell and Mojduszka, 1996).
The perceptions and beliefs about the given information are inuenced by various factors, like individual
characteristics, situational, behavioural and attitudinal
factors, as well as product category involvement (Nayga,
1999). In some instances, consumers consider price as
an important indicator of product quality. Those who
associate a price premium with better quality are likely to
display a higher willingness-to-pay. However, Zeithaml
(1988) pointed out that the association between price and
quality is not obvious and also many other studies could
not conrm this relationship. The associated relationship
is inuenced by the product category, the individual
characteristics of the consumer and the availability of
other information about the product. Consequently, it
appears very dicult to isolate the relationship between
perceived price and perceived quality, and in the case at
hand, it remains to be investigated whether quality
labelled tomatoes are perceived as more expensive than
regular non-labelled tomatoes.
Most literature dealing with the role and consumer
perception of food quality labelling is situated in the
area of animal proteins like meat and sh (Verbeke
et al., 2002; Jary et al., 2004; Bredahl, 2003; Schroder
and McEachern, 2004; Verbeke and Viaene, 1999;
Bernues et al., 2003; Enneking, 2004). Nevertheless,
the use of quality labelling is also widespread in the fresh
fruit and vegetable category. In the specic case of fruit
and vegetables, taste and freshness are the primary
quality properties (Wandel and Bugge, 1997; ter Hofstede et al., 1996). Appearance and nutritional value are
also given high priority by a substantial part of the
consumers. Furthermore, the study by Wandel and
Bugge (1997) indicates that environmental concerns in
the evaluation of food quality are more prominent with
regard to fruit, vegetables and potatoes than meat. In
line with this, most consumer studies about fruit and
vegetable labelling initiatives focus on origin or specic
production methods taking environmental issues into
account, like organic or integrated pest management
practices (Anderson et al., 1996; Baker, 1998; Manhoudt et al., 2002; Vannoppen et al., 2001; Vannoppen
et al., 2002; Wandel and Bugge, 1997).
The objective of this study is, rst, to investigate

consumer perception and health benet beliefs about
tomatoes in general. Second, attribute importance
upon tomato purchase is analysed. Finally, consumer
perception of specic Flandria-labelled tomatoes is
assessed and compared across buyers and non-buyers.
Perceptions and benet beliefs about tomatoes in
general (as resulting from the rst parts of analysis)
are linked to individuals behaviour towards Flandria
in order to assess whether Flandria manages to appeal
to particular consumer needs or interests when purchasing tomatoes.
To understand the potential role of the Flandria
tomato label, distinct features of Flandria tomatoes are
presented in the rst part of this paper. In the second
part, details about the research method and empirical
ndings are presented and discussed.
Flandria-labelled tomatoes
The label Flandria is a quality label, which was

established in 1995 by combined eorts of producers,
auctioneers, sellers, exporters, research institutes and
the Flemish organisation (VLAM) for promoting
agriculture, horticulture, sheries and the agro-alimentary sector in Belgium and abroad. From the onset, the
Flandria label was specically used for tomatoes,
which is the main justication for our focus on
tomatoes in this study. Today, more than 50 fruits
and vegetables are sold under the Flandria quality
label, including among others tomatoes, salad, broccoli
and apples.
The Flandria label stands for products of excellent
quality, cultivated by family farm businesses in Flanders, for which traceability is perfectly feasible owing to
the use of unique product codes. The intrinsic quality of
Flandria goes beyond the European rst class norms
because the label includes more rigorous requirements
for the production of fruits and vegetables concerning
freshness, rmness, uniformity, taste and nutritional
value. The Flandria standard is inherently dierent from
the organic standard because the former allows the use
of chemical fertilizers as well as synthetic pesticides,
however only as an ultimate solution.
Flandria growers have to follow a strict code of
conduct for the cultivation of Flandria fruits and
vegetables, which was compiled by a co-operation of
six fruit and vegetable auctions in Belgium, called
LAVA (Logistic and Administrative Auction Association) (Lava (Logistieke en Administratieve Veilingassociatie), 2004). The code of conduct describes clearly
and precisely the production and hygienic conditions,
the grading and quality requirements, the requirements
for the package and labelling, the control system and the
consequences in the case of abuses. On top of this, the

cultivation method must be integrated in an environmental friendly and sustainable production, which
means for example that the pollination is done by
bumble-bees. In case of risk for diseases and plagues
natural remedies will be used, like ladybirds as a natural
predator. These integrated farming practices are less
restrictive compared to the organic standard. Only when
natural remedies prove inecient, the use of chemical
pest control is allowed. However, only control measures
with very specic eects are allowed, whereas preventive
and total-surface pest control measures are forbidden.
Owing to the family business character of the production process, producers are very conscious of their
individual responsibility. Furthermore, auctioneers and
specialised research centres assist them with technical
and administrative support. Flandria tomatoes are
cultivated in glasshouses, because the climate in Belgium
is not suitable for economically viable outdoors cultivation. Basically, the Flandria quality label guarantees
environment-friendly, Belgian tomatoes with a more
beautiful presentation and higher utility value. There is
a stricter control on production and integrated pest
management is practised. No specic claim on healthiness or nutritional value is made to dierentiate
Flandria from other tomatoes.
The quality and other properties of the Flandria
products are both internally and externally controlled.
Internal control is performed by the auctioneer upon
product delivery. The inspectors of the fruit and
vegetable auction check the registration forms and if
these are incorrect, incomplete or delayed, the inspectors
hold back the quality label. The external control
organisation monitors the auctioneers and checks the
products for contaminants and impurities as well as
extrinsic quality signs like appearance, freshness, colour,
calibration, uniformity and other properties (Flandria,
2004). The Flandria label is promoted by LAVA and
VLAM through generic advertising campaigns, distribution of table mats to restaurants, aps for shopping
carts, advertisements in magazines, video clips at retail
outlets, gratication for wholesalers who order large
amounts of Flandria tomatoes and farm visit days at the
growers. Individual tomatoes do not carry stickers with
the Flandria label, but Flandria is signalled to consumers through labels on the shelf, and on the bin and the
cover of the bin holding the tomatoes.
Consumer survey
Cross-sectional data were collected in February-March

2003 through a consumer survey in Flanders, which is
the northern Dutch-speaking region of Belgium. Participants were personally contacted through mall intercepts
in urban areas and a random-walk in selected streets in
rural areas, and asked to participate in the study.
1595
1596
Locations for recruitment were chosen based on convenience and respondents were selected through nonprobability quota sampling taking predetermined quota
on age into account. Questionnaires were self-administered at home and returned or collected by the
researchers after completion, which took 1520 min. A
total of 440 respondents was approached, of whom 413
respondents completed the questionnaire. The valid
response amounted to 373 after a quality check. Table 1
gives an overview of the sample characteristics for
gender, age, education, living environment and presence
of children. All respondents were responsible for the
purchase of food within their household. As a result, the
sample prole reects the primary role of women as
responsible person for purchase within the family.
Although the sample is not statistically representative in the sense that locations were not randomly
chosen, and respondents were not selected randomly
from the population, it should be noted that distributional characteristics of the sample closely match
with census data. For example, the age distribution of
the sample matches well with the overall population
distribution from census data, despite a slight over
sampling of the younger age groups (NIS (National
Institute for Statistics), 2002). Also with respect to the
presence of young children, a good match between the
sample and population distribution is realised. It
should be noted that the sample is biased towards
higher education. Nevertheless, consumers in the
<18 years education category still constitute a substantial part of it.
Table 1 Socio-demographic characteristics of the valid sample
(n = 373)
Frequency
(n)
Gender
Male
Female
Age
< 26 years
2640 years
4050 years
>50 years
Education
<18 years
>18 years
Living environment
Urban area
Rural area
Children <12 years
No
Yes
Frequency
(%)
Census data
(%) (NIS (National
Institute for
Statistics), 2002)
99
274
26.5
73.5
57
103
100
113
15.3
27.6
26.8
30.3
18.4
32.8
22.6
26.2
156
217
41.8
57.9
67.4
32.6
234
139
62.7
37.3
286
87
76.7
23.3
Measures
First, behaviour with respect to consuming tomatoes

was assessed using a consumption frequency scale (how
many times a week tomatoes are eaten), distinguishing
between summer and winter period. Second, consumers
perception of tomatoes in general was assessed on
ve-point interval scales on eight relevant attributes.
Perceived health benets from eating tomatoes were
measured by means of a ve-item Likert scale anchored
from Strongly disagree to Strongly agree. Relevant
attributes and benets were selected based on literature
review and are provided in Tables 2 and 3, respectively.
The third part of the questionnaire focussed on the
perceived importance of evaluative criteria upon tomato
purchase. Consumers rated 16 attributes on a ve-point
scale interval ranging from Not important at all to
Very important. Fourth, associations with Flandria
quality labelled tomatoes were probed for. A list of 24
attributes was drafted based on the Flandria product
information leaets. The statements about associations
with Flandria were phrased as: To what extent do you
associated a Flandria labelled tomato with and
evaluated on a ve-point interval scale ranging from
Not at all to Very much. Finally, relevant sociodemographic characteristics were included.
Empirical findings
Behavioural characteristics of the sample
Within the valid sample (n = 373), more than 75% of

the respondents buy tomatoes in the supermarket, about
40% of the respondents buy tomatoes in a small grocery
store (i.e. greengrocer), and about 20% on the street
Table 2 Principal component analysis of general perception of tomatoes, factor loadings >0.50 after Varimax rotation
% Variance explained
80.3
19.7
Tasteful
Delicious
Healthy
Nutritional value
Safe
Trustworthy
Shelf-life
Availability
Cronbachs alpha
Mean
SD
Factor 1
Nutritional
and sensory (43.1)
Factor 2
Credence
(16.6)
Factor 3
Convenience
(13.2)
0.87
0.86
0.77
0.64
0.91
0.90
0.83
3.95
0.64
0.86
3.30
0.70
0.90
0.62
0.41*
n.a.
n.a.
* Since the alpha value < 0.60, the item scores are not merged into a
single dimension score; n.a. = not applicable; Shelf-life mean (SD) = 3.27
(0.91); Availability mean (SD) = 4.30 (0.75).
Table 3 Health benet beliefs from eating

tomatoes, frequency distribution (%), mean
and SD on ve-point scale (n = 373)
Strongly disagree
Disagree
Neutral
Agree
Strongly agree
Mean
SD
market. One quarter of the respondents grow their own

tomatoes in summer. The socio-demographic prole of
those who grow tomatoes does not dier from those
who do not with respect to age, gender, education and
presence of children (all P > 0.10). Obviously, the share
of respondents growing their own tomatoes is signicantly lower in urban areas (17.5%) compared to rural
areas (38.8%) (v = 20.99; P < 0.001).
Tomato consumption shows a clear seasonal pattern:
93.6% of the respondents claim to eat tomatoes at least
once a week in summer, vs. only 53.9% in winter. A
paired sample t-test conrms the dierence in the
average frequency of eating tomatoes in summer vs.
winter (P < 0.001). A positive and signicant correlation between the consumption frequency of tomatoes in
summer and in winter is found (Pearson r = 0.489,
P < 0.001).
Several socio-demographic variables associate with
the consumption frequency of tomatoes. Signicantly
more women (54.6%) eat tomatoes in summer as
compared to men (44.4%) (P < 0.001), whereas no
gender dierence is seen in winter. This suggests that the
vegetable consumption patterns of women are more
liable to seasonal changes than the eating habits of men.
Consumers with children eat tomatoes more frequently
than consumers without children (P < 0.001). In summer, the consumption of tomatoes is also related with
the age of the consumers: consumers older than 35 years
eat tomatoes more frequently than younger consumers
(P = 0.002). There is no age-related dierence in
winter. No association is found between the level of
education and tomato consumption frequency.
General perception of tomatoes
Attribute perceptions for tomatoes in general are

grouped on the basis of a principal components analysis,
which yields three factors with eigenvalues above one.
These three attribute factors account for 72.9% of the
variance in the original data and are referred to as (in
descending order of explained variance): nutritional
and sensory (43.1%), credence (16.6%) and convenience (13.2%). The nutritional and sensory factor
includes the attributes tasteful, delicious, healthy, and
Reduced risk
on CVD
Source of
dietary fibre
Reduced risk
on cancer
Nutritious
Source of
vitamins
4.7
11.0
56.7
24.0
3.6
3.11
0.82
2.5
7.7
42.5
37.3
10.1
3.45
0.87
4.1
10.2
54.8
21.8
9.1
3.21
0.90
0.8
4.1
30.6
47.3
17.2
3.76
0.81
0.5
1.4
18.1
54.1
25.9
4.04
0.74
nutritional value. The credence and the convenience

component only contain two attributes each, trustworthy and safe on the one hand and shelf-life and
availability on the other hand. The Cronbachs alpha
value, which is an indication of composite scale reliability, is high for the nutritional and sensory and
credence factors, which indicates that the separate
items can be merged into one dimension score. The
perception score on the nutritional and sensory
dimension is signicantly higher than on the credence
dimension (P < 0.001). The alpha value of the convenience items is below the cut-o value of 0.60 for
sucient internal reliability. Since these two items are
poorly correlated, they are not merged into one dimension score.
An independent samples t-test shows that there is only
a signicant gender dierence on the nutritional and
sensory dimension: men have signicantly lower perception scores on the nutritional and sensory value of
tomatoes as compared to women (P = 0.014). Consumption frequency is associated with the score on the
nutritional and sensory factor. The Pearson-correlation
coecient equals 0.185 (P < 0.001), which indicates
that consumers who nd tomatoes healthy and tasty eat
tomatoes more frequently than consumers who share this
belief only to a lesser extent. No such relationship with
the credence or convenience items is found. This nding
is indicative that beliefs related to nutrition and taste are
the main motivations of consumers for eating tomatoes.
Health benefit beliefs from eating tomatoes
Each health benet item scores on average higher than

3, which indicates that consumers have rather strong
beliefs in the health benets from eating tomatoes. The
benecial inuence of the vitamin intake (M = 4.04)
and the nutritional value (M = 3.76) from eating
tomatoes receive the highest mean score. Benet beliefs
from the intake of dietary bre through tomato
consumption scores also high with M = 3.45. The
beliefs that tomatoes reduce the risk for cardiovascular
diseases (CVD) and cancer score rather neutral, which
indicates that consumers are not strongly convinced of
these risk reducing benets (Table 3).
1597
1598
The scores of the ve items result in a Cronbachs

alpha of 0.74, which indicates that the ve variables can
be aggregated into one health benet belief construct
score. This aggregated construct score shows that
women (M = 3.56) have a signicantly stronger belief
in the benecial inuence of tomatoes on health than
men (M = 3.36) (P = 0.003), and that consumers with
children younger than 12 (M = 3.63) have a signicantly stronger benet belief than consumers without
young children (M = 3.42) (P = 0.005). Health benet
belief of tomatoes is signicantly correlated with the
nutritional and sensory, and with the credence factor
score: the Pearson correlation coecients are 0.240 and
0.268, respectively (both P < 0.001).
Importance of tomato attributes upon purchase
Principal components analysis of the importance scores

upon purchase for 16 tomato attributes reveals a fourfactor solution explaining 60.2% of the variance in the
original data (Table 4). A rst factor is based on
credence characteristics and is referred to as product
identity. The second factor pertains entirely to typical
search characteristics or product appearance. Factor
three includes the sensory expectations, i.e. attributes
like taste, juiciness and texture. Quantitative search
attributes like price and expiration date constitute the
fourth factor. The latter are often used as predominant
Table 4 Principal component analysis based on evaluation criteria
upon tomato purchase, factor loadings >0.50 after Varimax
rotation
% variance
explained
Health claim
Organic label
Eco label
Guarantee of origin
Label from Belgium
Brightness
Size
Presentation
Packaging
Colour
Taste
Juiciness
Texture
Price
Expiration date
Cronbachs alpha
Mean
SD
Factor 1
Identity
(25.9)
Factor 2
Appearance
(16.3)
Factor 3
Sensory
(11.1)
Factor 4
Behavioural characteristics towards Flandria-labelled

tomatoes
The respondents are subdivided in three groups based

on their claimed awareness and use of Flandria-labelled
tomatoes. The rst group contains the respondents who
claim to have never heard of the Flandria label (34.1%):
the unaware. The second group of respondents are
consumers who know the label and bought it sometimes
in the past, but do not buy it anymore: the aware nonbuyers (39.2%). The third group consists of respondents who claim to buy Flandria tomatoes on a regular
basis: the buyers (26.8%).
The three groups dier in terms of gender and age
(Table 5). Only 15.2% of the men belong to the group of
buyers, whereas this is 31.0% for the women. Also,
42.4% of the men are unaware of the Flandria label, vs.
only 31.0% of the women (v = 9.98; P = 0.007).
Although a similar amount of Flandria buyers is found
within the below (26.3%) as in the above 40 years
(27.2%) age categories, there is tendency (v = 4.80;
Table 5 Socio-demographic characteristics of the Flandria user
segments (%)
Data (6.9)
Aware
Unaware non-buyers Buyers Chi-square
(34.1%)
(39.1%)
(26.8%) P-value
0.86
0.84
0.83
0.83
0.69
0.71
0.67
0.64
0.6
0.57
0.75
0.74
0.67
0.88
3.13
0.96
quality cues in the case of unbranded products, or in

specic situations where consumers face heightened
levels of quality uncertainty (Zeithaml, 1988; Bredahl,
2003). The Cronbachs alpha value is high for the
identity, appearance and sensory factor, which
indicates that the separate items can be merged into
dimension scores. The sensory expectations dimension
receives the highest importance score upon tomato
purchase.
0.71
3.14
0.64
0.64
4.08
0.48
0.85
0.62
0.46*
n.a.
n.a.
* Since the alpha value < 0.60, the item scores are not merged into a
single dimension score; n.a. = not applicable; Price mean (SD) = 3.87
(0.89). Expiration date mean (SD) = 4.05 (0.79).
Gender
Female (n = 274)
Male (n = 99)
Age
< 40 years (n = 160)
> 40 years (213)
Education
< 18 years (n = 156)
> 18 years (n = 217)
Living environment
Urban area (n = 234)
Rural area (n = 139)
Children <12 years
No (n = 286)
Yes (n = 87)
Growing tomatoes oneself
No (n = 278)
Yes (n = 95)
0.007
31.0
42.4
38.0
42.4
31.0
15.2
28.8
38.0
45.0
34.7
26.3
27.2
33.3
34.7
39.7
38.4
27.0
26.9
32.9
36.0
37.6
41.7
29.5
22.3
34.8
31.8
39.7
37.4
25.5
30.8
31.3
42.1
41.0
33.7
27.7
24.2
0.091
0.955
0.317
0.616
0.156
P = 0.091) that younger respondents have a higher

awareness of the existence of Flandria tomatoes as
compared to the older respondents. The Flandria label is
known and purchased to almost the same extent by
consumers living in urban as in rural areas (v = 2.298;
P = 0.317). The number of Flandria buyers from
families with young children (30.8%) and without
young children (25.5%) is not signicantly dierent
(v = 0.97; P = 0.616). Finally, there is no signicant
association between the level of education, and whether
or not the respondents know and or buy Flandria
tomatoes.
According to One-way anova (F = 3.168;
P = 0.043), Flandria buyers eat tomatoes signicantly
more frequently in summer (3 to 4 days out of 7) than
the unaware (23 days a week). The aware non-buyers
eat on average three times a week tomatoes, which is not
signicantly dierent from the other two groups.
Although One-way anova reveals only a marginally
signicant dierence for tomato consumption in winter
(F = 2.759; P = 0.065), the post hoc Duncan test
indicates that Flandria buyers also eat tomatoes more
frequently in winter (12 days a week) than the unaware
(1 day or less than 1 day a week). Although there is
tendency that people who grow their own tomatoes in
summer are more unaware of Flandria (42.1% vs.
31.3% unaware among the non-growers), the observed
association between growing tomatoes oneself and user
segment membership is not statistically signicant
(v = 3.72; P = 0.156) (Table 5).
With respect to the perception of tomatoes in general,
the three user segments dier only in their perception of
the price tomatoes. Consumers who do not know about
Flandria (i.e. unaware) evaluate tomatoes signicantly
less expensive (M = 2.78) than the aware non-buyers
and buyers (both with M = 3.02 on 5) (F = 3.67;
P = 0.027) (data not shown in table format). Flandria
buyers have a signicantly stronger health benet belief
from eating tomatoes (M = 3.75 on 5) compared with
the unaware (M = 3.36) and the aware non-buyers
(M = 3.47) (F = 13.47; P < 0.001). Hence, Flandria
tomato buyers are more convinced that tomatoes are
benecial for human health. This corroborates the
observation that these consumers eat more tomatoes
and that the respondents who eat more tomatoes more
strongly believe in the health advantages (signicant
dierence in the winter).
The buyers have also a signicantly (P = 0.007) more
positive attitude (M = 3.37 on 5) towards labelled
tomatoes as compared to the aware non-buyers
(M = 3.21) and the unaware (M = 3.13) (F = 4.96;
P = 0.007). Upon tomato purchase, Flandria buyers
attach signicantly more importance to the factor
identity (F = 5.72; P = 0.004) and sensory
(F = 5.22; P = 0.006) than the aware non-buyers and
the unaware.
Regarding the characteristics, consumers associate

with Flandria tomatoes, the unaware and the aware
non-buyers have, with respect to all criteria, a neutral to
slightly negative attitude, whereas the buyers have in
each case a signicantly (P < 0.001) more positive
perception than the other two groups (Table 6). All
consumers are well aware of the fact that Flandrialabelled tomatoes have Belgian origin. The most
important characteristics for the buyers are a (perceived)
better quality, better taste and a stricter control on the
production of Flandria tomatoes. It should also be
noted that the buyers associate Flandria tomatoes more
strongly with an ordinary (i.e. average, of no exceptional
quality) tomato than the other groups and that they
perceive Flandria tomatoes as more expensive. The
association with better quality and ordinary at the
same time by the same buyer group indicates that these
consumers have become used to high quality tomatoes,
and that their consider high quality as quite normal, or
as their new standard. Only with respect to organic
Table 6 Association with Flandria-labelled tomatoes depending on
user segment
Unaware Aware
Buyers ANOVA
F
non-buyers
n = 127
n = 146
n = 100 P-value
Belgian origin
Better quality
Better taste
Stricter production control
Better flesh firmness
Better evenly colour
More trustworthy
Better shaped
More juicy
More shining look
Higher price
More natural
Loose tomato*
Integrated pest management
Cluster tomato*
Safer
Better presentation
User friendly packaging
Bigger fruits
Healthier
Environmental friendly
Ordinary tomato
Pesticide free
Organic production
2.72a
2.72a
2.62a
2.73a
2.64a
2.62a
2.66a
2.62a
2.65a
2.63a
2.70a
2.69a
2.66a
2.65a
2.67a
2.63a
2.65a
2.63a
2.63a
2.66a
2.64a
2.68a
2.65a
2.67a
3.18b
3.01b
2.87b
3.00b
2.93b
2.96b
2.93b
2.91b
2.90b
2.91b
3.00b
2.96b
2.92a
2.91b
2.79a
2.88b
2.95b
2.86b
2.99b
2.96b
2.88b
2.85a
2.93b
2.84a
4.04c
3.86c
3.71c
3.68c
3.66c
3.57c
3.54c
3.53c
3.50c
3.45c
3.44c
3.42c
3.40b
3.38c
3.38b
3.34c
3.29c
3.24c
3.23c
3.21c
3.18c
3.15b
2.97b
2.84a
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.013
0.212
a,b,c The mean scores with different superscripts are significantly

different based on a post hoc Duncan test.
* Cluster tomatoes are tomatoes sold on a vine stem, in contrast with
tomatoes sold in containers (referred to as loose tomatoes). Hence, the
terms cluster and loose refer to how the tomato is marketed. Note also
that it concerns in most cases different varieties of tomatoes.
1599
1600
production, the three groups have a similar perception

(P = 0.212), which matches with reality since Flandria
tomatoes are not organically produced (Table 6). It
should be noted that whereas aware non-buyers and
buyers can base their perception on knowledge or prior
experience, the unaware probably form an opinion
based on suppositions or predispositions.
The unaware and aware non-buyers were also asked
about their motivation to eventually shift towards
buying Flandria tomatoes in the future. For both
groups, the most important motivation would be if no
other tomatoes were available, followed closely by if
more information was provided and if proven to be
healthier. Neither product image nor price are perceived
as potential motives or barriers (Table 7).
Discussion and conclusions
After some consecutive food safety crises, several

initiatives were undertaken to restore consumer trust
in the food chain. Several food quality labels were
established. Most of these in the vegetable chain focus
on ecological and safety aspects of the production
method, i.e. adding credence characteristics to the
product, which are ultimately signalled by means of a
quality label. The Flandria quality label stands in the
rst place for high quality from Belgium, produced
through an environmental friendly production process.
In this study, we investigated the general perception,
beliefs and attribute importance with respect to tomatoes, and checked whether the Flandria prole matches
consumer expectations and demand. It is important to
draw attention to some limitations associated with the
study. First, the study was conducted in one region and
considering one specic quality label only. Hence,
results remain to be conrmed in other regions and
countries or with other labels. Second, the sample was a
convenience sample, non-representative for the Belgian
population at large. In future studies, it would be
Table 7 Potential motivations to buy Flandria tomatoes in the future
(%)
Unaware
If Flandria tomatoes were had
n = 127
Aware
non-buyers
n = 146
No alternative available
Better information
Proven to be healthier
More environmentally friendly
Better taste
Better availability
Better known
Cheaper
Better production control
79.3
76.6
73.3
64.4
71.7
61.0
52.5
51.3
61.2
78.5
76.4
75.8
64.4
63.6
63.6
59.9
58.4
58.4
important to employ larger samples, representative for

the Belgian population.
The general perception of tomatoes is explained by
three factors (in descending order of importance): nutritional and sensory, credence and convenience. The
two factors explaining most of the variance corroborate
with the characteristics found in the Flandria prole.
Apparently, through its positioning, communication
strategy and resulting image, the Flandria label meets a
particular market demand. Analysis of the behaviour of
the consumers towards Flandria-labelled tomatoes reveals that 73.3% of the consumers are familiar with the
label. Although the buyers have a more positive perception of Flandria than non-buyers, both groups provided a
similar ranking of attributes for Flandria (i.e. strongest
associations with the eective content of Flandria). For
both groups, the strongest associations are with the
Belgian origin, better quality and stricter control on
production. Interestingly, Flandria tomato buyers report
a stronger belief in health benets from eating tomatoes
in general as compared to non-buyers. Since Flandria
tomatoes have never been brand advertised as superior in
nutritional value, it is hypothesised that the healthiness
appeal of Flandria tomatoes among buyers although
only moderate results indirectly from perceptions of
the production process and stricter control. Although the
general perception of the consumers coincides with the
Flandria label characteristics, the question rises whether
this is reected in the purchasing decision process.
Principal component analysis shows that four factors
determine the purchase behaviour: product identity,
product appearance, sensory expectations and quantitative search attributes.
Concerning the purchase of Flandria-labelled tomatoes, the consumers can be divided in three groups:
unaware (34.1%), aware non-buyers (39.2%) and buyers
(26.8%). Taking into account the strong match between
consumers expectations about tomatoes, as exemplied
by their general perception and attribute importance
levels on the one hand, and the Flandria label prole on
the other hand, the group of buyers is rather small. As a
consequence, a larger market share should be possible,
especially given the high importance consumers attach
to sensory expectations upon tomato purchase, which is
a major point of dierentiation claimed by Flandria.
When respondents were probed about eventual barriers
for purchasing Flandria tomatoes, the strongest arguments were better information, a proven health advantage in comparison with other tomatoes and only if no
other alternatives are available. It is hard to believe,
however, that a lack of information explains the small
share of conscious buyers, especially seen the intensive
communication eorts for the label. It should be noted
that aware non-buyers dier largely from buyers in their
beliefs related to Flandrias Belgian origin, stricter
production control and trustworthiness. Perhaps
advertisements and information should also concentrate

more on production control, eventually in combination
with a guarantee of origin, which might improve
perceived trustworthiness and stimulate purchasing
among the aware non-buyers.
Another potential explanation for the relatively low
share of conscious Flandria buyers could be that after
the food safety crises and the consecutive initiatives by
the government, food industry and other stakeholders in
the food chain, consumers consider food quality and
food safety as a standard attribute. Findings indicate
that Flandria buyers are heavy tomato users, and at the
same time have the strongest perception of Flandrialabelled tomatoes as being an ordinary tomato. As
such, consumers seem not to perceive very specic added
value of the Flandria label. Findings indicate that this
label being regularly advertised and being used for
around 50 fruit and vegetable categories for 10 years,
and being positioned in the middle as compared to
non-labelled or other labelled tomatoes has become
quite standard and little dierentiated.
References
Anderson, M.D., Hollingsworth, C.S., Van Zee, V., Coli, W.M. &
Rhodes, M. (1996). Consumer response to integrated pest management and certication. Agriculture, Ecosystems and Environment, 60,
97106.
Baker, G.A. (1998). Strategic implications of consumer food safety
preferences. International Food and Agribusiness Management Review, 1, 451463.
Bernues, A., Olaizola, A. & Corcoran, K. (2003). Labelling information
demanded by European consumers and relationships with purchasing
motives, quality and safety of meat. Meat Science, 65, 10951106.
Bredahl, L. (2003). Cue utilisation and quality perception with regard
to branded beef. Food Quality and Preference, 15, 6575.
Caswell, J.A. & Mojduszka, E.M. (1996). Using informational labeling
to inuence the market for quality in food products. American
Journal of Agricultural Economics, 78, 12481253.
Caswell, J.A. & Padberg, D.I. (1992). Toward a more comprehensive
theory of food labels. American Journal of Agricultural Economics,
74, 460468.
Caswell, J.A. (1992). Current information levels on food labels.
American Journal of Agricultural Economics, 74, 11961201.
Codron, J.-M., Giraud-Heraud, E. & Soler, L.-G. (2005). Minimum
quality standards, premium private labels, and European meat and
fresh produce retailing. Food Policy, 30, 270283.
Darby, M.R. & Karni, E. (1973). Free competition and the optimal
amount of fraud. Journal of Law and Economy, 16, 6788.
Enneking, U. (2004). Willingness-to-pay for safety improvements in
the German meat sector: the case of the Q&S label. European Review
of Agricultural Economics, 31, 205223.
Flandria. (2004). Verse groenten en fruit Brussel: VLAM. http://
www.andria.vlam.be.
Grunert, K.G. (2002). Current issues in the understanding of
consumer food choice. Trends in Food Science & Technology, 13,
275285.
Grunert, K.G. (2005). Food quality and safety: consumer perception
and demand. European Review of Agricultural Economics, 32,
369391.
ter Hofstede, F., Steenkamp, J.E.B.M. & van Trijp, J.C.M. (1996).
Pan-Europese studie naar de consument van groenten en fruit. sHertogenbosch: Stichting Agro Keten Kennis.
Issanchou, S. (1996). Consumer expectations and perceptions of meat

and meat product quality. Meat Science, 43, 519.
Jary, S., Pickering, H., Ghulam, Y., Whitwarsh, D. & Wattage, P.
(2004). Consumer choices for quality and sustainability labelled
seafood products in the UK. Food Policy, 29, 215228.
Lava (Logistieke en Administratieve Veilingassociatie) (2004). LAVA
overkoepelt 6 Belgische groente en fruitveilingen. Boortmeerbeek:
LAVA. http://www.lava.be.
Manhoudt, A.G.E., van de Ven, G.W.J., de Haes, H.A.U. & de Snoo,
G.R. (2002). Environmental labelling in the Netherlands: a framework for integrated farming. Journal of Environmental Management,
65, 269283.
Marette, S., Crespi, J.M. & Schiavina, A. (1999). The role of common
labelling in a context of asymmetric information. European Review
of Agricultural Economics, 26, 167178.
Miles, S. & Frewer, L.J. (2001). Investigating specic concerns about
dierent food hazards. Food Quality and Preference, 12, 4761.
Nayga, R.M. (1999). Toward an understanding of consumers
perceptions of food labels. International Food and Agribusiness
Management Review, 2, 2945.
Nelson, P. (1970). Information and consumer behavior. Journal of
Political Economy, 78, 311329.
Nilsson, H., Tuncer, B. & Thidell, A. (2004). The use of eco-labeling
like initiatives on food products to promote quality assurance - is
there enough credibility? Journal of Cleaner Production, 12, 517526.
NIS (National Institute for Statistics) (2002). Household composition
statistics. Brussels: NIS.
Przyrembel, H. (2004). Food labelling legislation in the EU and
consumers information. Trends in Food Science & Technology, 15,
360365.
Salaun, Y. & Flores, K. (2001). Information quality: meeting the needs
of the consumer. International Journal of Information Management,
21, 2137.
Schroder, M.J.A. & McEachern, M.G. (2004). Consumer value
conicts surrounding ethical food purchase decisions: a focus on animal welfare. International Journal of Consumer Studies, 28, 168177.
Teague, J.L. & Anderson, D.W. (1995). Consumer preferences for safe
handling labels on meat and poultry. The Journal of Consumer
Affairs, 29, 108127.
Teisl, M.F., Roe, B. & Levy, A.S. (1999). Ecocertication: why it may
not be a eld of dreams. American Journal of Agricultural
Economics, 81, 10661071.
Vannoppen, J., Verbeke, W. & Van Huylenbroeck, G. (2002).
Consumer value structures towards supermarket vs. farm shop
purchase of apples from integrated production in Belgium. British
Food Journal, 104, 828844.
Vannoppen, J., Verbeke, W., Van Huylenbroeck, G. & Viaene, J.
(2001). Consumer valuation of short market channels for fresh food
through labelling. Journal of International Food and Agribusiness
Marketing, 12, 4169.
Verbeke, W. & Viaene, J. (1999). Consumer attitude to beef quality
labeling and associations with beef quality labels. Journal of
International Food and Agribusiness Marketing, 10, 4565.
Verbeke, W. (2005). Agriculture and the food industry in the
information age. European Review of Agricultural Economics, 32,
347368.
Verbeke, W. & Ward, R.W. (2006). Consumer interest in information
cues denoting quality, traceability and origin: An application of
ordered probit models to beef labels. Food Quality and Preference,
17, 453467.
Verbeke, W., Ward, R.W. & Avermaete, T. (2002). Evaluation of
publicity measures relating to the EU beef labelling system in
Belgium. Food Policy, 27, 339353.
Wandel, M. & Bugge, A. (1997). Environmental concern in consumer
evaluation of food quality. Food Quality and Preference, 8, 1926.
Zeithaml, V. (1988). Consumer perceptions of price, quality and value
A means-end model and synthesis of evidence. Journal of
Marketing, 52, 222.
1601
1602
Original article
Colour improvement of common carp (Cyprinus carpio) fillets by
hydrogen peroxide for surimi production
Ali Jafarpour,1 Frank Sherkat,1 Brian Leonard2 & Elisabeth M. Gorczyca3*
1 Food Science, School of Applied Sciences, RMIT University, Melbourne, Vic. 3001 Australia
2 Biology and Biotechnology, School of Applied Sciences, RMIT University, Melbourne, Vic. 3001 Australia
3 RMIT University, School of Applied Sciences, Food Science, City Campus, GPO Box 2476 V Melbourne, Vic. 3001, Australia
(Received 6 September 2006; Accepted in revised from 30 April 2007)
Summary
The preferred colour for surimi is white, but surimi prepared from light llets of common carp (Cyprinus
carpio) is slightly pink. Hydrogen peroxide (H2O2; 13% v v) with and without sodium tri-polyphosphate
(STP; 12% w v) was added to a sodium carbonate bath (pH 7.011.5) resulting in a nal pH range of 4.4
10.1 which was injected into carp llets. After soaking and tumbling for 30 min at 410 C, the llets were
evaluated for colour and water holding capacity (WHC). Fillets tumbled with treatment solution with
dierent pH levels (7.011.5), but with no H2O2 or STP added, had improved colour with signicantly
(P < 0.05) higher L* compared with untreated llets as the control. However, the colour improvement [(L*
and colour deviation (DE)] was not signicantly dierent (P > 0.05) within the pH levels (7.011.5) trialled.
With increasing H2O2 levels (13%), llets became lighter and DE increased signicantly (P < 0.05),
especially with a 3% H2O2 treatment at pH of 10.5 (adjusted pH before H2O2 addition, actual pH after H2O2
addition was 8.2). The whiteness (L*)3b*) of kamaboko produced from treated (3% H2O2, pH 10.5)
common carp light llets was not signicantly dierent to that of kamaboko from Alaska pollock and
threadn bream. Treatments combining H2O2 (3%) with STP (12%) signicantly reduced the L* value
obtained in comparison with llets treated with only H2O2 (3%). Similarly, llets treated with STP (1%)
alone, resulting in lower L* values, irrespective of treatment pH (7.011.5). WHC, an indicator of the quality
of the llet texture, increased from 816 g kg at pH 7.0 without STP to 841 g kg at pH 11.5 with 1% STP.
Treatment with H2O2 (without STP) decreased the WHC of the llets.
Keywords
Carp, colour, hydrogen peroxide, sodium tri-polyphosphate, surimi, water holding capacity.
Introduction
Surimi is a stabilised minced sh esh, in which,

according to Park & Morrissey (2000), the shelf-life
has been extended by water washing followed by
blending with cryoprotectants to prevent denaturing of
protein. As colour and texture are regarded as the most
important quality parameters of surimi, white-eshed
sh species with high quality textural characteristics,
such as Alaska pollock (Theragra chalcogramma), are
highly valued for surimi manufacturing.
Dark-muscle sh species, which currently make up
4050% of the total sh catch in the world (Hultin &
Kelleher, 2000) and cultured freshwater shes, such as
common carp, could be utilised as raw material for
surimi-based products provided the esh colour could
e-mail: l.gorczyca@rmit.edu.au
be lightened. The normally lower L-value (which is a

measure of lightness) for dark muscle (Jiang et al., 1998;
Wang et al., 2002) is associated with a high concentration of haeme proteins (Lanier, 2000) in blood (as
haemoglobin) and red muscle (as myoglobin) compared
with that present in white sh muscle. According to
Lanier (2000), removal of the coloured heme-containing
[sic] moiety from the muscle during rening depends on
maintaining the heme [sic] proteins in a nearly native or
undenatured state. If the haeme component of a protein
is denatured, the colour of surimi will be darkened due
to binding between haeme and myobril proteins.
Lanier (2000) noted that, as myoglobin is located within
the muscle cells, it is more dicult to leach than
haemoglobin.
Himonides et al. (1999) reported that some researchers have tested vegetable fat-based agents and hydrophilic colloids, such as milk, gum hydrocolloids,
mixtures of sugars, surfactants and fats as sh-esh
doi:10.1111/j.1365-2621.2007.01622.x
Colour improvement of common carp (Cyprinus carpio) fillets by hydrogen peroxide for surimi production A. Jafarpour et al.
whitening agents. Treatments with titanium dioxide,

calcium carbonate, soybean oil have also been proposed
(Benjakul et al., 2004).
Hydrogen peroxide also has been used to whiten sh
esh (Sims et al., 1975; James & McCrudden, 1976;
Young et al., 1979; Raksakulthai et al., 1983; Brown
et al., 1993; Himonides et al., 1999). Sims and his coworkers (1975) reported that H2O2 (600 ppm) whitened
herring esh. Research by James & Mccrudden (1976)
found that 0.85% H2O2 at a pH of 10.5 improved the
colour of cod mince after 15 min exposure at 1520 C.
James & Mccrudden (1976) also claimed that the
addition of STP to a bleaching solution had a synergistic
eect on both colour and texture; provided that the
temperature was kept below 15 C. Brown et al. (1993)
evaluated the bleaching eect of injected H2O2 and STP
over a pH range of 4.510.05 on the dark muscle of
Alaska pollock llets and reported that alkaline pH
values were more eective at improving the colour of
both llets and resultant mince. In a more recent study
by Himonides et al. (1999), H2O2 (58 g L) applied for
1.5 h was found to eectively bleach immersed cod and
haddock aps, without any signicant textural dierences in cooked sh mince produced from untreated and
treated aps.
This investigation was carried out to evaluate the
feasibility of using H2O2 as a bleaching agent to whiten
the light llets of the female common carp, and as a result
be able to produce an acceptable surimi and kamaboko in
terms of colour, comparable to that produced from
Alaska pollock or threadn bream (Nemipterus sp.). The
eect on WHC of llets, surimi and kamaboko as a result
of the application of H2O2 and STP also was investigated
to indicate any changes in texture.
Female common carp (length of 65 2 cm, weight of

4850 175 g) were transported, packed in ice, to
Melbourne directly from the commercial sheries (K &
C Fisheries) usually within 12 h of harvesting from the
Gippsland Lakes (Victoria, Australia). On arrival at the
pilot plant at RMIT University, < 24 h after harvesting, the sh were headed, gutted and lleted to produce
llets of c. 450 g each. Each llet was trimmed to
produce two rectangle portions of approximately
15 10 cm (predominantly dorsal muscle), and ve
sites selected on each portion as ve replicate samples.
For other trials, llets were chopped to c. 2 2 cm.
The following treatment solutions were prepared:
1. pH solutions
Solutions of distilled water (pH 5.5), tap water (pH 6.2) or
pH solutions of 7.0, 8.5, 10.5 or 11.5 (as stock solutions)
prepared from distilled water with sodium carbonate
anhydrous (NaCO3 = 105.99 from AnalaR, Product
No. 10 240, BDH Chemicals [Australia] Pty, Ltd, Princes

Highway, Port Fairy, VIC., 3284).
2. Chemical solutions
Hydrogen peroxide (35% active ingredient, from
SigmaAldrich Pty, Ltd. 12 Anella Avenue, Castle Hill.,
NSW 2154, Australia) at three concentrations (1%, 2%
and 3%) referred to as the bleaching solution.
Sodium tri-polyphosphate (Na5P3O10=367.93, Product No. D3247 from AJAX Chemical Ltd, SydneyMelbourne, Australia) at two concentrations (1% and
2%) with or without added H2O2.
The above-mentioned chemical solutions were added

to pH adjusted solutions as per Table 1.
All treatment solutions were prepared immediately
prior to use, and an aliquot (10 mL) of the appropriate
solution was injected into each of ve sh sites (i.e.
50 mL for ve replicates) to a depth of c. 0.20.5 cm.
Injection sites were spaced at c. 0.5 cm intervals over the
surface of the llet. The injected samples were next
placed into a bag with an appropriate treatment solution
at a ratio of 1:4 (sh weight:solution volume), and after
vacuum sealing, the bags were tumbled at a speed of
4 rpm (tumbler L60 cm, i.d. 35 cm) for c. 2530 min
between 4 and 10 C.
Samples were removed from the treatment solution
and washed in cold (410 C) running tap water and left
in a cold (410 C) water-bath for 1020 min to remove
any residual H2O2, which might be trapped interstitially.
During washing, the water-bath was replenished with
fresh water several times until eervescence of oxygen
was no longer observed. As well as water washing, the
method of Brown et al. (1993) was used to remove
residual H2O2. Washed llets were immersed in a
catalase solution (#C1345, SigmaAldrich) of 0.01%
(w v) at a ratio of 1:4 (sh weight:solution volume).
Table 1 The nal pH of treatment solutions
pH1 of treatment solutions
H2O2
(%)
added
STP
(%)
added
Initial pH adjustment
7.0
8.5
10.5
0
1
2
3
3
3
0
0
0
0
0
0
1
2
1
2
7.0
5.0
4.4
4.4
8.3
8.3
11.2
11.2
8.5
6.4
5.7
5.5
8.3
8.3
11.2
11.3
10.5
8.7
8.4
8.2
8.6
8.4
11.3
11.3
Distilled water
Tap water
11.5
pH 5.5
pH 6.2
11.5
10.5
10.3
10.1
10.1
10.2
11.5
11.5
5.5
NA2
NA
NA
NA
NA
NA
NA
6.2
NA
NA
NA
NA
NA
NA
NA
Distilled water was adjusted with sodium carbonate to produce initial

stock solutions at pH 7.0, 8.5, 10.5 and 11.5, and then further pH modified
by addition of H2O2 (1%, 2% or 3%), H2O2 (3%) plus STP (1% or 2%).
2
Not Applicable.
1603
1604
Residual H2O2 was then determined according to a

procedure (#47.3.22) from the Association of Ocial
Analytical Chemists (AOAC) in which 1020 drops of
vanadium pentoxide solution (1 g V2O5 in 100 mL
H2SO4) was applied directly to the catalase-treated sh
sample. This method, with sensitivity of 0.0006% (w v),
indicated the presence of H2O2, a positive reaction, by
the development of a red or pink colour (Munday, 1957).
Colour evaluation of treated fish fillets
The reected colour of untreated and treated sh

samples was measured using the Minolta Chromo-meter
(CR-100); a tristimulous colour analyzer (Konica
Minolta Sensing, Inc., Sakai, Osaka, Japan). In this
system, the degree of lightness, redness, greenness,
yellowness, or blueness are represented by L*, +a*,
)a*, +b*, )b*, respectively and are reported as L* a*
and b* values (Lanier, 1992). The instrument was
calibrated using a standard white tile (L-value of
92.28, a value of 1.95 and b value of 1.35) placed under
the orice of the instrument. However, before measuring
the colour, the surface of the sh samples was lightly
patted dry with paper toweling. This minimised the
shininess eect, caused by surface moisture, which
would articially increase the lightness and whiteness
value recorded, not as a result of the treatment trialled
but as an artefact of the measurement method. As the
colour of each sample was recorded in duplicate, there
were a total of 10 results (two readings and ve replicate
sh samples) per treatment. The eectiveness of a
treatment to whiten samples was determined by calculating whiteness (Park, 1994; Esturk, 2003; Luo et al.,
2004) and DE as colour deviation (Lanier, 1992), thus:
Whiteness L 3b
DE DL 2 Da 2 Db 2 1=2
and
Water holding capacity
Water holding capacity (WHC) of untreated and treated

llets was determined by the method of Himonides et al.
(1999). For each treatment, three samples (each c.
10.00 0.50 g) were wrapped in individual Whatman
lter paper (No. 41, dia. 7 cm; Whatman International
Ltd, Maidstone, UK) and centrifuged (Beckman, GS15R centrifuge; Beckman Coulter, Inc., Fullerton, CA,
USA) at 1700 g for 30 min at 8 C. The quantity of
water expelled from the sh tissue was estimated from the
weight dierence of the lter paper before and after
centrifugation. The WHC of the sh tissue was determined by the following equation (Himonides et al., 1999):
WHC (g/kg) 1 Mw =Ms 1000
where Mw is the mass (grams) of expelled water and Ms

is the initial mass (grams) of the sample.
Surimi and surimi gel preparation
Frozen surimi from Alaska pollock and threadn

bream was obtained from Austrimi Seafood Co.,
Victoria-Australia, and kept at )20 C. Carp surimi
was prepared from light llets of common carp;
untreated (control) and treated (by injecting, soaking
and tumbling for 30 min in a solution adjusted to pH
10.5 to which 3% H2O2 was added but without STP)
using the following procedure: llets were removed
from the treatment solution and after washing with
running tap water, catalase solution was applied
according to the method of Brown et al. (1993)
as described above. Fillets were then chopped
(c. 2 2 cm) and minced c. 300 g in a food processor
(Moulinex, master-chef 350) for 90 s. During mincing,
temperature was maintained at 410 C by adding ice
water. Once minced to a paste, the cryoprotectants,
sugar, sorbitol and STP, were incorporated to a nal
concentrations of 4%, 4% and 0.3%, respectively into
the paste and homogenizing continued for a further
60 s. Then, the resultant surimi was placed into a bag,
and formed into a block of c. 15 10 2 cm, vacuumsealed and blast frozen ()20 C) until required for gel
preparation.
To prepare the gel, frozen surimi was rst chopped
to smaller pieces and tempered at room temperature (c.
22 C) for about 1.5 h. The surimi pieces were further
reduced in size and adjusted in composition in the food
processor in a two-stage process. In the rst stage, the
surimi pieces were chopped for about 2 min to create a
homogeneous paste, and in the second stage, salt and
ice water were sprinkled over the mince to adjust the
paste to 2% salt and 80% moisture, whilst the
processor continued to homogenize the paste for
another 60 s with the temperature maintained at 4
10 C throughout the process. Then, the paste was
compressed into stainless steel tubes (L 20 cm, i.d.
2.5 cm), which had been sprayed internally with canola
oil to reduce surface stickiness. Both ends of the tubes
were then sealed with screw-thread caps. To obtain low
temperature setting (Lanier, 1992) as required by the
Park and Morrissey method (2000), tubes were refrigerated (4 C) for some 18 h. To convert the surimi to
kamaboko, the tubes were heated in a steam jacketed
kettle set at 90 C 2 C for 30 min to bring the
temperature of the samples to c. 90 C at the geometric
centre. To stop any further eect of heat on texture,
the tubes were immediately transferred to an ice-water
bath. Once cooled (after 1520 min) to 410 C, the
gels were removed from the tubes with a plunger and
sliced to required dimensions for measuring colour and
WHC.
Colour and WHC data were analysed by one-way

analysis of variance (anova) and signicance of dierence between mean values were determined by least
signicant dierence (LSD), and two sample t-test using
a SPSS (SPSS, Inc. Headquarters, 233 S, Wacker Drive,
11th oor, Chicago Illinois 60 606, Version 13.0) statistical package.
Effect of treatment solutions on colour
The eect of water (tap water or distilled water), H2O2

(13% v v), STP (12%) and adjusted pH (7.011.5) on
the colour of light (as this muscle is more pinkish than
white) llets of common carp was evaluated. Untreated
llets of common carp and threadn bream (Fig. 1)
were used as controls in that their L*, a* and b* values
could be used as a baseline measure to determine colour
deviation (DE); an index of colour improvement after
bleaching (Table 2). Threadn bream llets were whiter
than carp llets as the L* value was signicantly
(P < 0.05) greater and b* value signicantly
(P < 0.05) lower than that obtained for carp llets
(Table 2). Usually, L-values were used to indicate
dierences in lightness (Lanier, 1992; Brown et al., 1993;
Park & Morrissey, 2000) and even whiteness (Himonides
et al., 1999), however, a dierence of 4 units in L*
values (Table 2) between threadn bream and common
carp does not convey the dierence in visual whiteness
between these llets. By contrast, the whiteness formula
(L*)3b*) resulted in c. 19 units (Table 2) increase in
whiteness for threadn bream compared with the light
llets of common carp; a dierence that better represented the observed whiteness of these two sh species.
To whiten the llets, Himonides et al. (1999) applied
H2O2 (8 g L) to cod aps for 1.5 h, and reported that,
despite improvements in visual whiteness, there was no
signicant dierence (P > 0.05) between L-values of
cod mince from untreated (L of 47.3) and treated (L of
47.7) aps. The authors stated that the apparent

whiteness was mainly due to a reduction in redness (a
value) and yellowness (b value). In this study it was
found that the whiteness formula (L*)3b*) more
accurately reected visual whiteness than either eqn (2)
or L* alone.
Injecting, soaking and tumbling of common carp
llets with treatment solutions with and without H2O2 at
dierent pH levels improved the colour of llets from
the light muscle of common carp (Fig. 2 and Table 3).
Colour improvement, in the absence of H2O2, occurred
due to washing, and was much the same for the various
pH levels trialled, with no signicant dierences
(P > 0.05) across these treatments. The untreated
control samples had an average L* value of 33.21,
which was signicantly (P < 0.05) lower than that for
llets treated with solutions at pH 7.0, 8.5, 10.5 and 11.5
(Table 3). Injecting, (soaking and tumbling) with distilled water (pH 5.5) or tap water (pH 6.2) produced
very similar results (P > 0.05) to those produced by the
treatment solutions at various pH levels, only (Table 3),
suggesting that the whitening eect was due to washing
of blood from the surface of llets. Removal of
pigments, such as haemoglobin and myoglobin from
the intracellular and intercellular spaces, respectively, is
unlikely to have occurred.
The eect of H2O2 as a bleaching agent signicantly
(P < 0.05) improved the colour of carp llets. The DE
values increased from an average of 7.65 for llets at
various pH levels but not treated with H2O2 to an
average of 12.21, 19.13 and 25.93 for levels of 1%, 2%
and 3% H2O2, respectively. There appeared to be a
negative eect on colour caused by an interaction
between H2O2 and pH at higher levels of each, since DE
decreased (Table 3) signicantly (P < 0.05) for the pH
of 11.5, compared with other pH levels, when H2O2
concentration increased to 2% and also 3%. The strong
oxidation eect of H2O2 lowered the pH of the
treatment solution, thus, the actual pH that is associated with a reduction in DE for carp llets, after H2O2
addition of 2% and 3%, was ca. pH 10.110.3
(Table 1).
1
Figure 1 Untreated common carp (1),
threadn bream (2) llets and surimi from
untreated common carp llets (A), treated
[treatment with 3% H2O2 at a nal pH of
8.2 and no STP (Table 1)] common carp
llets (B), Alaska pollock llets (C) and
threadn bream llets (D), respectively.
1605
Table 2 Colour parameters of untreated

llets of common carp and threadn bream
Tristimulous colour values (SD1)
Species
Visual
assessment
L*
of colour
Common carp
Pink
(control)
Threadfin
White
bream (control)
a*
Whiteness
b*
DE2
33.21 0.74Z )0.48 0.21Z )0.73 0.35Z 34.65 2.33Z 6.25 1.35
37.28 2.04Y )0.15 0.40Y )5.40 0.72Y 53.48 3.25Y
Standard Deviation.
DE represented colour difference between species tested.
Different superscripts in the same column indicated significant difference (P < 0.05) according to a
two sample t-test.
2
pH levels
8.5
10.5
11.5
2%
[H2O2]
1%
0%
7.0
3%
1606
According to James & Mccrudden (1976), who

investigated the eect of H2O2 on cod llets, the most
eective concentration of H2O2 for whitening was
0.85%. They found that higher concentrations of
H2O2 did not produce a signicant improvement in
colour and their optimum pH for whitening at any
temperature or H2O2 concentration was 10.5. They also
indicated that higher pH (11.0) damaged the texture of
the product and resulted in a springy mince. Young and
others (1979) evaluated the eect of various buer
solutions at dierent pH levels (2 to 89 and 10.5) with
mainly 0.75% H2O2 at ambient temperature on either
llets or mince, and the authors associated whitening of
the sh samples with an increase in pigment solubility,
particularly at alkaline pH. Brown et al. (1993) indicated that the highest L-value (lightness) of the dark
Figure 2 Common carp llets treated without and with H2O2 (13%) at dierent pH
levels (7.011.5) by injecting, soaking and
tumbling for 30 min at 410 C.
muscle of Alaska pollock llets was produced by

bleaching solution consisting of 2% H2O2 with 1%
STP at pH 10.5. Himonides et al. (1999) reported no
textural damage to the cod aps treated with 7.5 g L
H2O2 at neutral pH (even after 2 h immersion), but
there was irreversible damage at pH 10.5 possibly as a
result of high pH aecting protein functionality even
after washing the treated aps by cold running tap water
(15 2 C) and immersing in a water bath for about
10 min.
In our experiments, brous and spongy llets were
obtained only after treatment with higher concentrations of H2O2 (23%), and in any case, the texture
recovered after washing with tap water which also
reversed the pH to ca. 6.57.5 from the pH of the
treatment solution. A possible explanation for the
Table 3 Colour deviation (DE) levels for common carp llets
Colour deviation (DE)
Trial
A
Treatment solution
Initial pH2 adjustment
H2O2 (%)
added
STP (%)
added
7.0
0
1
2
3
3
3
3
0
0
0
0
0
0
0
0
1
2
0
1
2
7.09
12.35
20.41
25.17
25.17
13.59
13.79
7.09
1.36
NR
8.5
1.59z
0.70y
0.83x
0.46w
0.46w
2.68u
2.72u
1.59z
0.78y
8.03
12.02
20.78
25.83
25.83
14.03
15.35
8.03
1.50
NR
10.5
1.19z
1.33y
1.83x
1.20w
1.20w
1.50u
2.41u
1.19z
0.54y
7.04
12.10
20.65
29.49
29.49
14.44
14.80
7.04
1.45
NR
11.5
1.75z
2.73y
0.98x
2.60v
2.60v
3.10u
3.94u
1.75z
0.63y
8.51
12.41
14.65
24.23
24.23
10.48
14.69
8.51
1.11
NR
1.38z
0.40y
0.62y
5.50w
5.50w
3.60u
1.60u
1.38z
0.49y
Distilled water
Tap water
pH 5.5
pH 6.2
6.60 1.14z
NR3
NR
NR
NR
NR
NR
NR
NR
NR
7.58 1.63z
NR
NR
NR
NR
NR
NR
NR
NR
NR
1
Fillets treated with distilled water (DW pH 5.5), tap water (TW pH 6.2) or treatment solution without and with H2O2 (13%) and STP (12%) at different pH
levels by injecting, soaking and tumbling for 30 min at 410 C.
2
pH of treatment solutions before H2O2 addition. For actual pH of treatment solutions after H2O2 addition refer to Table 1.
3
Not relevant.
Within each trial, different superscripts in the same column and row indicated significant difference (P < 0.05) according to a one-way ANOVA, and LSD
test.
dierence in the results obtained in this study and that

carried out by Himonides et al. (1999) is that Himonides
used cod aps whereas light llets of common carp were
used in this study; a dierence in both sh species and
tissue tested.
Table 1 shows that addition of 2% or 3% H2O2
modied the pH of the bleaching solutions to range
from 4.4 to just over 10 depending on the initial pH.
While the pH ranges tested were virtually identical, the
whitening eect of 2% and 3% H2O2 was dierent. The
3% H2O2 treatment produced signicantly (P < 0.05)
greater DE values than the 2% H2O2 solution (Table 3).
Sodium tri-polyphosphate had a negative impact on
colour of common carp llets treated with bleaching
solution (H2O2 3%) as DE decreased from ca. 2530 in
the absence of STP to ca. 1014 after STP (1% or 2%)
was added (Table 3). In contrast, James & Mccrudden
(1976) indicated that the addition of 1% STP to
bleaching solution had a synergistic eect on both
whiteness and texture. In support of our study, Brown
et al. (1993) reported that, although the best results for
whitening of Alaska pollock llets came after the
immersion of llets in a bleaching solution of 2%
H2O2 and 1% STP at pH 10.5, this improvement was
not due to the STP. In the absence of H2O2, applying
STP (1%) at adjusted pH levels (7.011.5) signicantly
(P < 0.05) reduced the DE values of the carp llets
compared with llets treated with only adjusted pH
solutions. The DE value decreased from an average of
7.6 to 1.3 with the addition of 1% STP. The inuence of
STP on colour was not signicantly (P > 0.05) aected
by the pH of the treatment solutions trialled (Table 3).
The colour of the resultant surimi from untreated

common carp llets (Fig. 1a) was pinkish, compared
with the white surimi derived from treated (with 3%
H2O2 at initial pH 10.5 and no STP) common carp llets
(Fig. 1b), Alaska pollock surimi (Fig. 1c) and threadn
bream surimi (Fig. 1d). Untreated and treated carp
surimi, Alaska pollock surimi and threadn bream
surimi were next cooked at 90 C 2 C for 30 min to
produce kamaboko, and the colour values of the
kamaboko were measured (Table 4). The only signicant change in colour was a reduction in b* value
(yellowness) which decreased from 0.09 for untreated
carp kamaboko to )5.03 for treated carp kamaboko.
The white appearance of the kamaboko from treated
carp, both visually and quantitatively, was therefore
mainly due to a reduction in yellowness (b*), since
whiteness is based on the dierence between L* and b*
values (specically L*)3b*). This trend is in agreement
with that of Hultin & Kelleher (2000) who reported that
colour improvement of light muscle of mackerel was
primarily due to lower b* values (from 7.2 to 2.0).
Signicantly, the whiteness of the treated carp kamaboko was not dierent (P > 0.05) to that of the Alaska
pollock but was signicantly (P < 0.05) whiter than
threadn bream kamaboko (Table 4).
Like Himonides et al. (1999), this study found that
the whitening eect of H2O2, irrespective of concentration trialled, was supercial, and even injecting the
treatment solution into the llets to improve the
eciency of bleaching did not result in an even
whitening in the centre (depth) of the llets. Penetration
of H2O2 was improved by chopping the llets to ca.
1607
1608
Table 4 Colour parameters of kamaboko

from samples tested
Tristimulous colour values (SD)

Kamaboko from
L*
Untreated common carp surimi

Treated1 common carp surimi
Alaska pollock surimi
Threadfin bream surimi
71.21
70.83
70.25
65.78
a*
)1.70
0.88Z
2.04Z
1.52Y
b*
)2.55
)1.97
)2.51
)1.84
Whiteness
z
0.32
0.29Z
0.65Z
0.23Z
0.09 0.47
)5.03 0.26y
)5.00 0.40y
)4.11 0.57y
70.94
85.92
85.32
78.11
1.64z
0.80y
0.85y
1.04x
1
Treatment with 3% H2O2 at a final pH of 8.2 and no STP (Table 1).
Different superscripts in the same column indicated significant difference (P < 0.05) according to a
one-way ANOVA and LSD test.
2 2 cm to increase the surface area of the sample, thus

producing a more even bleaching eect. This approach
resulted in a colour improvement for the treated
samples which was not signicantly (P > 0.05) dierent
from that obtained by the injection method (data not
presented); however, increasing the surface area (rather
than injecting) and soaking is a far more practical
method of pre-treating the sh llets for surimi manufacturing. Brown et al. (1993) reported that injecting or
immersing the samples with bleaching solutions (4v w)
were equally eective in reducing the colour of dark
muscle sh.
of H2O2 (13%) signicantly (P < 0.05) decreased the

WHC of llets compared with llets treated only at
dierent pH levels without H2O2 (Table 5). However,
no signicant dierence (P > 0.05) in WHC was found
for llets treated with 1%, 2% and 3% of H2O2.
Furthermore, the interaction eect of H2O2 and pH on
WHC was not signicant (P > 0.05).
The combined eect of STP (1% and 2%) and H2O2
(3%) on WHC also was evaluated. Adding STP to the
treatment solution improved the WHC signicantly
(P < 0.05), compared with the same treatment without
STP (Table 5). The positive eect of STP on WHC of
various sh species was reported in other studies (James
& Mccrudden, 1976; Brown et al., 1993). To conrm the
positive eect of STP on WHC, treatment with STP
alone, without H2O2, was carried out. WHC increased
signicantly (P < 0.05) with added STP (1%) but
without any signicant (P < 0.05) dierence within
the pH levels (Table 5).
Effect of treatment solutions on water holding capacity
Increasing the pH of the treatment solution to 11.5

increased the WHC to 832 g kg (Table 5) which was
signicantly greater (P < 0.05) than the WHC value for
llets treated with distilled water (pH 5.5). The addition
Table 5 Water holding capacity (WHC) of common carp llets
Water holding capacity (g kg)
Trial
A
Treatment solution
Initial pH2 adjustment
H2O2 (%)
added
STP (%)
added
7.0
0
1
2
3
3
3
3
0
0
0
0
0
0
0
0
1
2
0
1
2
815.98
805.51
806.07
802.88
802.88
833.71
842.07
815.98
832.28
NR
8.5
5.32y
3.94z
3.47z
0.73z
0.73y
2.36z
3.70w
5.32y
2.00z
820.87
814.62
810.98
804.95
804.95
833.07
840.27
820.87
836.58
NR
10.5
3.63y
3.40w
1.65zw
1.30zv
1.30y
1.38z
4.50w
3.63y
3.54zw
828.60
814.91
813.92
809.88
809.88
836.84
841.77
828.60
838.26
NR
11.5
3.75x
2.58w
4.57wy
6.63ywv
6.63x
3.40z
1.80wz
3.75x
4.00w
832.29
816.31
815.68
811.87
811.87
834.76
847.52
832.29
840.87
NR
3.88x
2.39wy
2.05wy
4.47y
4.47x
2.70z
4.60w
3.88x
0.30w
Distilled water
Tap water
pH 5.5
pH 6.2
805.44 2.65z
NR3
NR
NR
NR
NR
NR
NR
NR
NR
808.62 3.68z
NR
NR
NR
NR
NR
NR
NR
NR
NR
1
After treating fillets with distilled water (DW pH 5.5), tap water (TW pH 6.2) or treatment solution without and with H2O2 (13%) and STP (12%) at
different pH levels by injecting, soaking and tumbling for 30 min at 410 C.
2
pH of treatment solution before H2O2 addition. For actual pH of treatment solutions after H2O2 addition refer to Table 1.
3
Not relevant.
Within each trial, different superscripts in the same column and row indicated significant difference (P < 0.05) according to a one-way ANOVA and LSD
test.
Water holding capacity of the resultant kamaboko

made from common carp llets treated with H2O2 (3%)
at a nal pH of 8.2 (initial pH 10.5), without STP was
measured following the method of Himonides et al.
(1999). WHC of treated carp kamaboko was slightly but
signicantly (P < 0.05) reduced to 874 2.5
(mean standard deviation, n = 3) g kg from
880 3.5, 883 3.2 and 880 2.6 g kg for kamaboko from untreated common carp, Alaska pollack and
threadn bream, respectively, which indicated that H2O2
has a negative eect on the structure of the resultant
kamaboko.
Conclusion
Injecting, soaking and tumbling of light muscle of

common carp with a solution of H2O2 (3%) at an initial
pH of 10.5 (actual pH of 8.2) without STP for around
30 min at 10 C produced the best whitening of the
llets. The colour and WHC of the resultant carp
kamaboko was comparable to kamaboko produced
from Alaska pollock and threadn bream.
Studies should be conducted to evaluate the eect of
bleaching solutions on the quality of the texture,
rheology, and microstructure of llets, and the resultant
surimi and kamaboko.
Acknowledgments
We are grateful to Mr Keith Bell, the Manager of K & C

Fisheries, for supplying carp and taking interest in the
project. Special thanks also to the Managing Director of
Austrimi Co., Mr Shinji Narasaki, for his guidance and
providing commercial surimi samples.
References
Benjakul, S., Visessanguan, W. & Kwalumtharn, Y. (2004). The eect of
whitening agents on the gel-forming ability and whiteness of surimi.
International Journal of Food Science and Technology, 39, 773781.
Brown, P., Rasco, B.A. & Borhan, M. (1993). Colour removal from
the dark muscle of Alaska Pollock (Theragra chalcogramma) llets
and minces using peroxide. Journal of Aquatic Food Products

Esturk, O. (2003). Characterization of rheological properties and
thermal stability of sh myobrillar proteins. PhD thesis: Oregon
State University, p. 147.
Himonides, A.T., Taylor, K.A. & Knowles, M.J. (1999). The improved
whitening of cod and haddock aps using hydrogen peroxide.
Journal of the Science of Food and Agriculture, 79, 845850.
Hultin, H.O. & Kelleher, S.D. (2000). Surimi processing from dark
muscle sh. In: Surimi and Surimi Seafood (edited by J.W. Park).
New York NY: Marcel Dekker Inc, Pp. 5977.
James, A.L. & Mccrudden, J.E. (1976). Whitening of sh with
hydrogen peroxide. In: The Production and Utilisation of Mechanically Recovered Fish Flesh, Aberdeen, UK: Torry Research Station,
Pp. 5455.
Jiang, S.-T., Ho, M.-L., Jiang, S.-H., Lo, L. & Chen, H.-C. (1998).
Colour and quality of mackerel surimi as aected by alkaline
washing and ozonation. Journal of Food Science, 63, 652655.
Lanier, T.C. (1992). Measurement of Surimi Composition and
Functional Properties. In: Surimi Technology. (edited by T.C. Lanier
& C.M. Lee). New York: Marcel Dekker, INC, Pp. 123163.
Lanier, T.C. (2000). Surimi Gelation Chemistry. In: Surimi and Surimi
Seafood. (edited by J.W. Park). New York: Marcel Dekker Inc, Pp.
237265.
Luo, Y.K., Kuwahara, R., Kaneniwa, M., Murata, Y. & Yokoyama,
M. (2004). Eect of soy protein isolate on gel properties of Alaska
pollock and common carp surimi at dierent setting conditions.
Munday, W.H. (1957). Report on Hydrogen Peroxide in Milk. Journal
of AOAC International, 40, 789792.
Park, J.W. (1994). Functional protein additives in surimi gels. Journal
Park, J.W. & Morrissey, M.T. (2000). Manufacturing of surimi from
light muscle sh. In: Surimi and Surimi Seafood (edited by J.W.
Park). New York: Marcel Dekker, Inc, Pp. 2358.
Raksakulthai, N., Aksnes, A. & Njaa, L.R. (1983). Eects of hydrogen
peroxide and of sulphite and humidity on amino acid composition
and digestibility of sh protein. Journal of Science of Food and
Sims, G.G., Cosham, C.E. & Anderson, W.E. (1975). Hydrogen
peroxide bleaching of marinated herring. Journal of Food Technology, 10, 497505.
Wang, X., Fukuda, Y., Chen, S. et al. (2002) Development of an
intermediate foodstu from freshwater sh in China. 9th JIRCAS
International Symposium 2002, Value- Addition to Agricultural
Products.
Young, K.W., Neumann, S.L., Mc Gill, A.S. & Hardy, A. (1979). The
use of dilute solutions of hydrogen peroxide to whiten sh esh. In:
Advances in Fish Science and Technology. (edited by J.C.J. Farnham). Aberdeen, UK: Fishing New Books Ltd, Pp. 243250.
1609
1610
Original article
Use of a toasted durum whole meal in the production
of a traditional Italian pasta: chemical, mechanical,
sensory and image analyses
Antonietta Baiano,1,2* Clara Fares,3 Giorgio Peri,4 Roberto Romaniello,4 Antonella M. Taurino,5 Pietro Siciliano,5
Giuseppe Gambacorta,1,2 Carmela Lamacchia,1,2 Sandra Pati1 & Ennio La Notte1,2
1 Department of Food Science, University of Foggia, Via Napoli, 25 - 71100 Foggia, Italy
2 Istituto per la Ricerca e le Applicazioni Biotecnologiche per la Sicurezza e la Valorizzazione dei Prodotti Tipici e di Qualita`, University of Foggia,
Via Napoli, 25 - 71100 Foggia, Italy
3 CRA ex Istituto Sperimentale per la Cerealicoltura, S.S. 16 km. 675, 71100 Foggia, Italy
4 Department of Production, Engineering and Economics of Agricultural Systems, University of Foggia, Via Napoli, 25 - 71100 Foggia, Italy
5 Microelectronics and Microsystems Institute, CNR, Via per Arnesano, 73100 Lecce, Italy
(Submitted 16 November 2006; Accepted in revised form 22 March 2007)
The characterisation of traditional Italian pasta obtained by mixing amounts of toasted whole meal with remilled semolina and other ingredients was obtained by means of physico-chemical, rheological, mechanical,
sensory and image analyses. The toasted meal showed higher ash, bre and protein contents than re-milled
semolina. The replacement of percentages of re-milled semolina with the toasted meal and soft our
increased tenacity and decreased extensibility and strength, making the dough less suitable for pasta-making.
The P L values were indices of high starch damage. The replacement of part of re-milled semolina and water
with toasted whole wheat meal, soft our and eggs increased the optimal cooking time and the amount of
water absorbed during cooking but made the other cooking parameters worse. The image analysis provided
evidence of the changes induced by the use of toasted wholemeal, soft our and eggs in the microscopic
structure of pasta protein and starch.
Summary
Keywords
Characterisation, cooking quality, durum wheat, image analysis, pasta, toasting, traditional cereal-based product.
Introduction
Pasta is a cheap food representing an important source

of complex carbohydrates, and is obtained by kneading,
extrusion, pressing through a die and eventually dehydration of mixtures of meals and water.
Among meals usable for the production of pasta,
semolina is the favourite one because of its high gluten
content. Gluten is a complex protein network forming
between glutenins (polar) and gliadins (non polar)
during kneading and subsequent resting time (Graveland & Henderson, 1987; Milatovich & Mondelli, 1990).
Quantity and quality of protein and gluten aect pasta
cooking quality, modifying strength, elasticity, rmness,
stickiness and release of organic matter into the cooking
water. High levels of starch damage resulted in an
increase in tenacity and a decrease in extensibility
(Preston et al., 1987). Furthermore, the starch damage
caused by semolina overgrinding and the addition of
*Correspondent: Fax: +39 881 589308;
e-mail: a.baiano@unifg.it
our (Dexter et al., 1983) can aect in a negative way

the pasta cooking quality producing an increase in the
amylose content (Grant et al., 1993) and, in the case of
drying at low temperature, an increase in stickiness.
Also processing and drying are involved in determining
the pasta cooking quality (Novaro et al., 1993; IcardVerniere & Feillet, 1999).
Pasta made from a toasted whole durum wheat meal
is produced in the southern Italy (Foggia and Bari
provinces, Apulia region) and is recognised as traditional by the Italian Laws (Cabinet Decree 18 0 2000)
but is also known abroad. It was traditionally homemade using a whole meal derived from grains of ears
that escaped the harvesting and the burning of the
stubble. Because of the burning, the meal produced from
the grain had a typical brown colour and a roasted,
aromatic avour. This meal was traditionally mixed
with semolina and other ingredients (durum wheat remilled semolina, wheat our and water or eggs) to
produce fresh and dried pasta richer in bre and protein
if compared with the industrial pasta. The health eects
of a diet rich in complex carbohydrate, bre and
doi:10.1111/j.1365-2621.2007.01632.x
Characterization of a toasted durum wheat pasta A. Baiano et al.
antioxidant compounds is well documented in literature

(Cummings et al., 1986; Englyst & MacFarlane, 1986;
Marlett, 1990; Gey, 1993; Gebhardt et al., 2001).
At present, the burning operation has been replaced
by the toasting of the grains previously harvested and
threshed, but no data are available about changes in
nutritional value induced by the toasting treatment.
The objective of this work was to perform a physical,
chemical and sensory characterisation of traditional
Italian pasta samples made of a whole meal deriving
from toasted durum wheat and other ingredients. With
the aim to better clarify the results, image analysis was
used.
Pasta production
Pasta in the form of gnocchetti (Fig. 1) was produced

in a 20-kg plant consisting of a pilot kneading-extruder
(pressure in the kneading chamber )0.8 atm, extrusion
pressure 90100 bar) and a pilot dryer (Afrem International, Lyon, France), according to traditional recipes as
described in the following. Traditional pasta (TP), made

of 68 g per 100 g re-milled semolina and 32 g per 100 g
water; toasted durum wheat pasta (BWP) made of 68 g
per 100 g of a mixture of re-milled semolina and a whole
meal deriving from toasted durum wheat (8:2) and
32 g per 100 g water; toasted durum wheat egg pasta
(BWPE) made of 68 g per 100 g of a mixture of remilled semolina, a whole meal deriving from toasted
durum wheat and commercial soft wheat our (4:4:2),
sodium chloride (0.1% on the dry ingredients) and
32 g per 100 g whole eggs.
Re-milled semolina and toasted durum wheat meal
were derived from the same commercial blend of Italian
durum wheat (Simeto, Ciccio, Arcangelo) and were
supplied by Molino Daddario Antonia (Cerignola,
Foggia, Italy). Re-milled semolina and meal derived
from the toasted durum wheat meal were submitted to
roller milling. The particle size distributions of the used
raw materials are reported in Table 1. Re-milled semolina and toasted durum wheat meal showed a similar
particle size distribution.
The temperature of the added water was 41 C. The
kneading time was 30 min. The size of gnocchetti size
was as follows: length 2.412 0.313 cm, width
1.066 0.152 cm, weight 0.885 0.150 g.
Half of fresh pasta divided into three fractions (named
F-TP, F-BWP and F-BWPE) was stored at )18 C until
analysis whereas the remaining half also divided into
three parts (named D-TP, D-BWP and D-BWPE) was
submitted to drying (1 h at 80 C and then 4 h at 90 C)
in order to investigate the change in the cooking
behaviour induced by drying.
The raw materials and nal products were submitted
to the following analyses.
Analyses of the meals
Figure 1 Picture of fresh gnocchetti samples made of the mixture

of 80% re-milled semolina and 20% whole meal from toasted durum
wheat produced in the pilot plant.
Table 1 Particle size distribution percentage

of re-milled semolina, whole meal from
toasted durum wheat and soft our
- Moisture, ash and protein (nitrogen-to-protein conversion factor 5.70) analyses carried out according to the
AACC methods (2003).
- Total dietary bre (TDF), determined according to the
methods of Prosky et al. (1988) on re-milled semolina
and toasted whole meal, in triplicate. The kit was
supplied by Megazyme (Bray, Ireland).
Particle size
distribution ranges
Re-milled
semolina
Whole meal from

toasted durum wheat
Soft flour
>355 lm
Between 200 and 355 lm
<100 lm
0.05
37.54
14.77
21.40
25.14
0.04
34.02
15.51
23.25
27.78
0.46
6.64
7.67
20.85
62.93
0.00
0.01
0.01
0.01
0.00
0.00
0.43
0.03
0.08
0.48
0.00
0.00
0.03
0.07
0.00
1611
1612
- Gluten index, performed according to the 38-12 AACC

method (1993).
- Water activity measured by means of an Aqualab,
Model Series 3 TE (Decagon, WA, USA).
- Colorimetric measurements, performed through a tristimulus colorimeter 12 (Chromameter-400, Minolta,
Osaka, Japan). Colour was expressed as L*, a* and b*
(brightness, red green balance, rst chromatic coordinate and yellow blue balance, second chromatic coordinate) values, respectively. The colorimeter was
calibrated on a standard white tile (L* = 93.5,
a* = )1.0, b* = 0.8) before each series of measurements.
- Acidity determination, following the 939.05 AOAC
method (1990).
Analyses of the dough

- The dough alveographic indices (P, resistance to
extension or tenacity or elasticity; L, extensibility; W,
deformation energy or our strength) were determined
according to the 54-30 AACC method (1993).
- The farinographic indices (water absorption, stability
index, softening index measured 10 min after the
beginning of the analysis and expressed in Brabender
Unit, dough development time) were obtained according to the 54-21 AACC method (1999).
Analyses of the pasta samples
- Moisture and water activity measurements were carried
out on ground pasta samples.
- Colour was determined as described for ours. The
colorimeter reading head diameter was chosen taking
into account the sample dimensions. Measurements
were performed on the convex side of the samples.
- Optimal cooking time: the gnocchetti were boiled in
distilled water with increasing time and the optimal
cooking time was considered as the time at which the
core disappeared (measured by crushing pasta between
two glass plates).
- Weight and volume increase: culture tubes containing
10 ml of distilled water were equilibrated at
100 0.5 C in a thermostatic bath. Afterwards,
gnocchetti were immersed into the tubes, one for each
tube. A sample were removed from the tube every 30 s,
rapidly blotted and weighed. At a given time, the
volume was determined according to the method of
Olfat et al. (1993) appropriately modied.
- Pasta cooking losses: culture tubes containing distilled
water were equilibrated at 100 0.5 C in a thermo-
static bath. Afterwards, gnocchetti were immersed

into the tubes [the ratio water pasta 20 : 1 (w w) was
chosen in order to avoid water being a limiting factor
for pasta hydration]. At a given time, cooking water
was withdrawn from each tube, dried at 105 C
overnight and weighed, in order to evaluate the
presence of dried matter passed from pasta. The
gnocchetti cooking losses were referred to the water
weight before cooking.
Sensory analysis: dried Gnocchetti samples at the
optimal cooking time were submitted to a panel of ve
trained tasters for the estimation of adhesiveness and
rmness on a 60 (very adhesive, not rm)90 (not
adhesive, very rm) scale. Adhesiveness was evaluated
as the amount of product adhering to teeth after
mastication. Firmness was judged as the force required
to compress the spaghetti between the teeth. Differences among samples were considered signicant if
higher than ve points. Each of the samples was
presented together with a control.
Mechanical analysis: dried gnocchetti samples were
analysed under compression mode by means of an
Instron Universal Testing Machine (model 5567;
Canton, MA, USA, maximum load 999 N). Tests were
carried out at a cross-head speed of 50 mm min)1. Two
peaks were taken into consideration, the rst at a sample
deformation of 30% and the other one at a deformation
of 90% that determined the sample breakdown.
Image acquisition: dried samples of gnocchetti were
scanned using a Meridian ULTIMA Z-laser confocal
microscope system (Genomic Solutions, Lansing, MI,
USA). The acquired images had a 720 lm 720 lm
dimension. The similar samples were also imaged by
means of a scanning electron microscope (SEM), model
jsm6500 (JEOL, Milan, Italy) equipped with a eld
emission gun in a secondary electron conguration at
very low acceleration voltage (3 kV).
Image analysis: the grey-level (0255) digital images
acquired by means of the laser confocal microscope
were saved in a .TIFF format without compression and
successively analysed by means of an algorithm written
in the MATLAB 7 environment (MathWorks, Inc.,
Natick, MA, USA). This algorithm segmented the
images on the base of a threshold method (Gonzalez
et al., 2004), allowing the protein region (which is
autouorescent) and the starch granules to be distinguished. For each of these regions, the algorithm
calculated the surface (as number of pixels) and three
statistical properties of the grey-level histogram: the
mean (or mean grey-level intensity), uniformity and
entropy (intended as the disorder of the grey-levels).

These statistical parameters were determined according
to the following formulae.
Mean
255
X
was made impossible in the toasted whole wheat meal

probably as a consequence of the irreversible changes
induced by toasting in its chemical groups and their
ability to form intermolecular bonds (Lupano & Anon,
1987). It is well known that heat treatments cause
changes in secondary and tertiary structures of proteins
(Michon et al., 1999; Aussenac, 2001; Tilley et al., 2001).
This process of denaturation comes through the formation of disulphide linkages, strong cross-linking bonds
(dityrosine formation) and the exposure of hydrophobic
groups on the surface making the proteins insoluble and
preventing their successive hydration and dough formation.
The soft wheat our was characterised by low ash
content and a medium protein and gluten content. Acidity
was low in all the meals. Concerning the colorimetric
parameters of the meals, the toasting operation reduced
the brightness (L*) (from 88.68 1.32 to 61.08 0.45),
caused a little increase in the red index (from
)1.51 0.37 to 0.90 0.09) and a strong decrease in
the yellow value (from 20.24 0.91 to 12.37 0.14).
Table 3 reports some rheological parameters of
dough. The dough made of 100% re-milled semolina
was more resistant to extension, having a P L ratio
higher than those produced with toasted whole wheat
meal and soft our. Both the 100% re-milled semolina
and 100% soft our dough showed good values of the
parameter W. The formation of a weak gluten network
in the dough made of 100% toasted durum wheat meal
resulted in low W value as a consequence of both the
protein denaturation mentioned above and the presence
of bran and germ particles that physically interfered
with the dough development which is well documented
in literature (Zhang & Moore, 1997; Manthey &
Schorno, 2002). The replacement of 20% of re-milled
semolina with toasted durum wheat meal caused an
increase in tenacity making the dough less suitable for
pasta-making. The alveographic indices were made
worse by the replacement of re-milled semolina with
40% of toasted durum wheat meal and 20% of soft
our. The farinographic indices were not detectable on
the 100% toasted durum wheat meal dough. Remarkable dierences were noticed among all the doughs. The
zi pzi
i0
Uniformity
255
X
p2 zi
i0
Entropy
255
X
pzi log2 pzi
i0
where zi is a generic intensity level (0 value 255)

and p(zi) is the number of pixels in the region having
the zi intensity of grey-level.
Statistical analyses
Analyses were generally repeated at least thrice (ve

times in the case of volume and weight, and ten times in
the case of colorimetric analyses). Mean and standard
deviations were determined. anova analysis, the Tukeys
and Holm tests at a condence level of 95% (P < 0.05)
were performed by means of the Kaleidagraph Statistical Software (ver. 3.6.2; Synergy Software, Reading,
PA, USA).
Table 2 reports physico-chemical characteristics and

composition of re-milled semolina, whole meal from
toasted durum wheat and soft our. Because of the
toasting treatment, the meal from toasted wheat was
characterised by lower moisture and water activity
values than conventional re-milled semolina and, being
a whole meal, by higher ash and bre content. Furthermore, the toasted whole meal also showed higher
protein content, probably because of the milling of the
whole grain and the subsequent integration of germ and
aleuronic proteins (Manthey & Schorno, 2002). The
gluten quantication (gluten content and gluten index)
Table 2 Physico-chemical characteristics of re-milled semolina, whole meal from toasted durum wheat
Flours
Re-milled semolina
Whole meal from toasted
durum wheat
Soft wheat flour
Moisture (%)
Water
activity (aw)
Ash
(% db)
TDF*
(% db)
Proteins
(% db)
Gluten
(% db)
Gluten
index
Acidity (mL NaOH

1 N per 100 g)
8.39 0.11
6.73 0.18
0.438 0.004
0.410 0.002
1.09 0.01
1.56 0.00
2.880.02
9.830.08
11.30 0.00
13.85 0.04
9.27 0.04
n.d.
82.0 1.4
n.d.
2.20 0.25
3.14 0.23
11.50 0.24
0.750 0.10
0.56 0.01
10.8 0.07
9.47 0.03
3.03 0.25
TDF, Total dietary fibre.

n.d., not determinable.
Data are reported as mean value standard deviation.
1613
1614
Table 3 Dough alveographic and farinographic indices

Alveographic indices
Dough
100% re-milled semolina
100% toasted whole wheat meal
100% soft flour
80% re-milled semolina 20% whole
meal from toasted durum wheat
40% re-milled semolina 40% whole meal
from toasted durum wheat 20% soft flour
P
(mm H20)
L (mm)
113
25
81
128
34
23
76
20
137
12
Farinographic indices
Water
absorption (%)
Stability
index (min)
Softening
index (BU)
Dough development
time (min)
3.32
1.09
1.07
6.4
150
28
221
118
66
56.3
61.9
2.5
2.6
71
64
66
1.7
1.6
1.7
11.42
74
60.6
7.4
27
2.5
PL
P, resistance to extension or tenacity or elasticity; L, extensibility; W, deformation energy or flour strength; BU, Brabender Unit.
highest water absorption was detected for the 100% remilled dough. The dough made of the meal mixtures
absorbed lower amounts of water, since the already
discussed protein denaturation. The 100% re-milled
semolina dough showed intermediate values of stability
time and the highest value of maximum consistency
(softening index). Concerning the mixture of re-milled
semolina toasted whole meal soft our, the discussion
of the farinographic indices (response) was made dicult by the concurrent eects of toasting and bre
content. The replacement of semolina with 40% toasted
whole wheat meal and 20% soft our increased the
dough development time (produced by the interference
of bres and the protein denaturation) but reduced the
maximum dough consistency as a consequence of the
formation of the already discussed weaker gluten
network. Therefore, the high stability index detected
in this sample had a limited rheological signicance as
it was referred to a very low value of maximum
consistency.
Table 4 reports some physical characteristics of the
pasta samples. Moisture and water activity were well
correlated (R = 0.956). The replacement of percentages
of re-milled semolina with the toasted whole wheat meal
determined a decrease in moisture and consequently in
water activity. This reduction was particularly evident
also for BWPE samples, in which a greater percentage of
re-milled semolina was replaced with the toasted whole

wheat meal and soft wheat our and water was replaced
with eggs. Concerning colour, the presence of the
toasted whole wheat meal and the drying operations
did not determine changes in brightness whereas a
dierent behaviour was put in evidence by a* and b*. In
particular, the addition of the toasted whole wheat meal
determined an increase in the red value and a decrease in
the yellow index for fresh pasta whereas it reduced the
red value for dried pasta. This nding is in accordance
with Manthey & Schorno (2002) who found an increase
in the reddish-brown surface in spaghetti made from
whole meal.
The toasted whole wheat meal increased optimal
cooking time for both fresh and dried pasta. This
behaviour could be explained with the changes in
protein and starch structure determined by the toasting
operation that are responsible for the delay in hydration
and melting of the starch crystals. This supposition is
consistent with the data reported in Table 3 and is
conrmed by Mohamed et al. (2004). According to them
heat damage may also cause aggregation of wheat
proteins and the loss of gluten extensibility, with a low
gluten water absorption and plasticisation.
As expected, the dried pasta showed a greater increase
in weight than fresh pasta, as a consequence of the
greater water absorption during cooking and overcoo-
Table 4 Moisture, water activity (aw), colorimetric measurements and optimal cooking times (OCT) of the gnocchetti samples
Gnocchetti samples
Moisture (g per 100 g)
aw
F-TP
F-BWP
F-BWPE
D-TP
D-BWP
D-BWPE
24.13
15.02
13.74
6.82
6.63
6.35
0.920
0.820
0.727
0.395
0.392
0.366
1.02
0.25
0.52
0.07
0.09
0.08
L*
0.005
0.020
0.033
0.002
0.008
0.008
47.73
43.37
43.27
44.94
33.58
44.26
a*
2.02
2.83
5.72
2.88
1.94
3.99
0.51
1.87
1.25
2.39
0.74
0.98
b*
0.14
0.58
0.26
0.27
0.11
0.26
17.67
10.82
8.95
18.52
5.34
9.26
OCT (s)
1.13
1.87
1.42
3.42
0.69
0.47
330
630
510
780
840
870

F-TP, fresh traditional pasta; F-BWP, fresh toasted durum wheat pasta; F-BWPE, fresh toasted durum wheat egg pasta.
D-TP, dried traditional pasta; D-BWP, dried toasted durum wheat pasta; D-BWPE, dried toasted durum wheat egg pasta.
(a) 1
0.9
Cooking losses %
0.8
F-TP
F-BWP
F-BWPE
0.7
0.6
a
a
0.5
0.4
a
a
0.3
0.2
a
a
a
a
a,b
a
200
400
a
a
0.1
0
0
600
800
Time (s)
1000
(b) 1
0.9
0.8
Cooking losses %
king. Among the fresh pasta, F-BWPE samples showed

the highest weight increase (probably because of the
same reason that determined a slower hydration and
gelatinisation) following by F-BWP and F-TP samples.
In particular, according to the Tukey test, which allows
comparison of all the possible combinations of sample
pairs, F-BWPE pasta signicantly diered from F-TP
(P < 0.0001) and F-BWP (P < 0.05) whereas no signicant dierences were detected among fresh BWP and
fresh TP. Concerning the dried pasta, D-BWPE showed
the highest weight increase, diering signicantly from
D-TP (P < 0.0002) and D-BWP (P < 0.0049) which
did not dier signicantly from each other.
The increase in volume during cooking and overcooking was about one-third higher for dried samples than
for fresh samples. For each pasta sample, a linear
correlation (correlation coecient, R > 0.94) was
found between the increase in weight and the increase
in volume during cooking and overcooking. The initial
volumes of BWPE and BWP pasta were higher than
those of the TP sample, probably as a consequence of
interferences of the bres with the protein network
architecture. Among fresh samples, and although the
weight increase was equal, the F-TP samples showed the
highest increase in volume, followed by F-BWP and
F-BWPE. These results can be explained by the dierences in the initial volume per unit of weight of the
samples: in the high porosity samples (F-BWP and
F-BWPE), the water absorbed during cooking occupies
the inner pores with a consequent signicant increase in
weight and a negligible increase in volume. On the
contrary, TP samples had a more dense structure and
thus the water absorption caused a remarkable volume
increase. Analogous considerations could be done for
dried pasta even if, the greater porosity induced by the
drying in all the samples reduced the magnitude of these
dierences.
Figure 2 shows the amount of cooking losses as a
function of the cooking time. The amount of solid
substance lost into cooking water represents one of
the indicators of the pasta cooking quality (Baiano
et al., 2006) as it means that an amount of starch is
lost from gluten network during gelatinisation forming
a surface layer that makes the pasta sticky (Feillet,
1988), and increases its tendency to clump. Concerning fresh pasta (Fig. 2a), the F-TP samples always
showed the lowest cooking losses. The use of several
percentages of toasted durum wheat meal determined
greater cooking losses as a consequence of the weaker
gluten network and of the highest bre content. In
fact, bres are known to cause the disruptions of the
gluten matrix (Manthey & Schorno, 2002). The
behaviour of dried pasta was quite dierent (Fig. 2b)
probably as a consequence of the ameliorative eects
of drying resulting in greater protein denaturation and
rmer gluten network (Lamacchia et al., in press) also
b
b
0.6
0.5
0.4
0.3
0.1
0
0
1400
a
D-TP
D-BWP
D-BWPE
0.7
0.2
1200
a
a
a
200
a
a
a
400
a
a a
a b
a
b
b
b
600
800
Time (s)
a
a
b
1000
1200
1400
Figure 2 Pasta cooking losses (weight of organic matter in the cooking

water weight of uncooked pasta)*100): (a) fresh gnocchetti; (b) dried
gnocchetti. F-TP, fresh traditional pasta; F-BWP, fresh toasted
durum wheat pasta; F-BWPE, fresh toasted durum wheat egg pasta.
D-TP, dried traditional pasta; D-BWP, dried toasted durum wheat
pasta; D-BWPE, dried toasted durum wheat egg pasta.
in the formulations containing whole meal from

toasted durum wheat.
The dried gnocchetti samples were submitted to the
sensory analysis. The best judgement was obtained by
the D-BWP and D-TP sample that were considered rm
(80 for both the samples) and not sticky (65 and 60,
respectively). D-BWPE sample showed a low rmness
(60) and a high stickiness (50). The stickiness defect is
due to the amylose not incorporated into the gluten
network which leaks onto the pasta surface This
behaviour could be explained taking into account that,
although the presence of eggs in the dough produced an
increase in protein content and an improvement of the
dough elasticity, the D-BWPE formulation also contained the highest amount of meal from toasted durum
wheat and an amount of soft wheat our that supplied a
high quantity of damaged gluten and starch. In fact, it is
well known that pasta rmness and stickiness depend
not only on protein content but also on both gluten
content and quality (which were impossible to evaluate
in the meal from toasted durum wheat).
1615
1616
Table 5 Mechanical analyses of pasta samples

Gnocchetti
samples
Considered cooking
time (s)
Maximum force at
30% deformation (N)
Maximum force at
90% deformation (N)
Breakdown
energy (N * s)
Elasticity
D-TP
D-BWP
D-BWPE
660
420
540
2.5 0.1
2.8 0.2
2.6 0.2
3.2 0.3
3.8 0.3
2.9 0.1
21.9 4.5
24.4 2.0
10.6 1.3
43.4 3.0
42.9 1.6
34.8 2.3

D-TP, dried traditional pasta; D-BWP, dried toasted durum wheat pasta; D-BWPE, dried toasted durum wheat egg pasta.
The dried gnocchetti was also submitted to mechanical analysis at the same cooking times used for the
sensory analysis. The sample shape was taken into
account in determining the values of the mechanical
indices. The D-TP and D-BWP samples gave the best
results (Table 5), showing the highest maximum force at

90% deformation and energy breakdown (correlated to
the rmness). The elasticity index was calculated as the
ratio between the area of the second half of the peak at
30% deformation and the area of the rst half of the
(a)
(b)
(c)
Figure 3 Microscopic images of gnocchetti

samples acquired by means of the confocal
microscope (on the left) and of the scanning
electron microscope (on the right): (a) dried
traditional pasta (D-TP); (b) dried pasta
made of the mixture of 80% re-milled semolina and 20% whole meal from toasted
durum wheat (D-BWP); (c) dried pasta made
of the mixture of 40% re-milled semolina and
40% whole meal from toasted durum wheat
and 20% soft our and eggs (D-BWPE).
same peak. Both D-TP and D-BWP samples showed a

solid-like behaviour (high elasticity value) whereas the
D-BWPE pasta showed a plastic-like behaviour and,
thus, when submitted to stress, suered permanent
change to its shape.
Figure 3 shows images of inner portions of the dried
gnocchetti samples acquired by means of a confocal
microscope and a SEM. In the dried TP pasta, gluten
forms large strands and includes the starch granules in a
thick lm. The presence of the whole meal from toasted
durum wheat determined a dierent microstructure: the
starch granules (those deriving from the toasted meal
look smaller) are now surrounded by an untidy network
consisting of damaged gluten in which only a small
number of short gluten strands are visible. This texture
was emphasised in the dried BWPE sample because of
the presence of: non-gluten proteins coming from eggs,
gelatinised as a consequence of kneading and extrusion
and stabilised by disulphide bonds during drying
(Alamprese et al., 2005); a high number of short gluten
strands deriving from the great amount of toasted whole
meal and soft wheat our; several damaged starch
granules deriving from the toasted meal and the soft
wheat our as a consequence of toasting and re-milling
operations.
Images acquired by means of the confocal microscope
were analysed and the results for gluten network and
starch granules are reported in Table 6. The replacement
of an amount of re-milled semolina with the mixture of
toasted whole wheat mealsoft our and the addition of
eggs resulted in a detectable increase in the surface of the
uorescent region (represented by proteins) and thus in
a decrease of the starch granule surface. This could be
easily explained by considering that the toasted whole
wheat meal had high protein content and that the
Table 6 Image analysis: surface and statistical properties of the greylevel histogram referred to the uorescent area (protein region) and the
starch granule region
Statistical properties of the
grey-level histogram
Gnocchetti
samples
Surface (%) of
the region
Flourescent area protein region

D-TP
16.97a
D-BWP
30.08b
D-BWPE
44.42c
Starch granule region
D-TP
83.03a
D-BWP
69.92b
D-BWPE
55.58c
Mean
Uniformity
Entropy
78.57a
175.75b
227.08c
0.0174a
0.0273b
0.2724c
6.1808a
6.5128b
4.186c
18.44a
67.76b
67.76c
0.0296a
0.0142b
0.0104c
5.3542a
6.2562b
6.6966c
The different letters indicate statistically significant differences

(P < 0.05).
D-TP, dried traditional pasta; D-BWP, dried toasted durum wheat pasta;
D-BWPE, dried toasted durum wheat egg pasta.
addition of eggs increased the total protein content. This

protein network encapsulated starch, making it less
assessable by microscopy technique detection. The mean
value of the grey-level histogram is an index of the
brightness of the considered region. The autouorescent
region is always brighter than the starch granule region
even if in a dierent amount (+326%, +160% and
+141% for D-TP, D-BWP and D-BWPE, respectively),
as it is well known that increasing protein will decrease
brightness (Chun et al., 2004). The bright intensity
was more uniform in the autouorescent region than in
the starch region for all the considered samples with the
exception of the dried TP pasta. In particular, the
uniformity of the autouorescent region increased when
toasted whole wheat meal and eggs were added. This is
probably because of the formation of the untidy protein
gel in which the gluten strands are plunged. On the
contrary, the uniformity of the starch region decreased
in the BWP and BWPE samples as a consequence of the
change in the starch granules structure induced by
toasting and the presence of starch granules derived
from the soft wheat our. Entropy, which is the opposite
of uniformity, obviously decreased in the autouorescent region because of the contribution of proteins
derived from the toasted meal and eggs and their
increase in the starch region due to the contribution of
starch granules from toasted meal and the soft our.
Conclusions
This study allowed the complete characterisation of

traditional Italian pasta whose main ingredient is
represented by a whole meal deriving from toasted
durum wheat. Raw materials, dough and pasta were
submitted to physico-chemical, rheological, mechanical,
sensory and image analyses. The toasted whole wheat
meal supplied triple the total dietary bre content
compared with the re-milled semolina and higher ash
and protein contents. The replacement of amounts of remilled semolina with the toasted meal or the mixture of
toasted mealsoft our determined a general decrease in
the dough pasta-making properties explainable with the
starch and protein damage induced by toasting. The
contribution of protein from eggs was unable to
counterbalance the toasting eects as these proteins
formed an untidy lm less eective in holding the starch
granules. The addition of toasted whole wheat meal, soft
our and eggs increased the optimal cooking time and
the amount of water absorbed during cooking and,
giving a more porous structure, determined a lower
increase in volume. The replacement of both re-milled
semolina (with the toasted meal or the mixture toasted
mealsoft our) and water (with eggs) strongly worsened the pasta-cooking quality although the defects
were greatly evident in the fresh samples. In fact, it is
well known that drying is able to improve the protein
1617
1618
network stability. The results of the image analysis

underlined the changes in the microscopic structure of
gluten and starch due to the use of toasted meal, soft
our and eggs. The regular gluten network of the pasta
made from re-milled semolina was changed, in the BWP
and BWPE samples, in an untidy protein lm wrapping
starch granules damaged by toasting and re-milling.
In the light of these results, the best formulation was
the mixture of 80% re-milled semolina 20% whole meal
deriving from toasted durum wheat.
References
AACC. (1993). Approved Methods. Method 38-12, 54-30. St. Paul,
MN: The Association.
AACC. (1999). Approved Methods. Method 54-21. St. Paul, MN: The
Association.
AACC. (2003). Approved Methods. Methods 44-15A, 08-01, 46-10, 3025. St. Paul, MN: The Association.
Alamprese, C., Iametti, S., Rossi, M. & Bergonzi, D. (2005). Role of
pasteurization heat treatments on rheological and protein structural
characteristics of fresh egg pasta. European Food Research and
AOAC. (1990). Ofcial Methods of Analysis of AOAC International,
15th Edition. Washington: Association of Ocial Analytical
Chemists.
Aussenac, T., Carceller, J.L. & Kleiber, D. (2001). Changes in SDS
solubility of glutenin polymers during dough mixing and resting.
Cereal Chemistry, 78, 3945.
Baiano, A., Conte, A. & Del Nobile, M.A. (2006). Inuence of drying
temperature on the spaghetti cooking quality. Journal of Food
Engineering, 76, 341347.
Cabinet Decree 18/0/2000 (2000). Ministero delle Politiche Agricole e
Forestali. Ordinary Supplement of the Ofcial Journal of the Italian
Republic n 194. 21st August 2000.
Chun, W., Kovacs, M., Fowler, D.B. & Holley, R. (2004). Eects of
protein content and composition on white noodle making quality:
color. Cereal Chemistry, 81, 777784.
Cummings, J.H., Englyst, H.N. & Wiggins, H.S. (1986). The role
of carbohydrates in lower gut function. Nutrition Review, 44, 5054.
Dexter, J.E., Matsuo, R.R. & Morgan, B.C. (1983). Spaghetti stickiness:
some factors inuencing stickiness and relationship to other cooking
quality characteristics. Journal of Food Science, 48, 15451551.
Englyst, H.N. & MacFarlane, G.T. (1986). Breakdown of resistant
starch and readily digestible starch by human gut bacteria. Journal
of Science and Food Agriculture, 37, 699.
Feillet, P. (1988). Protein and enzyme composition of durum wheat.
In: Durum Wheat: Chemistry and Technology (edited by C. Lintas).
Pp. 93119. St. Paul, MN, USA: AACC Inc.
Gebhardt, E., Dongowwski, G. & Huth, M. (2001). Resistant starch a
component in bifunctional high bre extrudates. Ernaehrung, 25,
202208.
Gey, K.F. (1993). Vitamin E and other essential antioxidants regarding
coronary heart disease: risk assessment studies. In: Vitamin E in
Health and Disease (edited by L. Packer & J. Fuchs). Pp. 589633.
New York: Marcel-Dekker.
Gonzalez, R.C., Woods, R.E. & Eddins, S.L. (2004). Digital Image
Processing Using MATLAB. Upper Saddle River, USA: Pearson
Prentice Hall.
Grant, L.A., Dick, J.W. & Shelton, D.R. (1993). Eects of drying
temperature, starch damage, sprouting, and additives on spaghetti
quality characteristics. Cereal Chemistry, 70, 676684.
Graveland, A. & Henderson, M.H. (1987). Structure and functionality
of gluten proteins. In: Proceedings of the 3rd International Gluten
Workshop, Budapest, Hungary, 9-12 May (edited by R. Lasztity &
F. Bekes). Pp. 238246. Singapore: World Scientic.
Icard-Vernie`re, C. & Feillet, P. (1999). Eects of mixing conditions on
pasta dough development and biochemical changes. Cereal Chemistry, 76, 558565.
Lamacchia, C., Di Luccia, A., Baiano, A., Gambacorta, G., la Gatta,
B., Pati, S. & La Notte, E. (2007). Changes in pasta proteins induced
by drying cycles and their relationships to cooking behaviour.
Journal of Cereal Science, 46, 5863.
Lupano, C.E. & Anon, M.C. (1987). Denaturation of wheat endosperm proteins during drying. Cereal Chemistry, 64, 437442.
Manthey, F.A. & Schorno, A. (2002). Physical and cooking quality of
spaghetti made from wholewheat durum. Cereal Chemistry, 79, 504
510.
Marlett, J.A. (1990). Analysis of dietary ber in human foods. In:
Dietary Fiber: Chemistry, Physiology, and Health Effects (edited by
D. Kritchevsky, C. Boneld & J.W. Anderson). Pp. 3148. New
York: Plenum.
Michon, T., Wang, W., Ferrasson, E. & Gueguen, J. (1999). Wheat
prolamine crosslinking through dityrosine formation catalyzed by
peroxidases: improvement in the modication of a poorly accessible
substrate by indirect catalysis. Biotechnology and Bioengineering, 63,
449458.
Milatovich, L. & Mondelli, G. (1990). La tecnologia della pasta
alimentare. Pinerolo, Italy: Chiriotti editore.
Mohamed, A., Rayas-duarte, P., Gordon, S.H. & Xu, J. (2004).
Estimation of HRW wheat heat damage by DSC, capillary zone
electrophoresis, photoacoustic spectroscopy and rheometry. Food
Novaro, P., DEgidio, M.G., Mariani, B.M. & Nardi, S. (1993).
Combined eect of protein content and high-temperature drying sistems on pasta cooking quality. Cereal Chemistry, 70, 716
719.
Olfat, Y.M., Yaseen, A.A.E. & Aziza, I.A. (1993). Enrichment of
macaroni with cellulose-derivative protein complex from whey and
corn steep liquor. Die Nahrung, 37, 544552.
Preston, K.K., Kilborn, R.H. & Dexter, J.E. (1987). Eects of starch
damage and water absorption on the alveograph properties of
Canadian hard red spring wheats. Journal Association of Ofcial
Analytical Chemists, 20, 7580.
Prosky, L., Asp, N.G., Schweizer, T.F., De Vries, J.W. & Furda, I.
(1988). Determination of insoluble, soluble, and total dietary bre in
food products. Journal Association of Ofcial Analytical Chemists,
71, 10171023.
Tilley, K.A., Benjamin, R.E., Bagorogoza, K.E., Okot-Kotber, B.M.,
Prakash, O. & Kwen, H. (2001). Tyrosine cross-links: molecular
basis of gluten structure and function. Journal of Agriculture and
Zhang, D. & Moore, W.R. (1997). ect of wheat bran particle size on
dough rheological properties. Journal of the Science of Food and
Original article
The role of volatile compounds on aroma and flavour perception
in coloured raw carrot genotypes
Stine Kreutzmann,* Anette K. Thybo, Merete Edelenbos & Lars P. Christensen
Department of Food Science, Faculty of Agricultural Sciences, University of Aarhus, DK-5792 Aarslev, Denmark
(Received 22 January 2007; accepted in revised form 30 July 2007)
Summary
Quantitative analyses of volatile compounds isolated from raw carrots were combined with sensory analysis
in order to identify the role of these compounds on aroma and avour perception in coloured carrots. A
sensory map of carrots with dierent colours was developed, the content of the isolated volatiles was
determined and the role of these compounds for harsh avour perception in raw coloured carrots was
evaluated using multivariate data analysis. The sensory map showed that the coloured carrots formed
distinct groups within the sensory prole. The orange genotypes were characterised by having signicantly
higher intensities in carrot avour and aroma, while the reverse was true for the yellow genotypes. The purple
genotype was characterised by having signicantly higher intensity in sickenly sweet avour and nutty
avour, and the red genotype was characterised by having signicantly higher intensities in green aroma and
avour, bitterness and burning aftertaste. From the multivariate data analysis it was concluded that the
isolated terpenes do correlate to the harsh avour attributes.
Keywords
Aroma compounds, carrot, Daucus carota L, sensory proling, volatiles.
Introduction
In the last decades, consumers have shown an increased

interest in healthy food and an increased demand for
diversity in vegetables (Jongen, 2000). Therefore, quality
is important in marketing of vegetables. In order to be
able to act on changing consumer demands, it is
important to have quantitative means for assessing
quality and quality changes in vegetables. Carrot
genotypes with dierent colours (orange, red, yellow,
purple and white) are now available on the market and
they show a large diversity in quality (Alasalvar et al.,
2001; Kreutzmann et al., 2006). However, there is a
limited knowledge on sensory quality variation between
these genotypes.
The eating quality of carrots can be measured directly
by sensory methods or indirectly by chemical, mechanical or optical measurements. Genetic variation and
environmental conditions largely inuence the sensory
quality of carrots, as quality varies in carrots when
grown under controlled conditions and sorted with
respect to size and shape (Simon et al., 1980; Baardseth
et al., 1996; Hogstad et al., 1997; Seljasen et al., 2001;
Rosenfeld et al., 2002).
*Correspondent: Fax: + 45 89 99 34 95;
e-mail: Stine.Kreutzmann@agrsci.dk
doi:10.1111/j.1365-2621.2007.01662.x
Sensory quality of carrots is dependent on the amount

of volatile compounds and non-volatile bitter taste
compounds and sugars. Bitter and harsh taste in carrots
often causes consumer rejection and is one of the main
reasons for low preference scores when carrots are
evaluated by sensory methods (Alasalvar et al., 2001).
During handling and distribution from eld to consumer, carrots are able to produce sporadic bitterness
and other sensory attributes such as harsh avour
which is caused by stress during growing, harvesting,
transportation, storage and processing (Lafuente et al.,
1996; Seljasen et al., 2000; Talcott et al., 2001). The
sensory characteristic harsh avour was rst introduced by Simon et al. (1980) to describe a strong
burning turpentine-like avour most clearly perceived at
the back of the throat during and after chewing. Harsh
avour seems to be correlated with high content of
volatiles and low content of sugars (Simon et al., 1980).
The majority of volatile constituents emitted from raw
carrots are mono- and sesquiterpenes. These compounds
can comprise up to 97% of the total volatile mass and
have been shown to contribute signicantly to the
aroma and avour of carrots (Simon et al., 1980;
Howard et al., 1995; Shamaila et al., 1996; Alasalvar
et al., 1999; Kjeldsen et al., 2001).
The volatile pattern is largely inuenced by the
genotype. Large genotype variations in carrots include
1619
1620
Aroma and flavour perception in coloured carrots S. Kreutzmann et al.
variations in the total content of mono- and sesquiterpenes and in their qualitative and quantitative distribution. Consequently, sensory quality is inuenced by
genetic variation (Yoo et al., 1997; Kjeldsen et al., 2001;
Ulrich et al., 2003). Simon et al. (1980) emphasised the
importance of both volatile terpenes and sugars for the
avour of raw carrots. Their ndings suggested that
sweetness and overall preference are enhanced by sugars
and reduced by volatile compounds.
The objectives of the present study were: (i) to develop
a sensory map of carrots with dierent colours, (ii) to
determine the content of volatiles in coloured raw
carrots, (iii) to evaluate the role of these compounds for
harsh avour perception in raw carrots dened by
sensory proling using multivariate data analysis and
(iv) to analyse if the sugar level interacts with harsh
avour perception in carrots.
Plant material and sample preparation
Eleven genotypes (Purple Haze, Nottingham, Nebula, Soprano, Nairobi, Mello Yello, Tornado,
Yellow Stone, White Satin, Bolero and Line 1)
were selected to represent a large variation in odour and
taste by sensory screening of twenty-four carrot genotypes. The genotypes were grown in Denmark, Norway
and the Netherlands during 2004 and harvested at the
end of October 2004. The roots were transported to the
Faculty of Agricultural Sciences, Aarslev, Denmark and
stored at 1 C until February 2005 at >95% relative
humidity in an ethylene-free atmosphere. The genotypes
varied in root weight from approximately 50 to 150 g.
The most representative size within each genotype was
selected for analysis. Representative samples (8 kg) of
carrots were taken from each genotype and divided into
subsamples of 1.52.0 kg carrots of rst class quality,
i.e. carrots with no visible damage representing each
replicate. The carrots were then carefully washed,
manually hand-peeled and trimmed. Approximately
0.651.00 mm of the periderm was removed by peeling
and 2 cm of the tip and 2 cm of the top by trimming.
The peeled carrots were cut into 2 2 20-mm sticks
using a food processor (Robot Coupe CL50, Vincennes
Cedex, France), carefully mixed and were taken for
immediately analysis of volatile compounds and sensory
evaluation. All analyses were carried out in three
replicates. The rest of the raw carrots were frozen at
)24 C until analysed for sugars and dry matter
2 months later.
Sensory analysis
Quantitative Descriptive Analysis (QDA) was performed in a sensory evaluation laboratory according
to international standards (ASTM, STP 913 1986). The

panel consisting of ten assessors (ve females ve males,
aged from 26 to 54 years) was screened for sensory
ability (basic taste, odour detection, colour vision) and
ability to communicate sensory descriptions of products
as recommended in ISO 85861: 1993. The panel was
trained to evaluate terpene avour on a test mixture of
a-pinene, b-caryophyllene (Fluka, Steinheim, Germany)
and terpinolene (Carl Roth GmbH, Karlsruhe, Germany) known to have a turpentine-like avour (Burdock, 2002). They were trained in soapiness perception
by introduction of a reference consisting of soft soap. In
an introductory discussion with the panel the assessors
agreed on a consensus list of attributes for sensory
proling and on the denition of each attribute
(Table 1). The fourteen attributes were: terpene aroma,
carrot aroma, green aroma, faded aroma, terpene
avour, carrot avour, green avour, faded avour,
nutty avour, soapiness, sickenly sweet avour, bitterness, sweetness and burning aftertaste. All samples were
served in triplicate in coded glasses with three digit
numbers in random order. Each panellist was served
25 g of shredded carrot at 20 2 C in a 200-mL glass
closed with an odour-free lid to avoid evaporation of
volatiles. The sensory laboratory and the computer
screens were illuminated with red light during evaluation
to mask visual dierences between samples. The panellists evaluated the samples at individual speed by
descriptive analysis on an unstructured 15-point line
Table 1 List of sensory attributes of carrots with descriptions

Characteristics
Odour
Carrot aroma
Terpene aroma
Green aroma
Faded aroma
Flavour
Carrot flavour
Terpene flavour
Sickenly sweet
flavour
Green flavour
Nutty flavour
Soapiness
Faded flavour
Taste
Sweetness
Bitterness
Aftertaste
Burning
aftertaste
Description
Relate to positive carrot odour

Odour related to mixtures of turpentine-like
flavour (a-pinene, b-caryophyllene, terpinolene)
Smells like green carrot leaves
Relate to faded hay odour
Like the flavour of carrot
Flavour related to mixtures of turpentine-like
flavour
(a-pinene, b-caryophyllene, terpinolene)
Unpleasant, overly sweet flavour
Like the flavour of green carrot leaves
Flavour of fresh hazelnut (green)
Like the flavour of soft soap
Like the flavour of faded hay
Taste related to the taste of sucrose
Sharp taste related to the taste of caffeine
Related to sharp, burning taste in the mouth
after 60 s
scale with intensity ratings ranging from low (value 0) to

high intensity (value 15). All data were registered on a
direct computerised registration system (FIZZ, ver.
2.00M; Biosystemes, Couternon, France).
obtained from Fluka Chemie GmbH (Buchs, Switzerland), and from Sigma-Aldrich Chemie GmbH (Steinheim, Germany).
Extraction and quantification of sugars
Collection and analysis of volatile compounds
Volatile compounds were collected from 50 g fresh-cut

carrots by dynamic headspace sampling according to the
method described by Kjeldsen et al. (2001).
Volatile compounds were quantied by gas chromatography (GC) and identied by GC-mass spectrometry
(MS). For GC analysis a Hewlett-Packard 5890 Series II
Plus gas chromatograph (Hewlett-Packard, Avondale,
PA, USA) equipped with a split splitless injector
(200 C) and a ame ionisation detector (FID) operating at 230 C was used. Volatiles were separated on a
Chrompack (Middleburg, The Netherlands) WCOT
fused silica capillary column (50 m 0.25 mm i.d.,
DF = 0.2 lm liquid phase: CP-Wax 52CB). Helium
was applied as a carrier gas with a ow rate of
1.4 mL min)1 and 22 psi column head pressure. One
microlitre of the sample was injected in splitless mode.
The oven temperature was kept at 31 C for 1 min,
programmed to increase to 80 C at 1.5 C min)1, from
80 C to 125 C at 1 C min)1 and further to 190 C at
18 C min)1, followed by a constant temperature for
10 min. The individual volatiles were tentatively quantied from the FID peak areas relative to that of the
internal standard [(E)-2-hexen-1-ol]. The response factor
was set to 1 for all compounds.
GCMS analyses were performed on a Thermo
Finnigan Trace dual-stage quadrupole (DSQ) single
quadrupole mass spectrometer operated at 70 eV and
directly coupled to a trace GC ultra gas chromatograph
equipped with a split splitless programmed temperature
vaporizing (PTV) injector (split ow 10 mL min)1,
200 C) and a CP-Wax 52CB column were used. Helium
was the carrier gas at a ow rate of 1.4 mL min)1 and 22
psi column head pressure. GCMS was performed on
1 lL of concentrated samples. The oven temperature
programme was as described before for the GC analysis.
The temperature of the column and the transfer line was
200 C. The MS was operated in scan mode over a mass
range from 39 to 650 amu. Compounds suggested by the
MS database (NIST., 1998) were veried by comparison
of the relative retention indices (RI) and MS of
authentic reference compounds unless noted. RI of the
isolated components was determined externally with a
series of n-alkanes (C10C25) as described by Kjeldsen
et al. (2001, 2003). a-thujene and a- and b-farnesenes
were obtained from Wako Chemicals Ltd. (Tokyo,
Japan) and (E)-c-bisabolene from TCI Tokyo Organic
Chemicals Ltd. (Tokyo, Japan). ()-b-bisabolene was
synthesised as described previously (Kjeldsen et al.,
2001). Other authentic reference compounds were
Samples of 75 g frozen carrots were homogenised for

4 min with 75 g ultra pure water. A 2-g subsample was
diluted to 50 mL using ultra pure water, kept at room
temperature for 3 h and ltered through a 0.45-lm
AcetatePlus Cameo lter (Bie and Berntsen, Rdovre,
Denmark). Samples for analysis (3 mL) were diluted to
50 mL with ultra pure water and analysed by analytical
high-performance anion exchange chromatography
(HPAEC) according to the method of Kaack et al.
(2004). Quantication was performed on a Dionex Series
300DX ion chromatograph (Dionex Denmark, Rdovre,
Denmark) coupled with pulsed amperometric detection
(PAD) and equipped with a borate trap (4 50 mm), an
amino trap (4 50 mm) and a CarboPac PA10
(4 250 mm) analytical column. Separations were performed by gradient elution with solvent A (200 mm
NaOH) and solvent B (water). The elution prole was
0 min (12% A), 7.0 min (15% A), 13.0 min (30% A),
24.034.8 min (100% A) and 35.0 min (12% A).
Dry matter
Dry matter was determined gravimetrically by weighing

before and after drying in a ventilated oven at 80 C for
20 h (Lytzen A S, Herlev, Denmark).
The chemical and the sensory data were analysed by the

General Linear Model (GLM) or the MIXED procedures (SAS version 8e; SAS Institute, Cary, NC, USA)
and by multivariate data analysis (Principal Component
Analysis, PCA) using Unscrambler (Windows version
8.0 software package; CAMO A S, Trondheim, Norway). The sources of variances were genotypes. In the
sensory analysis least signicant dierence at P 0.05
was determined taking the combined residual structure
into account. All experiments were carried out in
triplicate on subsamples from each genotype. For PCA
all data were standardised, and complete cross-validation was used.
Sensory analysis
The sensory proles varied signicantly between the

carrot genotypes. Mean values and results from the
statistical analysis on the sensory data are summarised
in Table 2. Except for soapiness and faded avour, the
1621
Table 2 Mean values for the eleven coloured carrot genotypes

Colour
Purple
Orange
Genotype
Purple
Haze
Nottingham
Nebula
Soprano
Nairobi
Tornado
Carrot aroma
Terpene aroma
Green aroma
Faded aroma
Carrot flavour
Terpene flavour
Sickenly sweet flavour
Green flavour
Nutty flavour
Soapiness
Faded flavour
Sweetness
Bitterness
Burning aftertaste
6.6
4.8
1.7
2.9
8.0
4.8
4.4
2.1
4.4
1.3
1.7
9.0
2.1
2.8
8.7
7.2
3.3
1.2
10.8
7.8
1.6
4.9
3.5
1.7
0.8
9.6
2.9
3.1
8.5
7.6
3.1
2.3
9.0
6.6
2.1
4.0
3.0
1.3
1.6
6.5
2.9
2.6
9.2
8.8
3.4
1.8
7.9
7.1
1.8
5.5
2.8
1.4
1.9
5.9
2.6
3.5
9.7
10.4
4.4
0.8
10.0
8.4
1.6
6.5
2.8
1.3
1.4
6.3
5.1
4.1
9.5
7.5
3.4
1.8
11.3
7.8
1.7
4.7
3.4
1.1
0.8
10.8
2.7
3.0
Red
Yellow
White
Bolero
Line 1
Mello
Yello
Yellow
Stone
White
Satin
LSD
8.7
9.6
4.6
2.4
11.0
9.1
1.9
6.0
3.0
1.8
1.4
9.3
4.0
4.4
10.4
9.0
5.8
1.0
9.3
10.0
1.2
8.9
1.5
2.1
1.9
4.1
10.4
5.5
8.7
7.4
3.9
2.3
7.8
6.6
1.5
7.2
2.3
2.2
1.4
5.9
5.3
4.3
7.0
6.5
3.2
2.2
6.5
6.6
2.8
5.6
2.5
2.5
2.1
5.1
3.4
3.7
8.8
9.1
3.3
0.8
9.3
8.6
1.2
6.9
2.7
1.4
1.4
6.4
5.7
3.9
1.5
1.9
1.4
1.4
1.4
1.6
1.4
2.1
1.3
NS
NS
1.5
1.5
1.6
Least significant difference (LSD) at P 0.05 is used to assess the significant differences. NS, non-significant.
eect of genotype was signicant for all attributes,

caused by the fact that the genotypes spun as much of
the variation in the plant material as possible. Gills et al.
(1999) and Martens et al. (1985) reported limited
dierences between avour attributes for orange carrot
genotypes while Simon et al. (1982) and Seljasen et al.
(2001) found that genotype largely inuenced the
variation in sensory attributes of orange carrots. In
contrast, Rosenfeld et al. (1997) found that the sensory
variation mainly was related to the variation in sugarand texture-related attributes more than related to a
genetic variation.
PCA was applied on the fourteen attributes and the
eleven genotypes to investigate the structural variation
Yellow stone
1000
in the data. PCA was performed on mean values and a

bi-plot (PC1 vs. PC2) is visualised in Fig. 1, showing the
rst two PC to explain 56% and 26% of the variation.
The attributes faded aroma, sickenly sweet avour and
nutty avour were positively correlated and inversely
correlated to burning aftertaste, bitterness, terpene
avour and aroma, green avour and aroma and carrot
aroma. Thus, genotypes placed on the right side of the
bi-plot were carrots with relative high intensity of
bitterness, burning aftertaste, terpene aroma and avour, green aroma and avour, carrot aroma and
avour. Inversely, genotypes placed on the left side
had relatively high intensity in sweetness, sickenly sweet
avour, nutty avour and faded aroma. In conclusion
Faded f
Soapiness f
500
PC2 (26%)
1622
Faded a
Mello yello
Sickenly sweet f
Line1
Burning aftert Bitterness
Purple haze
Soprano
Green f
Green a
White satin
Nairobi
Bolero Terpene f
Carrot a Terpene a
Nebula
Nutty f
500
Nottingham
Sweetness
Tornado
Carrot f
1000
1000
500
500
PC1 (56%)
1000
Figure 1 Sensory mapping of eleven carrot

genotypes and fourteen sensory attributes by
Principal Component Analysis (PCA bi-plot,
PC1 vs. PC2). ( ), genotypes; ( ), sensory
attributes (a, aroma; f, avour).
the bi-plot shows that the red genotype (Line 1) was

characterised by having signicantly higher intensities in
carrot aroma, green aroma and avour, terpene avour,
bitterness and burning aftertaste, while signicantly
lower intensity was found in sickenly sweet avour and
nutty avour. The purple genotype (Purple Haze) was
characterised by having signicantly higher intensity in
sickenly sweet avour and nutty avour, while signicantly lower intensity was found in terpene avour, green
avour, bitterness and burning aftertaste. The yellow
genotypes (Mello Yello, Yellow Stone) were characterised by having signicantly lower intensities in carrot
aroma and avour, and sweetness, while the orange
genotypes (Nottingham, Tornado, Bolero, Nairobi)
were characterised by having signicantly higher intensities in carrot avour and aroma and sweetness (Fig. 1,
Table 2). According to Table 2, the most bitter genotype
was the red genotype Line 1 (average score 10.4), and
the sweetest genotypes were the orange genotypes
Tornado (10.8), Nottingham (9.6), and Bolero (9.3)
and the purple genotype Purple Haze (9.0).
Previously, Kreutzmann et al. (2006) analysed the
overall sensory prole of coloured carrot genotypes and
showed that taste and colour are highly correlated
indicating that genes coding for taste and colour are
related. Alasalvar et al. (2001) examined four dierent
coloured carrots for sensory quality. They found significant dierences among the colour groups for a carrot
top attribute resembling petrol and sweetness with the
purple carrot having the highest score, and that the
orange and the white genotypes had higher levels of
terpenes than the other genotypes. In contrast to our
observation they found no signicant dierence among
the colour groups for bitterness and aftertaste.
Volatile compounds in coloured carrots
A total of thirty compounds were detected repeatedly

and quantied in the carrot headspace extracts
(Table 3). Twenty-ve of these volatiles were identied
by comparison of their MS and RI data with those of
authentic reference compounds. The remaining volatiles
were tentatively identied by comparison of their MS
data with those of the NIST database (NIST., 1998).
Mono- and sesquiterpenes accounted for >99% of the
total volatile mass in the investigated carrots and most
of the volatiles have previously been isolated from
carrots (Buttery et al., 1979; Alasalvar et al., 1999, 2001;
Kjeldsen et al., 2001, 2003; Seljasen et al., 2001; Varming et al., 2004). However, estragole and a-chamigrene
have not been reported by these authors. Our study
showed that the major monoterpenes were a-pinene,
sabinene, b-pinene, b-myrcene, c-terpinene, terpinolene
and limonene. The major sesquiterpenes were b-caryophyllene, a-humulene and (E)-c-bisabolene. These
ndings agree well with previous studies on volatiles in
raw carrots (Alasalvar et al., 1999; Kjeldsen et al., 2001,

2003; Varming et al., 2004).
Results in Table 3 show large genotypic dierences
between coloured carrots in their contents of volatile
compounds. Total volatile mass ranged from
1838 ng g)1 fresh weight to 7683 ng g)1 fresh weight,
being highest in Bolero and lowest in Purple Haze.
Simon et al. (1982) found terpinolene to be the most
abundant terpene in carrots and concluded that terpinolene eected harsh avour, sweetness and preference.
Terpinolene was also the most abundant terpene in our
experiment. However, terpinolene was far more abundant in red and orange carrots than in white and yellow
carrots. b-caryophyllene and (E)-c-bisabolene were also
present in high amounts (Table 3). The large dierences
in total volatiles are in accordance with previous studies
(Simon et al., 1982; Alasalvar et al., 1999; Kjeldsen
et al., 2001; Toth-Markus & Takacs-Hajos, 2001).
Contribution of volatiles to harsh flavour in coloured
carrots
The PCA gives an overview on how the variation in

sensory quality and chemical composition are inuenced
by the colour of the genotypes. The results are visualised
by means of a bi-plot (Fig. 2), showing the rst two PC
explains 50% of the variation. A large proportion of the
variation in the sensory attributes and volatiles are
found in PC1 (35%), whereas PC2 (15%) mainly
explains a chemical variation in dry matter content,
sucrose, glucose and fructose (Table 4 shows contents of
dry matter and detailed sugar composition). The sensory
variation was mainly described by the variables, carrot
aroma, terpene aroma, burning aftertaste, green avour
and aroma, soapiness and bitterness versus faded
aroma, sweetness and nutty avour.
Terpenes generally cause a harsh or bitter avour
sensation in carrots (Simon et al., 1980; Seljasen et al.,
2000; Rosenfeld et al., 2002). In the present study most
of the harsh avour attributes (terpene avour, green
avour, bitterness and burning aftertaste) showed
increasing values with increasing terpene content
(Fig. 2). High relations were found between the sensory
attributes terpene avour, burning aftertaste and bitterness and the levels of p-cymene, c-terpinene, limonene,
(E)-b-farnesene, a-phellandrene and terpinolene. Terpene aroma, green aroma and avour and carrot aroma
were positively related with b-myrcene, a-terpinene and
b-phellandrene.
The odour descriptions in Table 3 are taken from
Kjeldsen et al. (2003). This and other studies show that
the terpenoides and especially the monoterpenes, are
considered to be the most important volatile compounds
(Howard et al., 1995; Shamaila et al., 1996; Kjeldsen
et al., 2003). Kjeldsen et al. (2003) found that the
volatile compounds with an odour sensation could be
1623
1624
Table 3 Odour description and concentration (ng g
)1
fresh weight) of volatile compounds isolated from coloured carrots by dynamic headspace
sampling
Colour
Orange
Volatile compounds
a-pinene
a-thujene
camphene
b-pinene
Sabinene
a-phellandrene
b-myrcene
a-terpinene
Limonene
b-phellandrene
(Z)-b-ocimene
c-terpinene
(E)-b-ocimene
p-cymene
Terpinolene
(E,E)-2,4-heptadienalb
Camphor
b-caryophyllene
(Z)-b-farnesene
a-humulene
p-allylanisole (estragole)
(E)-b-farnesene
c-cadineneb
a-chamigreneb
b-bisabolene
(E)-c-bisabolene
Cuparene
c-elemeneb
Caryophyllene oxide isomerb
()-caryophyllene oxide
Total
RI
Odour descriptions
1008
1010
1044
1086
1105
1147
1153
1162
1183
1191
1226
1230
1241
1252
1266
1494
1508
1576
1632
1640
1648
1650
1671
1704
1708
1737
1776
1816
1951
1969
Pine, carrot top
Carrot top, fresh green

Carrot-like, fresh green
Herbaceous, green, carrot top
Green, terpene-like
Sweet, citrus, fruity
Herbaceous, citrus, fruity

Carrot top
Sweet, fruity, citrus
Terpene-like, spicy, woody

Woody
Sweet
Soapy, spicy
Citrus
Yellow
Purple
Min
Max
Red
Min
Max
White
CV (%)d
107
2.1
1.0
17.0
3.9
0.4
11.2
0.7
5.2
2.5
3.9
78.9
3.8
19.3
37.8
NQd
1.2
727
0.6
35.3
NQe
13.6
10.1
9.3
43.0
671
24.4
0.3
4.5
3.0
1838
15.2
1.9
1.4
23.3
2.2
0.7
43.5
0.9
7.7
2.0
0.4
99.5
2.9
23.5
94.6
NQ
3.3
178
0.6
9.6
NQ
28.9
13.3
11.1
31.8
370
1.1
0.3
3.1
0.8
1920
145
20.9
5.3
128
199
25.6
87.7
12.8
139
6.7
44.3
760
8.5
96.8
2319
51.6
9.9
3229
10.8
167
11.1
60.2
60.3
17.6
1078
1384
86.5
3.2
27.9
11.5
7683
214
6.1
15.8
496
15.1
24.4
116
5.8
163
8.7
77.6
625
32.8
121
2634
1.2
23.5
284
1.7
16.2
2.0
83.2
2.8
12.1
54.4
768
23.4
0.5
1.4
1.4
5829
19.6
13.3
0.9
49.0
212
3.9
101
6.2
22.8
5.2
0.9
154
7.1
35.9
268
0.6
5.8
254
0.4
17.7
2.3
28.5
41.0
14.4
64.6
931
9.7
0.8
2.5
1.4
2312
37.8
24.3
1.9
52.7
405
7.0
173
9.7
48.9
9.4
5.1
303
8.7
64.0
580
2.1
9.8
266
1.2
53.4
2.9
32.6
140
27.3
82.1
1299
42.5
1.0
14.5
2.2
3666
73.3
18.1
2.4
59.4
241
10.5
33.1
7.3
60.2
4.6
0.8
253
0.4
38.4
850
0.7
3.2
1463
3.1
87.2
1.3
31.6
13.8
4.5
189
15.2
55.0
0.5
2.6
7.8
3530
34
45
25
23
43
33
43
29
26
28
41
29
35
26
29
24
28
42
41
35
32
29
38
26
23
24
69
49
42
51
Mass spectra (MS) and gas chromatography (GC) retention indices (RI) were consistent with those of the reference compounds unless noted.
Tentatively identified. No standard available but the MS is consistent with published data in the MS database (NIST., 1998).
c
Odour described by GC-olfactometry by Kjeldsen et al. (2001).
d
Mean coefficient of variance (CV) for three replicates of each genotype.
e
NQ, not quantified (less than 0.1 ng g)1).
b
divided into three distinct odour groups based on their

GColfactometry (GCO) description: carrot top,
fruity and spicy-woody. They found that the characteristic carrot top odour was related to a-pinene and
b-pinene, sabinene, a-phellandrene, b-myrcene and
p-cymene. The monoterpenes limonene, c-terpinene,
and terpinolene had citrus-like, and fruity, notes,
while the major sesquiterpenes, b-caryophyllene,
a-humulene, b-bisabolene and (E)- and (Z)-c-bisabolene
were found to have spicy and woody notes (Kjeldsen
et al., 2003). Compared with the PCA analysis, where
avour sensation for individual terpenes is impossible to
predict, it can be concluded that the isolated terpenes do
relate to the harsh avour attributes. However, identication of the terpenes that contribute to carrot top,
fruity or spicy-woody sensations cannot be deduced
from the present results. Clearly further investigations
are needed in order to correlate specic sensations to
individual terpenes, or if the dierent sensations are
because of synergistic eects among terpenes.
Interaction of sugar level with harsh flavour perception
of the terpenes
Total sugar content was positively related with sweetness and sickenly sweet avour, whereas sucrose and
Glucose
Mello yello
Fructose
Gamma-cadinene
Sabinene
Alfa-chamigrene
Caryophyllene oxide Nairobi Alfa-thujene
Beta-myrcene
Gamma-bisabolene
Terpene a
Gamma-elemene
Green f
White satin
Nebula
Soap f
Carrot a
Soprano
Green
a
Nottingham
Burning aftert
Yellow stone
Terpene f
Tornado
Bolero
Carrot f
Sweetness faded a
Bitterness
Faded f
Nutty f
Total sugar Cuparene
Beta-pinene
Sickenly sweet f
Alfa-pinene
Dry matter
Line1
Sucrose
Alfa-humulene
Purple haze
Beta-caryophyllene
1000
PC2 (15%)
500
Figure 2 Variation in the eleven carrot genotypes, chemical content and sensory attributes
analysed by Principal Component Analysis
(PCA bi-plot, PC1 vs. PC2). The circles represent groups of correlated volatiles [p-allylanosole, (E,E)-2,4-heptadienal, (Z)-bfarnesene, camphor, b-bisabolene, a-terpinene, b-phellandrene, p-cymene, c-terpinene,
limonene, terpinolene, a-phellandrene,
(E)- and (Z)-b-ocimene, (E)-b-farnesene, and
camphene]. ( ), genotypes; ( ), chemical
compounds and sensory attributes (a, aroma;
f, avour).
500
1000
Caryophyllene oxide isomere
1000
Genotype
Purple Purple Haze

Red
Line 1
Orange Nottingham
Nebula
Soprano
Nairobi
Tornado
Bolero
Yellow Yellow Stone
Mello yello
White
White Satin
Dry matter
Glucose
Fructose
Sucrose
14.58
10.91
11.38
11.51
10.45
10.80
11.57
11.83
11.22
9.70
9.11
0.89
0.94
1.87
1.82
1.65
2.06
1.71
1.96
1.45
2.49
1.60
0.83
0.49
1.65
1.64
1.52
1.93
1.62
1.75
1.27
1.69
1.49
7.60
4.60
4.60
3.83
3.60
2.88
4.28
4.44
3.70
2.79
3.26
(0.26)
(0.27)
(0.54)
(0.52)
(0.34)
(0.28)
(0.08)
(0.34)
(0.55)
(0.90)
(0.39)
(0.36)
(0.12)
(0.50)
(0.19)
(0.26)
(0.38)
(0.24)
(0.34)
(0.35)
(0.36)
(0.16)
(0.25)
(0.15)
(0.13)
(0.20)
(0.08)
(0.11)
(0.15)
(0.04)
(0.19)
(0.07)
(0.28)
(1.17)
(0.46)
(0.61)
(0.10)
(0.16)
(0.14)
(0.14)
(0.25)
(0.05)
(0.36)
(0.26)
total sugar content were negatively related to volatile

terpenes and terpene avour, green avour, bitterness
and burning aftertaste that express harsh avour
perception in carrots (Fig. 2). To test the hypothesis
that sugar level interacts with the perception of harsh
avour, the carrot genotypes were compared pair-wise
within nearly the same terpene level (Table 5).
Purple Haze and Soprano had similar terpene levels
(1838 and 1920 ng g)1 fresh weight, respectively) but
dissimilar sugar levels (9.32 and 6.77 g per 100 g fresh
weight, respectively). In agreement with this, Purple
Haze scored highest in sweetness, and it seemed that the
high sugar content masked some of the terpene avour
(harsh avour), as Soprano scored highest in terpene
avour, although the terpene content was similar.
Bolero and Line 1 are the two genotypes with the
highest terpene content (7683 and 5829 ng g)1 fresh
weight, respectively). Bolero has a higher sugar content
compared with Line 1 (8.14 and 6.04 g per 100 g fresh
500
1000
PC1 (35%)
Table 4 Dry matter (%) and sugar content (g per 100 g fresh weight) in
coloured carrot genotypes. Mean values and standard deviations
Colour
500
weight, respectively) and scored higher in sweetness.

Although Bolero has the highest terpene content, the
genotype obtained a lower score in terpene avour,
which again could indicate some masking of this
sensation because of the high sugar content. The
genotypes Nebula and Nairobi have similar terpene
content (4468 and 4282 ng g)1 fresh weight, respectively), sugar content (7.29 and 6.87 mg per 100 g fresh
weight, respectively) and sweetness scores (6.52 and
6.33, respectively). However, the genotypes have dissimilar scores in bitterness and terpene avour, which
indicate that substances other than terpenes such as
phenolic acids and or polyacetylenes may contribute to
the bitterness of the carrots (Alasalvar et al., 2001;
Czepa & Hofmann, 2003) and interact in the perception
of harsh avour. Rosenfeld et al. (2003) found that the
root tip was perceived sweeter than the crown although
the tip contained less sugar, and terpenes than the
crown. In carrot juice they also found an increase in
sweetness and a decrease in bitterness and terpene
content at a similar sugar level. These observations
indicate that interactions between sugar and volatiles
aect the sensory perception and are therefore a highly
relevant issue for consumer perception of coloured
carrots.
Conclusion
A sensory map was developed for coloured raw carrots.

The map described the sensory variation among purple,
orange, red, yellow and white genotypes of carrots in
terms of fourteen odour and taste attributes. The results
showed that the coloured raw carrots formed distinct
groups within the sensory prole. Diversity among the
raw material is to be visualised by a sensory mapping of
1625
1626
Colour
Genotype
Terpenes
(ng g)1 fresh
weight)
Purple
Orange
Orange
Red
Orange
Orange
Orange
Yellow
Orange
Yellow
White
Purple Haze
Soprano
Bolero
Line 1
Nebula
Nairobi
Nottingham
Yellow Stone
Tornado
Mello Yello
White Satin
1838
1920
7683
5829
4468
4282
2445
2312
3439
3666
3530
(477)
(121)
(1800)
(1334)
(1216)
(488)
(344)
(517)
(137)
(697)
(416)
Sugars (g per
100 g fresh
weight)
Terpene
flavour
Sweetness
Bitterness
9.32
6.77
8.14
6.04
7.29
6.87
8.12
6.41
7.61
6.97
6.35
4.8
7.1
9.1
10.0
6.6
8.4
7.8
6.6
7.8
6.6
8.6
9.0
5.9
9.3
4.1
6.5
6.3
9.6
5.1
10.8
5.9
6.4
2.1
2.6
4.0
10.4
2.9
5.1
2.9
3.4
2.7
5.3
5.7
(1.7)
(0.1)
(0.5)
(0.2)
(0.1)
(0.1)
(0.1)
(0.4)
(0.2)
(0.1)
(0.3)
(0.9)
(0.9)
(0.3)
(0.7)
(0.6)
(0.7)
(0.9)
(1.3)
(0.5)
(0.9)
(0.3)
coloured carrot genotypes, and will provide the consumers with information about gastronomic diversity
and possible applications. It was concluded that the
isolated terpenes related to the harsh avour attributes.
Further it was shown that interactions between sugar
and volatiles aect the sensory perception and are
therefore a highly relevant issue for consumer perception of coloured raw carrots. These relations between
sensory quality, chemical compounds and masking
eects are to be considered in the future development
of new genotypes with high consumer preference.
Acknowledgment
The authors gratefully acknowledge the excellent technical assistance of Birgitte Foged and Leon Hansen.
References
Alasalvar, C., Grigor, J.M. & Quantick, P.C. (1999). Method for the
static headspace analysis of carrot volatiles. Food Chemistry, 65,
391397.
Alasalvar, C., Grigor, J.M., Zhang, D., Quantick, P.C. & Shahidi, F.
(2001). Comparison of volatiles, phenolics, sugars, antioxidant
vitamins, and sensory quality of dierent colored carrot varieties.
Journal of Agricultural and Food Chemistry, 49, 14101416.
ASTM. (1986). Physical Requirements Guidelines for Sensory Evaluation (Special Technical Publication 913).
Baardseth, P., Rosenfeld, H.J., Sundt, T.W., Skrede, G., Lea, P. &
Slide, E. (1996). Evaluation of carrot varieties for production of
deep fried carrot chips II. Sensory aspects. Food Research
Burdock, G.A. (2002). Fenarolis Handbook of Flavor Ingredients, 5th
edn. London, UK: CRC Press.
Buttery, R.G., Black, D.R., Haddon, W.F., Ling, L.C. & Teranishi, R.
(1979). Identication of additional volatile constituents of carrot
roots. Journal of Agricultural and Food Chemistry, 27, 13.
Czepa, A. & Hofmann, T. (2003). Structural and sensory characterization of compounds contributing to the bitter o-taste of carrots
(Daucus carota L.) and carrot puree. Journal of Agricultuaral and
Gills, L.A., Resurreccion, A.V.A., Hurst, W.C., Reynolds, A.E. &
Pathak, S.C. (1999). Sensory proles of carrot (Daucus carota L.)
cultivars grown in Georgia. HortScience, 34, 625628.
(1.0)
(0.4)
(1.0)
(0.6)
(0.8)
(2.1)
2.1)
(1.3)
(0.5)
(1.4)
(0.3)
(0.21)
(0.51)
(0.78)
(0.81)
(0.62)
(1.01)
(1.17)
(0.74)
(0.53)
(0.71)
(2.46)
Table 5 Genotypes with nearly similar terpene levels are compared regarding to total
sugar content calculated from glucose, fructose and sucrose, and scores in terpene avour, sweetness and bitterness. Mean values
and standard deviations
Hogstad, S., Risvik, E. & Steinsholt, K. (1997). Sensory quality and

chemical composition in carrots: A multivariate study. Acta Agriculturae Scandinavica, Sect. B, Soil and Plant Science, 47, 253264.
Howard, L.R., Braswell, D., Heymann, H., Lee, Y., Pike, L.M. &
Aselage, J. (1995). Sensory attributes and instrumental analysis
relationships for strained processed carrot avour. Journal of Food
Science, 60, 145148.
ISO. (1993). Sensory Analysis General Guidance for Selection, Training
and Monitoring of Assessors. ISO 8586-1:1993. Geneva: ISO.
Jongen, W.M.F. (2000). Food supply chains: from productivity toward
quality. In: Fruit & Vegetable Quality. An Integrated View (edited by
R.L. Shewfelt & B. Bruckner). pp. 320. CRC Press LLC.
Kaack, K., Christensen, L.P., Hansen, S.L. & Grevsen, K. (2004).
Non-structural carbohydrates in processed soft fried onion (Allium
cepa L.). European Food Research and Technology, 218, 372379.
Kjeldsen, F., Christensen, L.P. & Edelenbos, M. (2001). Quantitative
analysis of aroma compounds in carrot (Daucus carota L.) cultivars
by capillary gas chromatography using large-volume injection
technique. Journal of Agricultural and Food Chemistry, 49,
43424348.
Kjeldsen, F., Christensen, L.P. & Edelenbos, M. (2003). Changes in
volatile compounds of carrots (Daucus carota L.) during refrigerated
and frozen storage. Journal of Agricultural and Food Chemistry, 51,
54005407.
Kreutzmann, S., Thybo, A.K. & Wender, B. (2006). Training of a
sensory panel and proling of winter hardy and coloured carrot
genotypes. Food Quality and Preference, 18, 482489.
Lafuente, M.T., Lopez-Galvez, G., Cantwell, M. & Yang, S.F. (1996).
Factors inuencing ethylene-induced isocoumarin formation and
increased respiration in carrots. Journal of the American Society for
Horticultural Sciences, 121, 537542.
Martens, M., Rosenfeld, H.J. & Russwurm, H. Jr. (1985). Predicting
sensory quality of carrots from chemical, physical and agronomical
variables: a multivariate study. Acta Agriculturae Scandinavica, 35,
407420.
NIST. (1998). The NIST EPA NIH Mass Spectral Database, version
6.0. Washington DC, USA: National Institute of Standards and
Technology.
Rosenfeld, H.J., Risvik, E., Samuelsen, R.T. & Rdbotton, M. (1997).
Sensory proling of carrots from northern latitudes. Food Research
Rosenfeld, H. J., Aaby, K. & Lea, P. (2002). Inuence of temperature
and plant density on sensory quality and volatile terpenoids of
carrot (Daucus carota L.) root. Journal of the Science of Food and
Rosenfeld, H.J., Vogt, G., Aaby, K. & Olsen, E. (2003). Interaction of
terpenes with sweet taste in carrots (Daucus carota L.). Acta
Horticulturae 637: XXVI International Horticultural Congress:

Advances in Vegetable Breeding.
Seljasen, R., Hoftun, H. & Bengtsson, G.B. (2000). Sensory quality of
ethylene-exposed carrots (Daucus carota L, cv Tukon) related to
the contents of 6-methoxymellein, terpenes and sugars. Journal of
the Science of Food and Agriculture, 80, 18.
Seljasen, R., Bengtsson, G.B., Hoftun, H. & Vogt, G. (2001). Sensory
and chemical changes in ve varieties of carrot (Daucus carota L) in
response to mechanical stress at harvest and post-harvest. Journal of
the Science of Food and Agriculture, 81, 436447.
Shamaila, M., Durance, T. & Girard, B. (1996). Water blanching
eects on headspace volatiles and sensory attributes of carrots.
Simon, P.W., Peterson, C.E. & Lindsay, R.C. (1980). Genetic and
environmental inuences on carrot avor. Journal of the American
Society for Horticultural Sciences, 105, 416420.
Simon, P.W., Peterson, C.E. & Lindsay, R.C. (1982). Genotype, soil,
and climate eects on sensory and objective components of carrot
avor. Journal of the American Society for Horticultural Sciences,
107, 644648.
Talcott, S.T., Howard, L.R. & Brenes, C.H. (2001). Factors contributing to taste and quality of commercially processed strained
carrots. Food Research International, 34, 3138.
Toth-Markus, M. & Takacs-Hajos, M. (2001). Flavour substances of
carrot cultivars. Acta Alimentaria, 30, 205218.
Ulrich, D., Nothnagel, T., Hoberg, E. & Straka, P. (2003). Analysis of
aroma types by rapid GC methods in carrot breeding research. In:
Flavour Research at the Dawn of the Twenty-First Century.
Proceedings of the 10th Weurman Flavour Research Symposium
(edited by J.L. Le Quere & P.X. Etievant). Pp. 440444. Intercept
Scientic Technical Medical Publisher.
Varming, C., Jensen, K., Mller, S. et al. (2004). Eating quality of
carrots correlations between avour compounds, sensory proling
analysis and consumer liking test. Food Quality and Preference, 15,
531540.
Yoo, K.S., Pike, L.M. & Hamilton, B.K.A. (1997). A direct headspace
sampling technique to rapidly measure low boiling point volatile
terpenoides in carrots. HortScience, 32, 714716.
1627
1628
Original article
Effect of film and temperature on the sensory, microbiological and
nutritional quality of minimally processed cauliflower
Ana Simon,1 Elena Gonzalez-Fandos2* & Domingo Rodrguez1
1 Servicio de Investigacion Agroalimentaria y Desarrollo Tecnologico de La Rioja, Crta. Logrono-Mendavia, NA-134 km 87, 26071 Logrono,
La Rioja, Spain
2 Departamento de Agricultura y Alimentacion, University of La Rioja, C Madre de Dios 51, 26006 Logrono, La Rioja, Spain
(Received 5 February 2007; Accepted in revised form 24 July 2007)
Summary
The eect of atmosphere modication, generated using three dierent packaging lms, on the quality of
cauliower minimally processed when stored at 4 or 8 C for up 20 days was evaluated. The colour, texture,
weight loss, sensory attributes, as well as microbial counts and sugars and ascorbic acid content were
determined. The atmosphere generated with the perforated polyvinylchloride (PVC) lm was hardly
modied, whereas the other two lms (non-perforated PVC and polypropylene lms) originated changes in
CO2 and O2 levels during storage. The dierent packaging conditions and storage temperature inuenced
yellowing. An increase in shear force was observed. Weight losses were below 5%. Mesophiles and
Pseudomonas counts were below 7 log CFU g)1, the populations being lower with lm B and lm C than
with lm A. Cauliower maintained an acceptable appearance in all the lms studied. Total sugars decreased
about 27% after 20 days of storage, whereas ascorbic acid did not change.
Keywords
Ascorbic acid, atmosphere packaging, food quality, minimally processed vegetables, sensory evaluation.
Introduction
Among minimally processed vegetables, cauliower can

be marketed cut in orets and packed in small units
overwrapped with an appropriate lm to avoid excessive
dehydratation and the generation of inadequate atmospheres. This product must also be kept under refrigeration during marketing in order to maintain its quality
and enhance its shelf life.
High quality cauliower is white to cream in colour,
without yellowing or browning caused by sun exposure. Other desirable attributes are good compactness,
no mechanical damage and no microbial or fungal
attacks (Forney & Toivonen, 2004). Sugar content in
addition to some glucosinolates has been associated
with cauliower avour (Schonhof et al., 2004). Cauliower is also known for its high vitamin C content
mainly in the form of ascorbic acid (Lee & Kader,
2000).
Whole cauliower storage at 0 C with 95%98%
relative humidity can maintain good quality for up to

e-mail: elena.gonzalez@unirioja.es
3 weeks or even more depending on the initial quality

(Forney & Toivonen, 2004). According to Romo-Parada et al. (1989), shelf life can be extended up to 52 days
in a controlled atmosphere of 3% O2 and 2.5%5%
CO2, at 1 C and 100% relative humidity, compared
with 25 days for cauliower kept in air. However, other
authors consider that the benets from controlled
atmosphere are modest (Kader, 1992; Suslow & Cantwell, 1999; Forney & Toivonen, 2004). Atmospheres
containing O2 levels below 2% and CO2 levels above 5%
can induce physiological disorders, symptoms of
injury include o-odours, softening and discoloration
(Romo-Parada et al., 1989).
As cauliower has a high respiration rate (Kader,
1992), packaging with an adequate lm is essential in
order to avoid the generation of a passive atmosphere
inside the package that may damage the product.
Dierent lms have been used to pack minimally
processed vegetables, the most common commercial
packaging being perforated and non-perforated polyvinylchloride (PVC). However, some European countries have limited the use of PVC as wrapping lm for
horticultural products. Recent developments in packaging technology have led to the production of new lms
whose O2 permeability can be selected in advance
doi:10.1111/j.1365-2621.2007.01672.x
The effect of film and temperature on minimally processed cauliflower A. Simon et al.
(P-Plus: microperforated polypropylene lms); these

lms could be an alternative to PVC for commercial
wrapping of cauliower.
The refrigeration temperature in shelves where minimally processed vegetables are marketed can normally
range between 4 and 8 C. There is little information on
the eect of these temperatures and the packaging lms
on the passive atmospheres created by cauliower and
its inuence on the quality and shelf life of this
vegetable.
The aim of this study was to evaluate the sensorial,
nutritional and microbiological quality of cauliower
packed in three dierent types of lms when stored at 4
or 8 C for up to 20 days. Sensory quality was evaluated
using the following parameters: colour, texture and
appearance evaluation. Weight loss was also evaluated.
Microbial quality was evaluated by the following
determinations: mesophiles, Pseudomonas, Enterobacteriaceae, coliforms, faecal coliforms, Escherichia coli and
moulds. Sugars and ascorbic acid content were also
determined.
Harvesting and processing of cauliflower
Cauliowers (Brassica oleracea var. Botritis cv. Lepini)

weighing 1 kg without leaves and 16 cm in diameter
were harvested in La Rioja (Spain). Inorescences
surrounded by a crown of well-trimmed leaves were
selected, in order to avoid yellowish because of the sun
exposure. Immediately after picking, they were cooled at
3 C and stored under refrigeration and 90% relative
humidity for 24 h.
Cauliowers were cut in orets and placed in polystyrene trays measuring 22.5 13.5 3 cm (approximately 400 g per tray). The trays were overwrapped with
three dierent lms whose characteristics are shown in
Table 1. PVC lms were provided by Borden Espana SA
(Madrid, Spain). Polypropylene (PP) oriented microperforated lm was provided by Amcor Flexibles
Europe (Gloucester, UK).
The cauliowers packages were divided into two
batches that were stored at 4 and 8 C, respectively
and in 90% relative humidity for up 20 days.
The gases determination (CO2 and O2) inside the

packages were carried out on days 1, 8, 13 and 20 of
storage. Samples were taken on days 0, 8, 13 and 20 of
storage and the following variables were determined:
colour, texture, weight loss, sensorial appearance, total
and reducing sugars and ascorbic acid content. Microbiological counts (mesophiles, Pseudomonas, Enterobacteriaceae, coliforms, faecal coliforms, E. coli and
moulds) were analysed on days 0 and 20 of storage. All
the parameters were measured using two trays each time
for each treatment.
Gas determination
Carbon dioxide and oxygen were determined using a

Checkmate model 9900 O2 and CO2 head space gas
analyser (PBI-Dansensor, Ringsted, Denmark). Samples
were taken automatically with a syringe.
Colour
Colour was determined by measuring the L*, a* and b*

parameters. The measurement was taken using a HunterLab MiniScan XE colorimeter (HunterLab, Reston,
VA, USA) with an 8-mm diameter diaphragm calibrated
with a white tile (X = 81.1, Y = 86.0 and Z = 91.8).
The measurement conditions were illuminant C with an
observation angle of 10. In each tray, four determinations were made. The mean was calculated for each tray.
Texture
Texture was measured using an Instron Universal

Testing Machine (Instron Model 1140; Instron Ltd,
High Wycombe, UK) equipped with a Kramer standard
shear cell. Each measurement was taken on 45 g of
cauliower placed on the Kramer cell. The shear press
displacement speed was 50 mm min)1 and maximum
shear force was measured in newtons (N). For each tray,
two measurements were taken and the mean calculated.
Weight loss
Cauliower weight losses were measured in two trays

from each treatment on days 8, 13 and 20. On days 8, 13
Table 1 Film characteristics
Treatment
Film type
Thickness
(lm)
O2 transmission rateb
(mL m)2day)1atm)1)
Permeability
CO2 : O2 ratio
WVTRc
(g m)2day)1)
A
B
C
Perforated PVCa
Non-perforated PVC
Microperforated oriented polypropylene
12
12
35
25 000
45 000
3:1
1:1
200
0.9
Data provided by manufacturer.

Pore size 1 mm diameter, one pore for each 2.5 cm2 of film surface; bAt 25 C; cWater vapour transmission rate at 25 C and 75% HR.
1629
1630
and 20, the dierences, compared with the initial weight,

were calculated and expressed as percentage weight loss.
Twenty-ve grams of cauliower were aseptically

weighed and homogenised in a Stomacher (IUL, Barcelona, Spain) for 2 min with 225 mL of sterile peptone
water (Oxoid, Basingstoke, UK). Further decimal dilutions were made with the same diluent. The total number
of mesophilic microorganisms was determined on plate
count agar (PCA; Merck, Darmstadt, Germany) following the pour plate method, incubating at 30 C for 72 h
(ICMSF, 1978). Enterobacteriaceae were determined on
violet red bile glucose agar (Difco, Detroit, MI, USA), the
plates were overlayed prior to incubation at 37 C for 18
24 h (ICMSF, 1978). Pseudomonas spp. were determined
on Kings B medium (King et al., 1954) with an incubation temperature of 25 C for 48 h. Coliforms were
determined by the MPN method for a three-tube series
using brilliant green bile lactose (BGBL) broth (Difco)
incubated at 30 C for 48 h (ICMSF, 1978). Faecal
coliforms were determined by the MPN method for a
three-tube series using BGBL broth incubated at 44 C
for 48 h; when gas was formed, subcultures were made
onto Levine agar (Merck) and incubated at 37 C for
48 h. The plates were then examined for suspected E. coli
colonies (ICMSF, 1978). Moulds were determined on
oxitetracycline gentamicine yeast extract (OGYE) agar
(Merck) with and incubation temperature of 25 C for
5 days (ICMSF, 1978). All analyses were in duplicate.
Sensorial evaluation
Samples were evaluated for appearance by a panel of

twelve judges who were regular consumers of cauliower.
A structured hedonic scale (Anzaldua-Morales, 1994)
with numerical scores 1 (I dislike it a lot), 2 (I dislike it
slightly), 3 (neither like nor dislike), 4 (I like it slightly) and
5 (I like it a lot) was used. A score of 3 was considered the
borderline of acceptability. The presence of unpleasant
odours when the package was opened was also observed.
colorimetric method described by Cakmak & Marschner

(1992), where ascorbic acid reacts with 2,2-dipyridyl,
giving a rose colour whose absorbance was read at
525 nm in a Perkin Elmer model Lambda 10 series UV
spectrophotometer (Perkin-Elmer, Norwalk, CT, USA).
The experimental design was a factorial experiment

(3 2 3) in which the factors were the type of lm,
temperature and the time of storage with two repetitions. For microbiological data, seven treatments were
considered: day 0 and the six treatments corresponding
to day 20 (three types of lm for two temperatures).
Analysis of variance was performed using the systat
program for Windows, Statistics version 7.0 (SYSTAT,
Inc., Evanston, IL, USA, 1992). The comparison of
means was performed using the LSD method for
P 0.05 (Dagnelie, 1975). This analysis was not applied
to the weight losses, as these were measured throughout
storage in the same packages.
Atmosphere composition
The three types of lms generated signicantly dierent

atmospheres (P 0.001) in CO2 and O2 concentrations
(Table 2). The atmospheres generated with the perforated PVC lm (A) were hardly modied, being very
similar to air at 4 or 8 C (Fig. 1). The greatest change
in the atmosphere generated by the other two types of
lms (B and C) was reached in the rst 24 h of storage.
The non-perforated PVC lm (B) generated CO2 levels
of around 7%, without being aected by storage
temperature, whereas the O2 levels were 3.9% and
1.4% at 4 and 8 C, respectively. No signicant eect of
temperature was observed with the microperforated
polypropylene lm (C) on the rst day of storage, with
CO2 levels of around 8% and O2 levels of 15% (Fig. 1).
Table 2 Results of the variance analysis
Factor
CO2
O2
L*
a*
b*
TS
RS
AA
Film
Temperature
Time
Film temp.
Film time
Temp. time
***
n.s.
***
n.s.
***
n.s.
***
***
***
*
***
n.s.
*
*
n.s.
n.s.
n.s.
n.s.
***
n.s.
n.s.
n.s.
n.s.
n.s.
***
***
*
**
n.s.
***
***
n.s.
n.s.
n.s.
n.s.
n.s.
*
n.s.
***
n.s.
n.s.
n.s.
n.s.
n.s.
**
n.s.
n.s.
n.s.
*
n.s.
**
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s
n.s.
Total and reducing sugars content
Sugars were analysed in an aqueous extract using the

reagents Carrez I and II. The reducing sugars were
determined by the Bertrand method (Lecoq, 1965),
whereas total sugar content was determined by the same
method after hydrolysis with concentrated hydrochloric
acid.
Ascorbic acid content
Ascorbic acid content was determined after extraction

with a metaphosphoric acid solution (5% w v) using the
Significance level for the following factors: type of film, temperature and
storage time on the following parameters: carbon dioxide (CO2), oxygen
(O2), colour (L*, a*, b*), shear force (F), sensorial score (S), total sugars
(TS), reducing sugars (RS), ascorbic acid (AA).
n.s., not significant.
*P < 0.05; **P < 0.01; ***P < 0.001.
15
18
21
According to Romo-Parada et al. (1989), CO2 levels

above 5% can cause physiological injuries in cauliower. In the present study, the non-perforated lm (B)
and the polypropylene lm (C) generated CO2 levels
above 5% on day 1. After 8 days of storage, CO2 levels
were below 5% with non-perforated PVC lm (B) for
both temperatures tested, whereas with polypropylene
lm (C) levels around 5% were reached at 4 and 7% at
8 C.
Oxygen concentration below 2% was only reached
with lm B at 8 C on day 1. All the other values
observed were above 2%, which is considered the limit
below o-odours and o-avours could be produce in
cauliower (Romo-Parada et al., 1989).
15
18
21
Colour
CO2 (%)
10
8
6
4
2
0
0
12
Storage (days)
25
O2 (%)
20
15
10
5
0
0
12
Storage (days)
A4
B4
C4
A8
B8
C8
Figure 1 Oxygen and carbon dioxide concentrations in minimally

processed cauliower packaged in three different types of lm
(A: perforated PVC, B: non-perforated PVC, C: microperforated
polypropylene) during storage at 4 and 8 C. LSD0.05 for
CO2 = 1.12, LSD0.05 for O2 = 2.12.
After day 1, CO2 concentrations decreased and O2

concentrations increased throughout storage with nonperforated PVC lm (B) and polypropylene lm (C) but
not with perforated lm (A) (Fig. 1). The CO2 and O2
changes during the storage period diered depending on
the lm type (Fig. 1), which explains the signicant
interaction between lm type and storage time
(P 0.01; Table 2). This change in CO2 and O2 levels
throughout storage may have been due to a decrease in
the respiration rate in cauliower as reported by RomoParada et al. (1989). O2 concentration in non-perforated
PVC lm (B) was lower at 8 C than at 4 C. However,
no signicant eect of temperature was observed with
the perforated PVC lm (A) and the microperforated
polypropylene lm (C) (Fig. 1). The O2 changes at 4
and 8 C diered depending on the lm type (Fig. 1),
which explains the signicant interaction between lm
type and storage temperature (P 0.05) for O2
(Table 2).
The O2 levels within the non-perforated PVC lm
were lower than in the polypropylene lm (Fig. 1), as
the transmission rate for this gas of polypropylene was
higher than in PVC (Table 1). The CO2 levels were
lower in the non-perforated PVC lm than in the
polypropylene lm (Fig. 1), as the CO2 transmission
rate was three times those for oxygen in the nonperforated PVC lm, whereas the transmission rate was
the same for both gases in the polypropylene lm.
The analysis of variance (Table 2) showed that lm type

and storage temperature had a signicant eect
(P 0.05) on the L* parameter (lightness). However,
the dierences were minimal and could not be visually
observed.
Film type had a signicant eect (P 0.001) on the a*
parameter (Table 2). The mean values of this parameter
were negative and close to 0 ()0.61 for lm A, )0.70 for
lm B and )0.95 for lm C).
The b* parameter showed greater change that could
be visually observed, and increases in this parameter
denote a yellowing of cauliower. The three types of
lms generated atmospheres that had dierent inuences on the b* parameter, depending on storage
temperature, with interaction being signicant between
lm type and storage temperature (P 0.01). No
changes in the b* parameter throughout storage at
4 C were observed in cauliowers kept in air (lm A),
whereas a decrease was observed when storage temperature was 8 C (Fig. 2). Discoloration was observed at
this temperature. With atmospheres generated by nonperforated lm (B) and polypropylene lm (C), an
increase in parameter b* was observed after 13 days of
storage. This increase remained after 20 days of storage
at 4 C, whereas a decrease was observed at 8 C
(Fig. 2). The dierent behaviour throughout storage,
depending on the temperature, accounted for the
signicant interaction between storage temperature
and storage time (P 0.001).
Discoloration of cauliower could be due to ethylene,
as this vegetable is extremely sensitive to this compound
(Suslow & Cantwell, 1999). Higher production of
ethylene could be expected at 8 C than at 4 C, which
explains the decrease of b* parameter in cauliowers
packaged in lms B and C and stored at 8 C. This
nding could have accounted for the dierent types of
behaviour depending on the temperature. Decrease in
the b* parameter of cauliower at 8 C in perforated
PVC lm could be due to temperature eect associated
1631
2800
30
A
Shear force (N)
Colour (b*)
25
20
15
0
9
12
15
Storage (days)
18
A4
B4
C4
A8
B8
C8
21
Figure 2 Colour (b*) of minimally processed cauliower packaged in

three different types of lm (A: perforated PVC, B: non-perforated
PVC, C: microperforated polypropylene) during storage at 4 and 8 C.
with increased respiration rate, because this lm is

considered as open air condition that cannot accumulate
any volatile compound inside the package.
These changes in cauliower colour have not been
reported elsewhere. Berrang et al. (1990) observed an
increase in chroma during the storage of cauliower at
4 C, but this eect was observed in both cauliower
stored in air and controlled modied conditions of
18% O2 and 3% CO2. On the contrary, Hodges et al.
(2006) found lower chroma values in cauliowers
stored in 3% O2 and 5% CO2 compared with those
stored in air. The dierent cultivars of cauliower used
in these experiments could explain the dierent behaviour. Chroma is correlated with the b* parameter in
cauliower.
Texture
Shear force increased after 8 days of storage compared

with the initial values, with signicant dierences
depending on the lm type (P 0.001) used for packaging (Table 2). The shear force values were higher in
cauliowers packed with perforated PVC lm (A) than
in those packed in non-perforated PVC and polypropylene lms (B and C). After 8 days of storage, no
signicant eect of storage time on shear force was
observed (Table 2) and the dierence between perforated lm (A) and the other two lms (B and C) were
maintained (Fig. 3).
The lower increase in shear force obtained in cauliowers packed in MAP could have been due to the
lower water loss during storage. The cauliowers packaged in perforated PVC lm (A) had greater water loss
than those packaged in the other two lms (B and C),
which probably caused more texture hardening or
increased shear force in cauliower with the lm A.
It can also be conrmed from the result of weight loss
(Fig. 4). Although Lipton et al. (1967) reported softening in cauliower in atmospheres with high CO2 content,
2600
C
2400
2200
2000
0
6
9
12 15
Storage (days)
18
21
Figure 3 Shear force changes in minimally processed cauliower

packaged in three different types of lm (A: perforated PVC, B: nonperforated PVC, C: microperforated polypropylene) during 20 days of
storage. The data are the mean values at both temperatures tested at 4
and 8 C. LSD0.05 = 234.
Weight loss (%)
1632
4
3
2
1
0
0
9
12
15
Storage (days)
18
A4
B4
C4
A8
B8
C8
21
Figure 4 Weight losses (%) of minimally processed cauliower packaged in three different types of lm (A: perforated PVC, B: nonperforated PVC, C: microperforated polypropylene) during 20 days of
storage at 4 and 8 C.
the levels reached in the present study does not seem the
cause of the lower texture hardening in the cauliowers
packaged in lms B and C.
Weight loss
Weight losses increased linearly with storage time,

depending rstly on the type of lm and secondly on
temperature (Fig. 4).
After 20 days of storage, the cauliowers packaged in
perforated PVC lm (A) had lost 2.7% and 3.2% of
their initial fresh weight at 4 and 8 C, respectively, and
were the highest weight losses observed. For the
cauliowers packaged in non-perforated PVC lm (B),
the weight losses were 2% and 2.3% at 4 and 8 C,
respectively. In contrast, the weight losses of the
cauliower packed in polypropylene lm (C) were
signicantly lower (0.61% for both storage temperatures). However, this lm caused condensation, which
could have facilitated microbial growth compared with

lm B.
Weight loss is mainly because of water loss by
vegetable transpiration and by water evaporation, which
is inuenced by the water vapour transmission rate of
the lm. As the polypropylene lms had a lower water
vapour transmission rate than the PVC lms (Table 1),
the weight losses were also lower, but the high relative
humidity within the package resulted in the condensation of liquid water that, due to the antifog treatment,
was accumulated inside the package. Higher weight
losses were observed at 8 C than at 4 C, mainly in
cauliower packed in lm A.
The weight losses found were below 5%, as considered by Artes & Mart nez (1999) as the permissible limit
in order to avoid wilting or dryness symptoms in
cauliower.
Microbiological quality
The initial mesophiles counts in cauliower were low,

2.76 log CFU g)1, being Pseudomonas the most numerous group (2.39 log CFU g)1). No Enterobacteriaceae,
coliforms or moulds were detected on day 0 (Table 3).
Our initial mesophile counts were lower than those
reported by Berrang et al. (1990) (5 log CFU g)1),
whereas Garg et al. (1990) reported mesophile counts
between 3.6 and 5.4 log CFU g)1 for packed cauliower. These higher counts in packed cauliower may
have been due to the higher initial counts or contamination during processing.
Mesophiles reached populations of 6.66.7 log
CFU g)1 after 20 days of storage in cauliower packed
in perforated PVC lm (A). Pseudomonas growth was
very similar as shown in Table 3. These values were
below the 7 log CFU g)1 established as the maximum
)1
Table 3 Initial microbial counts (log CFU g ) and after 20 days of
storage in minimally processed cauliower packed with three dierent
types of lm and stored at 4 or 8 C
Treatment
Mesophiles
Day 0
2.76a
Day 20 at 4 C
A
6.63e
B
3.95b
C
5.83d
Day 20 at 8 C
A
6.77e
B
4.94c
C
6.06d
Pseudomonas
Enterobacteriaceae
Moulds
2.39a
<1a
<2a
6.67f
4.00b
5.67d
2.82c
<1a
1.58b
3.02b
<2a
2.15a
6.81f
5.10c
6.09e
3.66d
3.32d
4.25e
3.64c
<2a
2.24a
A: perforated PVC, B: non-perforated PVC, C: microperforated polypropylene.

Means with different superscript alphabets are significantly different
(P 0.05).
LSD test.
mesophile count for ready-to-use-vegetables under

Spanish law (Royal Decree 3484, 2000).
In non-perforated PVC lm (B), mesophiles, Pseudomonas and Enterobacteriaceae counts were lower than
in the other two lms tested, at both temperatures (4 and
8 C; Table 3). The atmosphere generated with nonperforated PVC lm may have inhibited the growth of
aerobic microorganisms. In non-perforated PVC lm
(B), mesophiles were 2.68 and 1.83 log units lower than
in perforated PVC lm (A) at 4 and 8 C, respectively.
The higher inhibition at 4 C may have been due to the
fact that the inhibitory eect of CO2 is higher at low
temperature. Similar developmental patterns were displayed by Pseudomonas. In contrast, after 20 days of
storage at 4 or 8 C, mesophile counts were only 0.8 and
0.79 log units lower in polypropylene lm (C) compared
with perforated PVC lm (A), respectively. The lower
microbial inhibition with polypropylene lm (C) compared with non-perforated PVC lm (B) was probably
because of the higher O2 levels (Fig. 1), as well as higher
moisture content in the polypropylene lm (C). In
contrast, Berrang et al. (1990) did not observe any
decrease in mesophile growth in cauliower stored at
4 C in an atmosphere of 18% O2 and 3% CO2. The
eect of MAP on aerobic mesophilic bacteria is variable.
According to King et al. (1991), the mesophilic counts in
shredded iceberg lettuce were reduced in an atmosphere
of 19% CO2 after 14 days at 2.8 C. In contrast, other
authors have observed that modied atmosphere packaging only had a limited eect on mesophiles counts
(Berrang et al., 1990; Lopez Briones et al., 1992).
Signicant dierences (P 0.05) in mesophile and
Pseudomonas counts were observed when cauliowers
packed in non-perforated PVC lm (B) stored at 4 and
8 C were compared (Table 3). No signicant eect of
temperature was observed with the other two lms
types.
The highest mould counts were observed with perforated PVC lm (A) followed by polypropylene lm (C),
where a slight incidence of black spots was observed
after 20 days of storage, possibly because of Alternaria
spp. (Artes & Mart nez, 1999). With non-perforated
PVC lm (B), no mould or black spots were observed.
The high relative humidity within polypropylene lm
(C) can increase the risk of fungal development (Artes &
Mart nez, 1999), in spite of the high CO2 levels (around
8%) reached with this type of lm (Fig. 1).
The initial Enterobacteriaceae counts in cauliower
were below 1 log CFU g)1. The numbers of Enterobacteriaceae found on various fresh vegetables ranged
between <1 and 8 log CFU g)1 (ICMSF, 1998). In
general terms, however, Enterobacteriaceae usually
represent a small proportion of the bacterial contamination (Nguyen-the & Carlin, 1994).
After 20 days of storage at 4 C, the highest Enterobacteriaceae counts were found in cauliowers packed in
1633
perforated lm (A). Moreover, the Enterobacteriaceae

counts were lower at 4 C than at 8 C. The Enterobacteriaceae counts in cauliower packed in polypropylene lm (C) were lower than in those packed in
perforated PVC lm (A) after 20 days of storage at
4 C, possibly because of the inhibitory eect of CO2
in polypropylene lm. Facultative anaerobes such as
Enterobacteriaceae are less aected by low O2 levels
than other microorganisms such as Pseudomonas and
compete better with other types of bacteria in the
presence of low oxygen concentrations (Adams & Moss,
2000). On the contrary, Gram-negative bacteria, such as
Enterobacteriaceae, are more sensitive to CO2 (Hintlian
& Hotchkiss, 1986). However, at 8 C, the growth was
higher in cauliower packed in polypropylene lm (C)
than in those packed in perforated PVC lm (A). The
lower inhibitory eect of CO2 at high temperatures, and
mainly the high humidity maintained by polypropylene
lm, could have explained this higher increase in
microbial growth. After 20 days, Enterobacteriaceae
counts reached 1.58 log CFU g)1 in cauliower packed
in polypropylene when stored at 4 C and 4.25 log CFU g)1 than when stored at 8 C. This last gure
highlights the importance of storage temperature.
Coliforms, faecal coliforms and E. coli were not
isolated in any sample.
Sensorial analysis
Visual acceptance of cauliower decreased signicantly

(P 0.001) with storage time, although scores remain
3 after 20 days of storage (Fig. 5). This value is the
borderline of acceptability. No signicant eect of
temperature was observed (Table 2). However, the type
of lm had a signicant eect (P 0.05) with better
appearance being observed in cauliowers packed in
polypropylene lm (C) after 20 days of storage. The
judges indicated that the yellowing and black spots were
Visual acceptance (1)
1634
the aspects that had the greatest inuence on the

decrease of acceptability.
O-odours were not detected in any sample, although
CO2 levels around 7% were reached, particularly with
polypropylene lm (C) at 8 C (Fig. 1). These values are
above the 5% CO2 considered to be the recommended
limit for cauliower (Romo-Parada et al., 1989; Kader,
1992). As injury may not be apparent until cauliower is
cooked (Forney & Toivonen, 2004), it would be interesting to determine whether appearance and avour of
cooked product is aected.
Sugars and ascorbic acid content
Total sugar decreased signicantly (P 0.01) with

storage time. A reduction of 27% was observed after
20 days of storage (Table 4). No signicant dierences
in total sugar content were observed between the three
types of lms or storage temperatures.
The main sugars in cauliower were reducing sugars
(glucose and fructose) whose initial content was 2.08%,
which represented 88% of total sugar content. Signicant dierences in reducing sugar content were observed
with storage time (P 0.01) and lm type (P 0.05;
Table 2). The reducing sugar content was higher in the
non-perforated lm than in the perforated PVC lm.
However, no signicant dierences in reducing sugar
content were found between the polypropylene lm and
the other two lms tested (Fig. 6).
The decrease in sugars during storage may have been
due to the use of sugar as respiration substrate (Mertens
& Tranggono, 1989). Thus, the modied atmospheres
generated in non-perforated PVC lm (B) and polypropylene lm (C) may aect the respiration rate reducing
the decrease in sugar compared with perforated PVC
lm (A). In contrast, Hodges et al. (2006) did not
observe any eect of an atmosphere with 3% O2 and 5%
CO2 on sugar content.
Table 4 Total sugar content (%) and ascorbic acid content (mg per
100 g)1) in minimally processed cauliower during 20 days of storage
5
4
3
2
1
0
9
12
15
Storage (days)
18
21
Figure 5 Visual acceptance of minimally processed cauliower packaged in three different types of lm (A: perforated PVC, B: nonperforated PVC, C: microperforated polypropylene) during 20 days of
storage. Scores: 1: I dislike it a lot, 3: neither like nor dislike, 5: I like it
a lot. The data are the mean values at both temperatures tested 4 and
8 C. LSD0.05 = 0.63.
Day of
storage
Total sugar
content (%)
Ascorbic acid
(mg per 100 g)1)
0
8
13
20
2.35a
2.01b
1.73c
1.71c
52.37
50.11
48.95
51.98
n.s.
Means with different superscript alphabets are significantly different

(P 0.05).
LSD test.
n.s., not significant.
The data are the mean values over all treatments, film type (A, B and C)
and storage temperature (4 and 8 C).
The cauliowers packed in perforated PVC lm

resulted in the highest increase in shear force, the
highest weight loss and the highest mesophile, Pseudomonas and mould counts.
All the dierent atmosphere compositions studied
were suitable for the shelf life of cauliower as no oodours were detected in any package.
Reducing sugars (%)
2.5
2.0
A
1.5
B
C
Acknowledgment
1.0
0
9
12 15
Storage (time)
18
21
Figure 6 Reducing sugar content (%) in minimally processed cauliower packaged in three different types of lm (A: perforated PVC,
B: non-perforated PVC, C: microperforated polypropylene) during
20 days of storage. The data are the mean values at both temperatures
tested 4 and 8 C. LSD0.05 = 0.31.
Initial total sugar content was similar to that reported

by Schonhof et al. (2004) for cauliower. These authors
found a multiple correlation between the sweet avour
and sacarose, glucose, fructose and some glucosinolates
content.
No signicant eects of lm type, temperature or
storage time on ascorbic acid content were observed
(Table 2). The mean value reached was 50.3 mg per
100 g FM. This value is within the range of 50103 mg
per 100 g reported by DFCD (2003).
Losses in vitamin C in vegetables during storage
depend on the type of vegetable, the cruciferous
vegetables displayed higher ascorbic acid retention
(Lee & Kader, 2000). Albrecht et al. (1990 ) reported
an ascorbic acid retention level of 96.8% in cauliower
stored at 2 C for 3 weeks. These authors found a
correlation between initial ascorbic acid content and
sulphur compounds in vegetables. Thus, the sulphur
compounds may aect high ascorbic acid retention in
cruciferous vegetables.
Conclusions
The three types of lms studied for packaging minimally processed cauliower kept an acceptable appearance without o-odour over 20 days of storage at 4 or
8 C. After 20 days of storage at 4 or 8 C, a decrease
of 27% in sugars was observed, whereas ascorbic acid
content did not change.
The atmosphere generated with non-perforated PVC
lm reduced the microbial counts but increased cauliower yellowing compared with perforated PVC lm.
The cauliower packed in polypropylene lm resulted
in less weight loss and consequently less dehydratation,
but yellowing was observed. However, cauliowers
packed with this lm obtained the highest sensorial
score after 20 days of storage.
The authors thank the regional government of La Rioja

(Spain) for its nancial support.
References
Adams, M.R. & Moss, M.O. (2000). Food Microbiology, 2nd edn. pp.
479. Cambridge: Royal Society of Chemistry.
Albrecht, J.A., Schafer, H.W. & Zottola, E.A. (1990). Relationship of
total sulfur to initial and retained ascorbic acid in selected
cruciferous and noncruciferous vegetables. Journal of Food Science,
55, 181183.
Anzaldua-Morales, A. (1994). La evaluacion sensorial de los alimentos
en la teora y en la practica, pp. 198. Zaragoza: Acribia.
Artes, F. & Mart nez, J.A. (1999). Quality of cauliower as inuenced
by lm wrapping during shipment. European Food Research and
Berrang, M.E., Brackett, R.E. & Beuchat, L.R. (1990). Microbial,
color and textural qualities of fresh asparagus, broccoli and
cauliower stored under controlled atmosphere. Journal of Food
Protection, 53, 391395.
Cakmak, I. & Marschner, H. (1992). Magnesium deciency and high
light intensity enhance activities of suporoxide dismutase, ascorbate
peroxidase and glutatione reductase in bean leaves. Plant Physiology, 98, 12221227.
Dagnelie, P. (1975). La comparaison de p moyennes deux a` deux. In:
The`orie et methodes statistiques. Applications agronomiques, pp.
242245. Gembloux: Les presses agronomiques de Gembloux.
Danish Food Composition Databank (DFCD). (2003). Available
from: http://www.foodcomp.dk.
Forney, Ch. & Toivonen, P. (2004). Cauliower: the commercial
storage of fruits, vegetables and orist and nursery stocks. Available
from: http://www.ba.ars.usda.gov/hb66/048cauliower.
Garg, N., Churey, J.J. & Splittstoesser, D.F. (1990). Eect of
processing conditions on the microora of fresh-cut vegetables.
Hintlian, C.B. & Hotchkiss, J.H. (1986). The safety of modied
packaging: a review. Food Technology, 40, 7076.
Hodges, D.M., Munro, K.D, Forney, Ch.F. & McRae, K.B. (2006).
Glucosinolate and free sugar content in cauliower (Brassica
oleracea var.botrytis cv. Freemont) during controlled-atmosphere
storage. Postharvest Biology and Technology, 40, 123132.
ICMSF (1978). Microorganisms in Foods. 1. Their Signicance and
Methods of Enumeration, 2nd edn. Toronto: University of Toronto
Press.
ICMSF (1998). Microorganisms in Foods. 6. Microbial Specications
of Food Commodities. London: Blackie Academic and Professional.
Kader, A.A. (1992). Postharvest Technology of Horticultural Crops, Pp.
296. University of California, Special Publication 3311, Davis, CA.
King, A.D, Magnusson, J.A., Torok, T. & Goodman, N. (1991).
Microora and storage quality of partially processed lettuce. Journal
King, E.O., Ward, M.K. & Ranzy, E.D. (1954). Two simple media for
the demonstration of pyocyanin and uorescein. Journal of
Laboratory and Clinical Medicine, 44, 301307.
1635
1636
Lecoq, R. (1965). Manuel danalyses alimentaires et dexpertises

usuelles, Vol. 2, Pp. 1224. Paris: Doin-Deven et Cie.
Lee, S.K. & Kader, A.A. (2000). Preharvest and postharvest factors
inuencing vitamin C content of horticultural crops. Postharvest
Biology and Technology, 20, 207220.
Lipton, W.J., Harris, C.M. & Couey, H.M. (1967). Culinary quality of
cauliower stored in CO2-enriched atmospheres. Proceedings of the
American Society for Horticultural Science, 91, 852854.
Lopez Briones, G., Varoquaux, P., Chambroy, Y., Bouquant, J.,
Bureau, G. & Pascat, B. (1992). Storage of common mushroom
under controlled atmospheres. International Journal of Food Science
Mertens, H. & Tranggono, H. (1989). Ethylene and respiratory
metabolism of cauliower (Brassica oleracea L. covar Botrytis)
in controlled atmosphere storage. Acta Horticulturae, 258, 493
501.
Nguyen-the, C. & Carlin, F. (1994). The microbiology of minimally

processed fresh fruits and vegetables. Critical Reviews in Food
Romo-Parada, L., Willemot, C., Castaigne, F., Gosselin, C. & Arul,
J. (1989). Eect of controlled atmospheres (low oxygen, high
carbon dioxide) on storage of cauliower (Brassica oleracea L.,
Botrytis group). Journal of Food Science, 54, 122158.
Royal Decree 3484 (2000). Normas de higiene para la elaboracion,
distribucion y comercio de comidas preparadas, Pp. 14351441. Spain:
Bolet n Ocial del Estado.
Schonhof, I., Krumbein, A. & Bruckner, B. (2004). Genotypic eects
on glucosinolates and sensory properties of broccoli and cauliower.
Nahrung Food, 48, 2533.
Suslow, T. & Cantwell, M. (1999). Cauliower: recommendations for
maintaining postharvest quality. Available from: http://posharvest.
ucddavis.edu/produce/producefacts/veg/cauliower.
Original article
Effects of gelatinisation level, gum and transglutaminase on the
quality characteristics of rice noodle
Seda Yalcin & Arzu Basman*
Food Engineering Department, Faculty of Engineering, Hacettepe University, Beytepe, Ankara 06800, Turkey
(Received 17 May 2007; Accepted in revised form 18 July 2007)
Summary
The aim of the present study was to investigate the eects of gelatinisation level, gum (locust bean gum,
xanthan gum, 3%) and or transglutaminase (TG, 0.5%) on quality characteristics of rice noodle. In order to
improve the dough forming ability, rice our was gelatinised at levels of 15%, 20%, 25% and 30%. Noodle
samples were evaluated in terms of cooking loss, total organic matter (TOM), water absorption, swelling
volume, maximum force, colour, sensory properties, pasting properties. Noodle sample with a gelatinisation
level of 25% had better cooking and sensory properties. Gum and or TG were added to this noodle formula.
The noodle samples including xanthan gum had better cooking and sensory properties. TG caused a
signicant decrease in TOM. The samples including locust bean gum had signicantly higher maximum force
values. Xanthan gum caused decreases in some Rapid ViscoAnalyzer viscosity values of the noodle samples,
while locust bean gum caused increases.
Keywords
Gluten-free, gum, noodle, quality, rice, starch gelatinisation, transglutaminase.
Introduction
Coeliac disease is caused by an individuals susceptibility

to the gliadin fraction of wheat gluten and similar
alcohol-soluble proteins (prolamines) of barley and rye.
Coeliac disease occurs in about one in 10002000 people
in Europe. When asymptomatic or latent forms of the
disease were taken into account, the true frequency of
the disease is estimated as close to one in 300 people
(Denery-Papini et al., 1999). The ingestion of these
proteins causes an inammatory response, resulting in
the destruction of the villous structure of the small
intestine. This leads to the malabsorption of several
important nutrients including iron, folic acid, calcium
and fat-soluble vitamins. Currently, the only eective
treatment for coeliac disease is a lifelong strict adherence
to a gluten-free diet (Feighery, 1999; Inman-Felton,
1999; Green & Jabri, 2003).
Gluten is the major structure-forming protein, responsible for the viscoelastic properties of dough and
contributes to the texture and appearance of end
products. The lack of gluten in cereals that are safe for
gluten-free food formulations causes major problems in
dough processing, product quality, mouthfeel and
e-mail: basman@hacettepe.edu.tr
doi:10.1111/j.1365-2621.2007.01674.x
avour (Gallagher et al., 2004; Moore et al., 2006).

Technological diculties in production and lack of the
awareness of the number of coeliac patients in need of
gluten-free products might restrict the number of papers
related to the gluten-free bakery products. In recent
years, various approaches to overcome the technological
problems are arising. Searching for ingredients that have
the ability to mimic properties of gluten in the production of gluten-free bakery products is the most common
approach.
Gluten-free pasta or noodle is a good choice for
coeliac patients, because these products have long shelf
life and transportation of these products from a
controlled processing centre is easy. Production of
gluten-free pasta is dicult because of the lack of
gluten. Gluten contributes to a strong protein network;
and this results in limited release of exudates during
starch gelatinisation (Feillet, 1988; Marconi & Carcea,
2001).
Utilisation of starches, hydrocolloids and emulsiers,
various treatments such as gelatinisation of raw materials (Juliano & Sakurai, 1985; Lai, 2002; Gallagher
et al., 2004; Kaur et al., 2005) can be used to solve the
problems. Huang et al. (2001) reported that gluten-free
pasta having similar quality characteristics with wheatbased pasta was produced when higher levels of
modied starch, xanthan gum and locust bean gum
1637
1638
Gum and transglutaminase in rice noodle S. Yalcin and A. Basman
were added to tapioca starch, potato starch, corn our

and rice our mixtures. Lai (2002) studied the eects of
two emulsiers (glyceryl monostearate; GMS and a
commercial emulsier; KM 3000, Sun-An Co., Taiwan)
on the quality characteristics of rice pasta and reported
that GMS improved the cooking properties of rice
pasta, especially the pasta made with high amylose rice.
The problems related to processing and end product
quality might also be compensated by using a proteinmodifying enzyme, transglutaminase (TG). TG
(protein-glutamine c-glutamyl transferase, EC 2.3.2.13)
is a promising enzyme for the food industry as a protein
modier. TG is capable of catalysing the formation of
non-disulphide covalent crosslinks between peptidebound glutaminyl residues and e-amino groups of lysine
residues in proteins (Motoki & Seguro, 1998; Larre
et al., 2000). Studies have shown that TG catalyses the
crosslinking of a number of cereal proteins such as
wheat, barley and rice (Koksel et al., 2001; Basman
et al., 2002a,b, 2003; Gujral & Rosell, 2004; Koksel &
Basman, 2005). Improving eects of TG on various
cereal products such as bread (Basman et al., 2002b,
2003), yeasted croissant and pu pastry (Gerrard et al.,
2000), Chinese noodles (Sakamoto et al., 1996) and
spaghetti (Basman et al., 2006) were reported. Basman
et al. (2006) investigated the possibility of incorporating
high levels of bran into spaghetti using TG-catalysed
crosslinking, without deterioration in spaghetti quality.
They found that TG addition had a signicant improving eect on spaghetti quality, which was more obvious
in the weaker durum wheat cultivar; and it was possible
to overcome the deteriorative eects of bran on
spaghetti quality to some extent. To the best of our
knowledge, there is no study about the eects of TG on
quality characteristics of gluten-free rice noodles. The
aim of the present study was to investigate the eects of
gelatinisation level, various gums (locust bean gum,
xanthan gum) and or TG on quality characteristics of
gluten-free rice noodles.
Materials
Commercial rice grits were ground into our by

Brabender Quadrumat Junior laboratory mill (Duisburg, Germany) and sifted through a 212-lm sieve.
Locust bean gum (Incom A.S., Mersin, Turkey),
xanthan gum (Kuzey Kimya, Istanbul, Turkey) and
transglutaminase enzyme (100 units g)1) (Ajinomoto,
Teaneck, NJ, USA) were used in the present study.
Moisture, protein and ash contents of the rice our were
determined according to AACC (1990) Approved
Methods. The colour measurement (L*, a*, b*) was
carried out in duplicate on rice our (<212 lm) using
the CIE L*a*b* colour system, where L* is lightness, a*
is redness, and b* is yellowness. The instrument used

was a Minolta CM-3600d Spectrophotometer (Minolta,
Osaka, Japan).
Starch gelatinisation
Part of the rice our (15%, 20%, 25% and 30%) was
gelatinised by adding boiling water at a ratio of our to
water of 1:2 and incubating for 5 min in a boiling water
bath. Then the gelatinised sample was rested at room
temperature for 2 h.
Noodle preparation
The gelatinised rice our (used as a binder) and

remaining rice our (85%, 80%, 75% and 70%) were
mixed with water in a mixer (KitchenAid K45SS,
KitchenAid, St. Joseph, MI, USA) for 15 min. The
amount of water was adjusted in order to obtain a total
water content of 70%, by taking the amount of water in
the gelatinised sample into account. The dough was
rounded and rested at 35 C for 30 min. The dough was
passed through the reduction rolls of a noodle machine
(Ampia 150, Marcato, Padova, Italy) to produce a
uniform dough sheet. Long dough strips obtained from
the cutting rolls of the noodle machine were dried at
45 C for 22 h. The nal dried noodle samples contained
a maximum of 10% moisture. The noodle samples were
packed in plastic bags and rested for 2 weeks before
analyses. The samples were evaluated in terms of
cooking, textural and pasting properties, sensory characteristics and colour values in order to select the best
gelatinisation level in noodle production. Gum (locust
bean gum or xanthan gum) and or TG were added to
formulation of the best noodle sample at a level of 3.0%
(w w) and 0.5% (w w), respectively. The noodle preparation procedure was also carried out for these samples
except an additional resting period at 40 C for 1 or 2 h
before drying of the noodle samples including TG. This
temperature is suitable for TG activity. These noodle
samples were also evaluated in terms of cooking,
textural and pasting properties, sensory characteristics
and colour values.
Cooking properties and texture analysis of noodles
The cooking time of the rice noodle was determined as

the time required for disappearance of opaque central
core when squeezed gently between two glass plates.
Cooking loss is the amount of solid substance lost in
cooking water. To determine cooking loss, noodle
strands were cut into 4 cm long pieces and 25 g noodle
sample was cooked in a 250 mL boiling tap water for
cooking time. An aluminium foil was used to cover the
beaker during cooking to avoid evaporation losses. The
cooking water was collected in a tared beaker, dried in
an air oven at 98 C for 24 h. The residue was weighed

and reported as percentage of the starting material.
Total organic matter (TOM) was determined according
to the method of DEgidio et al. (1982) suggested for
pasta products and is the surface material released from
cooked product during exhaustive rinsing. Water
absorption (%) and swelling volume (%) of the noodle
samples were calculated as follows
Water absorption%
Weight of cooked noodle weight of uncooked noodle
100
Weight of uncooked noodle
Swelling volume %
Volume of cooked noodle volume of uncooked noodle
100
Volume of uncooked noodle
The texture of the cooked rice noodle samples was

analysed after cooking the noodles to optimum temperature. The cooked noodle strand was wound two times
around parallel rollers of a spaghetti-noodle testing
xture (FG SPAG) (Lloyd Instruments Ltd, Fareham,
UK) to anchor the ends of the noodle strand and hinder
slippage. The maximum force (N) was determined on a
Texture Analyzer (Lloyd Instruments Ltd, Fareham,
UK) at a test speed of 3 mm s)1. Maximum force is
dened as the force required to rupture a cooked noodle
strand.
Colour analysis
The dried noodle samples were ground to pass through

the 212-lm sieve. The colour measurement (L*, a*, b*)
was carried out in duplicate on ground noodle samples
using the CIE L*a*b* colour system, where
L* is lightness, a* is redness, and b* is yellowness.
The instrument used was a Minolta CM-3600d Spectrophotometer.
Sensory analysis
Sensory analysis was carried out after cooking the

noodle samples for 10 min and cooling them at room
temperature. The samples were served in dishes labelled
randomly with two digit numbers. The sensory evaluation of the noodle samples was performed with an
evaluation panel of ve members. All of them were
habitual consumers of wheat noodles and instructions
about rice noodles were given in full to panellists
beforehand. Sensory properties evaluated were: surface
properties (wetness, slipperiness and micro roughness),
chewing properties (hardness, cohesiveness, the sensation of starch between teeth after each chew), mouthfeel
after chewing (chalkiness, stickiness) and taste. The
ballot sheet (not given) was prepared by adapting some

of the parameters of Janto et al. (1998). All of the
parameters were evaluated by a score of 15. The lower
the scores for the sensory properties, the inferior the
noodle quality. Final judgement was obtained by
averaging the scores given by all panellists.
Pasting properties
A Rapid ViscoAnalyzer (RVA-4; Newport Scientic,

NSW, Australia) with data analysis software (Thermocline, Newport Scientic) was used to analyse pasting
properties of noodle samples. The dried noodle samples
were ground to pass through the 212 lm sieve. A 3.5 g
(14% moisture basis) ground noodle sample was added
to 25.0 g of distilled water (adjusted to correct for
sample moisture content) in an aluminium canister.
RVA pasting curves were obtained using a prole given
in Table 1. The RVA curves were evaluated in terms of
peak viscosity, trough viscosity, nal viscosity, breakdown viscosity and setback viscosity.
Data related to the cooking properties of the noodle

samples prepared at various gelatinisation levels were
statistically evaluated by one-way analysis of variance
procedure. For the noodle samples including gum
and or TG, cooking properties were analysed by twoway and one-way analysis of variance procedures. The
least signicant dierence test was applied to compare
mean values. Pearson correlation coecients (r) between cooking properties and RVA viscosity values
were calculated. Standard deviations for colour values
and sensory properties were determined using Microsoft Excel.
The protein (N 5.95, db) and ash (db) contents of the

rice our were 7.5% and 0.51%, respectively. The L*, a*
and b* colour values of the rice our were found to be
95.28, 0.21 and 7.36, respectively.
Table 1 The RVA prole used for evaluation of the rice noodle
samples
Time (h:min:s)
Type
Value
00:00:00
00:00:00
00:00:10
00:01:00
00:04:42
00:07:12
00:11:00
00:13:00
Temperature
Speed
Speed
Temperature
Temperature
Temperature
Temperature
Temperature
50 C
960 r.p.m.
160 r.p.m.
50 C
95 C
95 C
50 C
50 C
1639
1640
Effects of gelatinisation level on the quality characteristics

of rice noodle samples
All of the rice noodle samples showed an optimum

cooking time of 10 min. The cooking properties of the rice
noodle samples prepared at various gelatinisation levels
are given in Table 2. Statistical analysis showed that the
gelatinisation level used for the noodle production caused
signicant dierences in all of the cooking properties
(cooking loss, water absorption, swelling volume, TOM,
maximum force) of the rice noodle samples (P < 0.05)
(Table 2). The starch gelatinisation enhanced the cooking
quality of rice noodle; and the noodle sample with a
gelatinisation level of 25% exhibited the lowest cooking
loss (11.1%) and TOM values (1.40%), indicating the best
quality characteristics among the noodle samples tested.
However, increases in cooking loss and TOM values were
observed for the noodle sample with a gelatinisation level
of 30%. When the water absorption and swelling volume
values of the noodle samples were evaluated, the noodle
sample with a gelatinisation level of 25% had signicantly
higher values when compared with those of the noodle
samples with 15% and 30% gelatinisation levels. The
noodle sample with a gelatinisation level of 25% had
swelling volume and water absorption values comparable
to those of the noodle sample with a gelatinisation level of
20%. The maximum force value increased signicantly as
the gelatinisation level increased (P < 0.05). The noodle
samples with the gelatinisation levels of 25% and 30%
resulted in signicantly higher maximum force values
when compared with those of other samples, indicating
better cooking tolerance and cooking quality. Although
the noodle sample with the gelatinisation level of 30%
resulted in signicantly higher maximum force value than
that of the sample with the gelatinisation level of 25%, its
surface properties were inferior as indicated by higher
cooking loss and TOM values.
Colour values and sensory properties of the rice
noodle samples including rice our gelatinised at various
Table 2 Cooking properties of the rice noodle samples including rice

our gelatinised at various levelsa
Gelatinisation
level (%)
Cooking
lossb
(%)
Water
absorptionc
(%)
Swelling
volumec
(%)
TOMd
(%)
Maximum
forceb
(N)
15
20
25
30
15.4
12.3
11.1
15.2
108.0
116.5
117.5
108.0
140 b
157 a
158 a
142 b
2.15
1.50
1.40
1.83
0.51
0.54
0.62
0.69
a
b
c
a
b
a
a
b
a
c
c
b
c
c
b
a
TOM, total organic matter.

a
Values followed by the same letter in the same column are not
significantly different (P < 0.05); bMeans are based on triplicate analyses;
c
Means are based on duplicate analyses; dMeans are based on quadruplicate analyses.
levels are presented in Table 3. Slight changes were

observed in L*, a* and b* colour values of the noodle
samples tested. Incorporation of gelatinised rice our up
to 25% into noodle formulation generally led to a
progressive increase in all sensory scores.
The RVA curves of the rice noodle samples including
rice our gelatinised at various levels (15%, 20%, 25%
and 30%) are given in Fig. 1. The gelatinisation level
used for the noodle production did not cause considerable changes in viscosity values of the noodle samples.
Overall results indicated that the noodle sample
including rice our with a gelatinisation level of 25%
was found to have better cooking properties when
compared with other noodle samples. Therefore, the
gelatinisation level of 25% was used for the second part
of the study in which the eects of gum and or TG on
the quality characteristics of noodle samples were
investigated.
Effects of gum and or transglutaminase on the quality
characteristics of the rice noodles with a gelatinisation level
of 25%
Supplementation of gum and TG to noodle formula and

resting periods of 1 and 2 h in TG-supplemented
samples improved machining ability of the dough and
caused smooth noodle surface. The eects of gum
and or TG on the cooking properties of the rice noodle
samples including rice our gelatinised at the level of
25% are given in Table 4. The two-factor (gum and TG)
anova results showed that eects of interaction of gum
and enzyme treatment on the cooking properties of the
rice noodle samples were insignicant (data not presented). Therefore, it was decided to use one-factor
anova for statistical evaluation of the data. One-factor
(gum) anova results (Table 5) showed that supplementation of rice noodle samples with xanthan gum resulted
in signicant decreases in cooking loss and TOM values,
indicating improved noodle quality. Noodle samples
supplemented with xanthan gum had higher water
absorption, swelling volume and maximum force values
when compared with those of control sample. However,
the increases were insignicant (a = 0.05). Addition of
locust bean gum to noodle formula caused signicantly
lower TOM values and signicantly higher maximum
force values, indicating an improvement in noodle
quality. Although the cooking loss values of the noodle
samples including locust bean gum were slightly lower
when compared with those of control sample, the
decrease was not signicant (a = 0.05) (Table 5). The
swelling volumes of the noodle samples supplemented
with locust bean gum were found to be signicantly
lower than that of control noodle sample, indicating an
inferior quality. One-factor (TG) anova results showed
that TG addition followed by a resting period (1 or 2 h)
prior to drying did not cause a signicant change on
Table 3 Colour values and sensory properties of the rice noodle samples including rice our gelatinised at various levels
Sensory
properties
Colour
Gelatinisation level (%)
L*
15
20
25
30
89.54
89.58
89.10
89.50
a*
0.05
0.21
0.02
0.11
0.12
0.12
0.16
0.14
b*
0.02
0.01
0.01
0.01
6.86
6.88
7.38
7.35
0.45
0.17
0.30
0.11
Surface
properties
Chewing
properties
Mouthfeel after
chewing
Taste
3.0
2.9
3.5
3.2
2.5
2.8
3.3
3.2
2.3
2.5
3.4
2.8
2.0
2.1
3.6
3.2
0.58
0.32
0.69
0.30
0.34
0.06
0.70
0.14
0.27
0.50
0.22
0.43
0.00
0.22
0.55
0.45
Values are SD.
5000
100
Temperature
90
Viscosity (cP)
70
60
3000
50
40
2000
30
1000
Temperature (C)
80
4000
20
10
0
Figure 1 RVA curves of the rice noodle sam-
0
0
ples including rice our gelatinised at various

levels. (The bars indicate the standard deviations of the means).
10
12
Time (min)
Gelatinized rice flour (%);
15
20
25
30
Table 4 Cooking properties of the rice noodle samples supplemented with gum and or transglutaminase (TG) enzyme
Gelatinisation
level (%)
Gum
25
25
Locust bean gum
25
Xanthan gum
TG
(incubation time, hours)
Cooking lossa
(%)
Water
absorptionb (%)
Swelling
volumeb (%)
TOMc (%)
Maximum
forcea (N)
TG(1)
TG (2)
TG (1)
TG (2)
TG (1)
TG (2)
11.1
11.0
10.9
10.9
10.6
10.6
10.4
10.3
10.1
117.5
117.5
118.0
116.0
116.0
116.0
118.0
118.5
118.5
158
154
150
147
147
153
154
158
158
1.40
1.40
1.25
1.28
1.20
1.13
1.10
1.00
0.90
0.62
0.64
0.71
0.78
0.89
0.80
0.85
0.78
0.73

a
Means are based on triplicate analyses; bMeans are based on duplicate analyses; cMeans are based on quadruplicate analyses.
cooking loss, water absorption, swelling volume and

maximum force values (data not shown). However, both
1 and 2 h resting periods with TG caused a signicant
decrease in TOM values of the noodle samples. The
noodle samples rested for 1 and 2 h had average TOM
values of 1.21% and 1.09%, respectively, while the
noodle samples without TG had an average TOM value

of 1.26%. All of the noodle samples were found to be
signicantly dierent from each other in terms of TOM
values. Basman et al. (2006) also reported that TG
addition caused signicantly lower TOM values in bransupplemented spaghetti samples. Signicant decrease in
1641
1642
The RVA curves of the gum and or TG-supplemented

rice noodle samples including rice our gelatinised at a
level of 25% are given in Fig. 2. The viscosity values of the
TG-supplemented noodle sample (without gum) rested
for 1 h were similar to those of the noodle sample without
TG. However, resting period of 2 h caused slight
decreases in peak and nal viscosity values of the TGsupplemented sample. Addition of locust bean gum to
noodle formula generally caused increases in peak, trough
and nal viscosity values whereas xanthan gum caused
decreases in peak and nal viscosity values. TG addition
followed by a resting period of 2 h caused decreases in
peak, trough and nal viscosity values of noodle samples
including locust bean gum while the RVA curve for 1 h
rested noodle sample was similar to that of the noodle
sample including only locust bean gum. The peak, trough
and nal viscosity values of the noodle samples including
xanthan gum and TG decreased as the resting period with
TG increased. The correlation coecients between cooking properties and RVA viscosity values of the rice noodle
samples supplemented with gum and or TG are given in
Table 7. Cooking loss and TOM values were found to be
positively correlated with peak, breakdown, nal and
setback viscosity values. Correlations for breakdown and
setback viscosity values were signicant at 0.01 level.
Water absorption values were negatively correlated with
peak viscosity, trough viscosity and nal viscosity values
(P < 0.01). Negative correlations were also observed
between swelling volume and some of the viscosity values
(peak viscosity, trough viscosity, nal viscosity)
(P < 0.05). These results indicate good potential of the
RVA viscosity values to give an idea about some of the
cooking properties of the rice noodle samples including
gum and or TG.
Table 5 Effects of various gums on cooking properties of the rice

noodle samplesa
Cooking Water
Swelling TOMd Maximum
forceb
lossb
absorptionc volumec (%)
(%)
(%)
(%)
(N)
Gum
11.0 a
Locust bean gum 10.7 a
Xanthan gum
10.3 b
117.7 ab
116.0 b
118.3 a
154.0 a
149.0 b
156.7 a
1.35 a 0.66 b
1.20 b 0.82 a
1.00 c 0.72 b
Values followed by the same letter in the same column are not
significantly different (P < 0.05); bMeans are based on triplicate analyses;
c
Means are based on duplicate analyses; dMeans are based on quadruplicate analyses.
TOM values of the TG-supplemented noodle or spaghetti samples can be explained by the formation of
covalent crosslinks catalysed by TG probably causing
reduced amounts of solids released during cooking.
Crosslinked proteins might form a network around the
starch granules and encapsulate them during cooking
and restrict the diusion of starch.
The colour values and sensory properties of the rice
noodle samples supplemented with gum and or TG are
given in Table 6. Gum and or TG addition seemed to
cause slight changes in L*, a* and b* colour values of
the noodle samples. A slight increase was observed for
a* colour values of the noodle samples supplemented
with xanthan gum and or TG. Sensory analysis results
showed that the noodle samples including xanthan gum
generally gave the highest sensory scores among the
noodle samples evaluated. The noodle samples including
locust bean gum seemed to be similar to control noodle
sample in terms of surface properties, chewing properties and taste. TG addition followed by a resting period
(1 or 2 h) prior to drying generally gave slightly higher
scores for the surface, chewing and mouthfeel after
chewing properties of the noodle samples supplemented
with xanthan gum.
Conclusions
The noodle sample with a gelatinisation level of 25%

gave the best cooking and sensory properties among the
Table 6 Colour values and sensory properties of the rice noodle samples supplemented with gum and or transglutaminase (TG) enzyme
Colour
Gelatinisation
level (%)
Gum
25
25
25
Sensory properties
TG
(incubation time, hours) L*
TG
TG
Locust bean gum
TG
TG
Xanthan gum
TG
TG
(1)
(2)
(1)
(2)
(1)
(2)
89.10
89.48
89.38
89.33
89.31
89.02
89.22
88.77
88.34
a*
0.02
0.11
0.34
0.25
0.13
0.40
0.20
0.38
0.08
0.16
0.19
0.17
0.16
0.21
0.22
0.27
0.31
0.36
Surface
Chewing
Mouthfeel
properties properties after chewing Taste
b*
0.01
0.04
0.01
0.04
0.02
0.01
0.04
0.03
0.01
7.38
7.63
7.89
7.15
7.63
7.77
7.42
7.86
8.41
0.30
0.35
0.09
0.18
0.48
0.31
0.58
0.30
0.26
3.5
3.3
3.1
3.1
3.2
3.3
4.0
4.3
4.2
0.69
0.43
0.43
0.62
0.17
0.15
0.38
0.27
0.61
3.3
3.1
3.0
3.1
3.1
3.2
3.6
4.2
4.1
0.70
0.24
0.14
0.11
0.31
0.11
0.21
0.19
0.48
3.4
3.3
3.2
2.7
2.9
3.1
3.8
4.2
4.3
0.22
0.43
0.45
0.27
0.22
0.45
0.57
0.27
0.45
3.6
3.3
3.3
3.2
3.2
3.3
4.4
4.4
4.4
0.55
0.67
0.45
0.45
0.45
0.67
0.42
0.55
0.55
Values are SD.
100
Temperature
5000
90
Viscosity (cP)
70
60
3000
50
40
2000
30
20
1000
10
Figure 2 RVA curves of the rice noodle samples
(with a gelatinisation level of 25%) supplemented

with gum and or transglutaminase (TG), grf,
gelatinised rice our; TG1, incubation with TG
for 1 h; TG2, incubation with TG for 2 h. (The
bars indicate the standard deviations of the
means).
0
0
10
11
12
13
Time (min)
Peak
Trough
Breakdown Final
Setback
viscosity viscosity viscosity
viscosity viscosity
0.789*
0.948**
)0.852** )0.535
)0.729* )0.430
0.780*
0.925**

*P < 0.05; **P < 0.01.
rice noodle samples produced in the present study. Rice

noodle samples including xanthan gum were better when
compared with the ones including locust bean gum. TG
addition improved machining ability of the dough,
caused smooth noodle surface and improved rice noodle
quality in terms of TOM values. High correlation
coecients between the cooking properties and some
of the RVA viscosity values of the rice noodle samples
including gum and or TG showed that the RVA has
good potential to give an idea about some of the
cooking properties of the rice noodle samples produced
in the present study. However, further investigations are
required to support the conclusions derived from the
present study.
Acknowledgment
The present study was supported by Afyon Kocatepe

University Scientic Research Projects Section,
Afyon Turkey.
25% grf
25% grf, TG1
25% grf, TG2
Table 7 Pearson correlation coefcients between cooking properties

and RVA viscosity values of the rice noodle samples supplemented
with gum and or transglutaminase
Cooking loss
0.689*
0.293
0.860**
Water absorption )0.877** )0.891** )0.068
Swelling volume )0.704* )0.793* 0.102
TOM
0.703*
0.305
0.867**
Temperature (C)
80
4000
25% grf, locustbean gum

25% grf, locustbean gum, TG1
25% grf, locustbean gum, TG2
25% grf, xanthan gum

25% grf, xanthan gum, TG1
25% grf, xanthan gum, TG2
References
American Association of Cereal Chemists (AACC). (1990). Approved
Methods of the AACC. St Paul: AACC.
Basman, A., Koksel, H. & Ng, P.K.W. (2002a). Eects of transglutaminase on SDSPAGE patterns of wheat, soy and barley proteins
and their blends. Journal of Food Science, 67, 26542658.
Basman, A., Koksel, H. & Ng, P.K.W. (2002b). Eects of increasing
levels of transglutaminase on the rheological properties and bread
quality characteristics of 2 wheat ours. European Food Research
Basman, A., Koksel, H. & Ng, P.K.W. (2003). Utilization of
transglutaminase to increase the level of barley and soy our
incorporation in wheat our breads. Journal of Food Science, 68,
24532460.
Basman, A., Koksel, H. & Atli, A. (2006). Eects of increasing levels
of transglutaminase on cooking quality of bran supplemented
spaghetti. European Food Research and Technology, 223, 547551.
DEgidio, M.G., De Stefanis, E., Fortini, S. et al. (1982). Standardization of cooking quality analysis in macaroni and pasta products.
Cereal Foods World, 27, 367368.
Denery-Papini, S., Nicolas, Y. & Popineau, Y. (1999). Eciency and
limitations of immunochemical assays for the testing of gluten-free
foods. Journal of Cereal Science, 30, 121131.
Feighery, C.F. (1999). Coeliac disease. British Medical Journal, 319,
236239.
Feillet, P. (1988). Protein and enzyme composition of durum wheat.
In: Durum Wheat Chemistry and Technology (edited by G. Fabriani
& C. Lintas). Pp. 93119. St Paul: The American Association of
Cereal Chemists.
Gallagher, E., Gormley, T.R. & Arendt, E.K. (2004). Recent advances
in the formulation of gluten-free cereal-based products. Trends in
Gerrard, J.A., Newberry, M.P., Ross, M., Wilson, A.J., Fayle, S.E. &
Kavale, S. (2000). Pastry lift and croissant volume as aected by
microbial transglutaminase. Journal of Food Science, 65, 312314.
Green, P.H.R. & Jabri, B. (2003). Coeliac disease. The Lancet, 362,
383391.
Gujral, H.S. & Rosell, C.M. (2004). Functionality of rice our
modied with a microbial transglutaminase. Journal of Cereal
Science, 39, 225230.
Huang, J.C., Knight, S. & Goad, C. (2001). Model prediction for
sensory attributes of non-gluten pasta. Journal of Food Quality, 24,
495511.
1643
1644
Inman-Felton, A.E. (1999). Overview of gluten-sensitive enteropathy

(Celiac Sprue). Journal of the American Dietetic Association, 99,
352.
Janto, M., Pipatsattayanuwong, S., Kruk, M.W., Hou, G. &
McDaniel, M.R. (1998). Developing noodles from US wheat
varieties for the Far East market: sensory perspective. Food Quality
and Preference, 9, 403412.
Juliano, B.O. & Sakurai, J. (1985). Miscellaneous rice products. In:
Rice: Chemistry and Technology (edited by B.O. Juliano). Pp.
592599. St Paul: The American Association of Cereal Chemists.
Kaur, L., Singh, J. & Singh, N. (2005). Eect of glycerol monostearate
on the physico-chemical, thermal, rheological and noodle making
properties of corn and potato starches. Food Hydrocolloids, 19,
839849.
Koksel, H. & Basman, A. (2005). Eects of transglutaminase enzyme
on rice and amaranth proteins. In: Book of Abstracts: ICC-Jubilee
Conference Cereals the Future Challenge. Pp. 104. Vienna:
International Association for Cereal Science and Technology.
Koksel, H., Sivri, D., Ng, P.K.W. & Stee, J.E. (2001). Eects of
transglutaminase enzyme on fundamental rheological properties of
sound and bug-damaged wheat our doughs. Cereal Chemistry, 78,

2630.
Lai, H.-M. (2002). Eects of rice properties and emulsiers on the
quality of rice pasta. Journal of the Science of Food and Agriculture,
82, 203216.
Larre, C., Denery-Papini, S., Popineau, Y., Deshayes, G., Desserme,
C. & Lefebvre, J. (2000). Biochemical analysis and rheological
properties of gluten modied by transglutaminase. Cereal
Marconi, E. & Carcea, M. (2001). Pasta from nontraditional raw
materials. Cereal Foods World, 46, 522530.
Moore, M.M., Heinbockel, M., Dockery, P., Ulmer, H.M. & Arendt,
E.K. (2006). Network formation in gluten-free bread with
application of transglutaminase. Cereal Chemistry, 83, 2836.
Motoki, M. & Seguro, K. (1998). Transglutaminase and its use for
food processing. Trends in Food Science and Technology, 9, 204210.
Sakamoto, H., Yamazaki, K., Kaga, C., Yamamoto, Y., Ito, R. &
Kurosawa, Y. (1996). Strength enhancement by addition of
microbial transglutaminase during Chinese noodle processing.
Nippon Shokuhin Kagaku Kaishi, 43, 598602.
Original article
Quality factors and grades for the classification and
standardisation of complex ready pasta meals
Maddalena Kindt,* Palmira Mazzaracchio & Giancarlo Barbiroli
Department of Business Science, University of Bologna, Piazza Scaravilli, 2 40126 Bologna, Italy
(Received 4 July 2007; Accepted in revised form 24 October 2007)
Summary
The aim of this study is to arrive at a new classication of ready pasta meals, through the evaluation of their
global quality, since their growing number, present in the market, often may be cause of confusion to
consumers, because the parameters that contribute to the quality level are not clearly identiable and
measurable. After having explored several possible quality factors, we arrived at the conclusion that at least
ve must be taken into consideration: degree of pulping of the pasta, capacity to hold the sauce, weight
constancy, wholeness and defects. Then, three quality grades are here proposed and discussed. Particular
emphasis should be given to the sauce, because it highly inuences on the global quality of these foodstus.
Keywords
Cooking, food processing aspects, food quality, pasta, processed food, sauces, vegetables.
Introduction
Consumers are becoming increasingly aware of the

importance of safe and high-quality food products.
Interest grows as new products foods are introduced
onto the market and modern technologies are being
used (bakery products, frozen ready-to-eat meals,
organic and others special foods). The evaluation of a
food product, processed to be transformed, needs the
identication of several quality factors, such as
appearance, avour, nutritional properties, texture and
microbiological safety. The healthy features of frozen
ready-to-eat products have been assessed by nutritional
and microbiological analyses (Yookyung et al., 2005),
and microbiological guidelines for ready-to-eat foodstus, according to the Codex Alimentarius of the USA,
have already been developed in several countries
(Gilbert et al., 2000; Food Safety Authority of Ireland,
2001).
Among the wide varieties of frozen ready-to-eat
meals present in the market, pasta is a very popular
food. To obtain high-quality frozen food, one of the
variables that can inuence on product quality is the
freezing rate. Several researchers have studied this
subject to know to what extent this factor plays a role
in the nal quality (De Kock et al., 1995; Gormley
et al., 2002).
The shelf life of ready-to-eat meals is not only
determined by the microbiological and chemical stabil-

e-mail: maddalena.kindt@unibo.it
doi:10.1111/j.1365-2621.2007.01693.x
ity, but also by several physicochemical phenomena.

The sauce instability, the migration between sauce and
pasta and the consequent pasta modications, can lead
to products not only with a dierent texture and sensory
quality (Gonzales et al., 2000; Redmond et al., 2004;
Farley & Reed, 2005; Olivera & Salvadori, 2006), but
also with new quality factors, useful to make an
evaluation of the global merit.
In a previous study, new standards based on a wide
range of quality factors for foods such as cakes,
snacks, biscuits and pastries had been proposed
(Barbiroli & Mazzaracchio, 1994). Recent research
has been carried out on frozen ready pasta meals, to
study some new characteristics of these products
(Redmond et al., 2004, 2005), mainly connected with
the presence of fats (Kindt et al., 2006). Even if the
Italian pasta ready meals are made from separate
cooked pasta and sauces (in other countries the sauces
are mixed with the pasta), this research work can have
a methodological relevance, leading to general
assumptions, when implemented to any kind of pasta
samples. New data about the quality of these products
are here reported: three quality grades are proposed,
on the bases of specic standards. One of these
quality factors, the capacity to hold the sauce (CHS),
has been largely studied, to test the pasta samples
with several kinds of sauces. Because of the complexity of the composition of the sauces, it seemed to be
useful to add sauces made with similar basic ingredients with the corresponding dierent samples of pasta
to study the behaviour of pasta avoured with
dierent sauces and to identify the connections among
quality factors.
1645
1646
Quality factors and grades for the classification and standardisation M. Kindt et al.
Product definition
Twenty samples of ready pasta meals consisting of

partially cooked and frozen pasta (simply called pasta)
and cooked and frozen sauce were purchased in a
local supermarket in Bologna (Italy): eight samples of
pasta were made with durum wheat and eggs,
whereas the other samples were only made with
durum wheat. Pieces of cooked and frozen sauce with
dierent shapes (cubes, spheres or individual
ingredients) were present in the packaging, separate
from the pasta. The cooking methods described by the
manufacturer were traditional for all types of products. All products were tested within the shelf life
reported on the package. Table 1 lists the commercial
name of the products analysed, their ingredients, the
nutritional value and the package ratio. Three products (Tagliolini con gamberi e zucchine, Tagliolini
allo scoglio and Spaghetti alle vongole veraci) were
bought after the analysis of the other pasta samples to
develop some considerations drawn from the
experiments.
Ratio pasta sauce
Products are classied as pasta with egg and pasta

without egg, and frozen pasta and sauce were weighed
separately to verify the percentage of ratio reported in
the label (Table 1).
Composition
The composition of ingredients has been divided into

two classes: basic and complementary ingredients. This
classication describes the ingredients present in the
pasta and sauce, respectively.
Basic pasta ingredients:
Durum wheat
Water
Salt
Complementary pasta ingredients:
Sunower oil
Egg
Basic sauce ingredients:
Cheese sauce: cheese, vegetable fat and salt.
Tomato sauce: tomato (pulp or juice), vegetable oil and
salt.
Fish sauce: sh, vegetable oil and salt.
Mushroom sauce: mushroom, vegetable oil, onion and
garlic.
Meat sauce: meat, tomato pulp, vegetable oil, onion and
salt.
Complementary sauce ingredients:

Cheese sauce: skimmed milk, milk protein, egg, nutmeg,
wheat our, wheat starch and modied starch.
Tomato sauce: cheese, other vegetables, onion, garlic,
wheat starch, sugar, basil and pepper.
Fish sauce: shes broth, vegetables, onion, garlic,
parsley, wine, pepper, chili pepper and wheat starch.
Mushroom sauce: skimmed milk, egg, wheat starch,
parsley, wine, yeast, vegetable fat, cheese and cream.
Quality factors
On the basis of previous direct experience on various

foodstus, the research was oriented towards some main
factors, able to contribute to a global prole.
Moisture content
According to the AOAC Ocial method (AOAC,

2000, method 926.07), moisture was evaluated on
frozen pasta. Although the processing method used
to prepare the ready pasta meals was unknown, the
cooking point of these products was veried by
observing defrosted pieces of single units of pasta
squeezed under two slides: every sample analysed
showed an uncooked starch core, conrming the
manufacturers partially cooked description. Commercial pasta (20 g), by the same manufacturer, when
available, was subsequently immersed in a beaker with
boiling water (400 mL) maintained at 100 C. Small
pieces of units were put every minute between two
slides, squeezed and compared to the result obtained
for the defrosted samples, until the two samples seemed
the same. After obtained the optimal cooking time, the
moisture content of cooked sample pasta was measured, according to the AOAC Ocial method
(AOAC, 2000, method 926.07). This test was applied
in duplicate to commercial pasta samples of dierent
sizes, with and without egg, and their moisture data
were calculated to establish a standard range of water
content for cooked pasta, as frozen ready products.
The moisture data of all samples analysed are the mean
of two data and the SD is <1.5% for all samples
(Table 2).
Degree of pulping (DP)
This test was carried out on all samples considered,

according to a previous research work (Kindt et al.,
2006): frozen pre-cooked samples (20 g) of pasta were
placed in a ask with 100 mL of distilled water and
shaken for 5 min with B and T shaker (Biard and
Tatlock London, England). The liquid phase was
poured into a 200-mL volumetric ask. An aliquot
(100 mL) was transferred into a glass dish, evaporated
at 105 C overnight and placed in a desiccator. The
Table 1 List of ingredients, nutritional values and package ratio of the analysed products
Sample
Commercial
product
Pasta A
Sauce a
Nutritional values
(g 100 g) reported
on the label
Firm
Ingredients
Mezze Penne ai
formaggi
Pasta B
Sauce b
Mezze Penne
ai formaggi
Pasta C
Saice c
Mezze Penne
allarrabbiata
Pasta: durum wheat, water, sunflower oil, salt.

Sauce: skimmed milk, gorgonzola, mozzarella,
emmenthal, taleggio, vegetable fat, cream,
sunflower oil, wheat flour, egg, wheat starch,
citric acid, milk proteins, salt, nutmeg
Pasta: durum wheat, water, salt.
Sauce: skimmed milk, mozzarella, cream,
emmenthal, gorgonzola, butter, taleggio, edam,
milk proteins, palm oil, salt, modified starch,
sunflower oil, egg, nutmeg, pepper
Sauce: tomato pulp, extravirgin olive oil, onion,
salt, parsley, garlic, sugar, wheat starch, chilly, pepper
Pasta D
Sauce d
Mezze penne
con pomodoro
e mozzarella
Pasta E
Sauce e
Penne tricolore
Pasta F
Sauce f
Fusilli alla
Sorrentina
Pasta G
Sauce g
Pennette
Con melanzane,
pomodorini
e ricotta salata
Pasta H
Sauce h
Caserecce alla
Siciliana
Pasta I
Sauce i
Paccheri
Caserecci
Pasta J
Sauce j
Spaghetti
con calamari,
polpo e vongole
Pasta K
Sauce k
Linguine allo
scoglio
Pasta L
Sauce: l
Spaghetti con
le vongole

Sauce: tomato pulp, mozzarella, basil, sunflower
oil, garlic, salt
Sauce: tomato pulp, mozzarella, sunflower oil, salt,
basil, pepper, chilly
Sauce: tomato juice, mozzarella, sunflower oil,
onion, cream, salt, basil, wheat starch, sugar,
garlic, pepper, flavours
Sauce: tomato pulp, aubergins, tomatoes,
salt ricotta, vegetable oil, ricotta, onion,
extravirgin olive oil, pecorino cheese, basil,
salt, pepper
Sauce: tomato juice, aubergins, tomatoes, ricotta,
extravergin olive oil, sunflower oil, onion, salt,
flour, flavours, wheat starch, garlic, sugar
Sauce: tomato pulp, pecorino cheese, sunflower
oil, ricotta, onion, salt, garlic, basil, pepper, chilly
Pasta: durum wheat, water, salt. Sauce: tomato
pulp, tomatoes, squids, cod, mussels, extravirgin
olive oil, octopus, clams, parsley, white wine,
garlic, salt, chilly
Sauce: tomato juice, squids, octopus, cuttlefish,
mussels, tomatoes, shrimps, sunflower oil, onion,
white wine, concentrated tomato, clams soup,
extravirgin olive oil, wheat starch, parsley, garlic,
salt, natural flavours, anchovy paste sugar
Sauce: clams broth, clams, wine, garlic, parsley,
extravirgin olive oil, salt, chili pepper
Package
ratio (%)
Proteins: 6.5 g
Carbohydrates: 27.0 g
Total fat: 6.5 g
192 Kcal
Pasta: 64
Sauce: 36
Proteins: 7.6 g
Total fat: 5.5 g
185 Kcal
Pasta: 60
Sauce: 40
Proteins: 4.0 g
Carbohydrates:
Total fat: 3.0 g
131 Kcal
Proteins: 4.8 g
Carbohydrates:
Total fat: 3.3 g
162 Kcal
Proteins: 5.5
Carbohydrates:
Total fat: 4.0 g
158 Kcal
Proteins: 4.6 g
Carbohydrates:
Total fat: 3.0
129 Kcal
Proteins: 4.5 g
Carbohydrates:
Total fat: 4.2 g
143 Kcal
22.0 g
Pasta: 58
Sauce: 42
28.3 g
Pasta: 63
Sauce: 37
25.0 g
Pasta: 60
Sauce: 40
21.0 g
Pasta: 56
Sauce: 44
21.9 g
Pasta: 50
Sauce: 50
19.0 g
Pasta: 54
Sauce: 46
22.0 g
Pasta: 55
Sauce:45
19.3 g
Pasta: 55
Sauce: 45
16.0 g
Pasta: 55
Sauce: 45
Proteins: 4.0 g
Carbohydrates:
Total fat: 3.0 g
119 Kcal
Proteins: 5.0 g
Carbohydrates:
Total fat: 4.0 g
144 Kcal
Proteins: 4.5 g
Carbohydrates:
Total fat: 2.0 g
113 Kcal
Proteins: 7.0 g
Carbohydrates:
Total fat: 3.0 g
119 Kcal
Proteins: 4 g
Carbohydrates: 24 g
Total fat: 1.5 g
126 Kcal
Pasta: 61
Sauce: 39
1647
1648
Table 1 (Continued)
Sample
Commercial
product
Pasta M
Sauce m
Nutritional values
(g 100 g) reported
on the label
Firm
Ingredients
Linguine allo
scoglio
Pasta N
Sauce n
Pappardelle al ragu`
di cinghiale
Pasta O
Sauce o
Pappardelle
ai funghi
Pasta P
Sauce p
Tagliatelle ai funghi
Pasta Q
Sauce q
Taglierini con
gamberetti robiola
e scorzetta di limone
Pasta R
Sauce r
Tagliatelle
al salmone
Pasta S
Sauce s
Tagliolini
allo scoglio
Pasta T
Sauce t
Tagliolini con
gamberi e
zucchine
Pasta: durum wheat, eggs, water, sunflower oil, salt.

Sauce: octopus, tomato pulp, sunflower oil, brandy,
tomatoes, squids, shrimps, garlic, onion, parsley,
anchovy paste, salt, wheat starch, dehydrated
fish soup, chilly, white pepper, yeast extract, laurel.
Pasta: durum wheat, eggs, water, sunflower oil, salt.
Sauce: tomato pulp, wild boar meat, onion, double
concentred tomato, extravirgin olive oil, celery,
carrots, red wine, garlic, salt, natural flavours
Pasta: durum wheat, eggs, water, salt.
Sauce: water, cream, mushrooms (Boletus edulis,
Pholiata nameko), margarine, onion, extravirgin
olive oil, garlic, cheese, modified starch, white
wine, egg, yeast exstract, salt, skimmed
milk powder, flavours
Pasta durum wheat, eggs, water, water sunflower oil, salt.
Sauce: mushrooms (Agaricus Bisporus, Boletus Edulis)
skimmed milk rehydrated, cream, water sunflower oil,
onion, garlic, salt, wheat flour, parsley, eggs,
wheat starch, pepper.
Sauce: robiola cheese, prawns, skimmed milk,
parsley, extravirgin olive oil, wheat starch, lemon-peel,
garlic, salt
Sauce: cream, salmon, tomato, vegetable oil, milk,
shrimps squash, onion, fish soup, carrots, celery,
double concentrated tomato, wheat flour,
modified maize starch.
Tomatoes, shellfish (octopus, cuttlfish, squid,
mussels, prawns), chopped tomatoes, clam broth,
cod, sunflower oil, whit wine, wheat starch,
onions, parsley, anchovy paste, salt garlic, chili pepper.
Sauce: shelfishes and crustaceans broth, water,
prawns, vegetable marrow, white wine, potatoes,,
carrots, onion, sunflower oil, extravirgin olive oil,
garlic, brandy, celery, salt, parsley, chili pepper,
yeast extract, white pepper.
same treatment was applied to frozen samples after

having weighed and warmed them up for 5 min in a pan
on a hot plate, maintained at 200 C. Calculation of the
DP has been done as follows:
dec
100
DP
def
where DP = degree of pulping, dec = dried extract
from frozen sample after 5 min of warming (mg g)1
frozen sample), def = dried extract from frozen sample
(mg g)1 frozen sample).
Package
ratio (%)
Proteins: 6.5 g
Total fat: 5.0 g
166 Kcal
Pasta: 54
Sauce: 46
Proteins: 7.0 g
Total fat: 6.0 g
168 Kcal
Proteins: 8.4 g
Total fat: 6.2 g
199 Kcal
Pasta: 55
Sauce: 45
Pasta: 48
Sauce:52
Proteins: 5.5 g
Total fat: 8.0 g
190 Kcal
Pasta: 40
Sauce: 60
Proteins: 7.0 g
Total fat: 5.0 g
137 Kcal
Proteins: not reported
Carbohydrates: not reported
Total fat: 8.6 g
Calories: not reported
Pasta: 57
Sauce: 43
Pasta: 37
Sauce: 63
Proteins: 7.2 g
Total fat: 2.7 g
110 Kcal
Pasta: 57
Sauce: 43
Proteins: 5 g
Carbohydrates: 19 g
Total fat: 3 g
123 Kcal
Pasta: 50
Sauce: 50
Samples were tested in duplicate and the SD ranged

between 0% and 0.1%. Data of the DP test are reported
in Table 2.
Capacity to hold the sauce
To uniform the results about the CHS, pasta and sauce

ratio (dierent in the package ratio) has been used
considering a ratio of 1:1. Same amounts of frozen pasta
(20 g) and sauce (20 g) were weighed and subsequently
placed in a pan on a hot plate, following the method
described in a previous study (Kindt et al., 2006): frozen
samples of pasta were weighed and subsequently placed
Table 2 Moisture, capacity to hold the sauce (CHS) and degree of

pulping (DP) of the ready pasta meals analysed
Samples
Moisture (%)
DP (%)
CHS (%)
A
B
C
D
E
F
G
H
I
J
K
M
N
O
P
Q
R
50.6
54.5
49.8
55.2
54.2
51.4
50.7
57.6
49.5
60.1
62.5
55.2
50.6
30.5
52.6
61.5
31.3
0.74
1.35
1.01
1.26
0.87
1.39
1.27
1.49
1.07
1.06
1.10
0.90
1.09
0.73
0.93
1.18
1.04
25.2
40.6
22.8
34.5
29.2
27.9
45.4
42.0
13.6
41.8
33.2
32.6
41.2
51.5
40.0
35.6
43.4
with the corresponding sauce in a pan on a hot plate,

maintained at a temperature of 200 C. According to the
manufacturers instructions, the products were warmed
up for 5 min. After warming up, the samples were
transferred into a colander with a 12-cm diameter,
shaken for 5 min and weighed. The dierence between
the weight of pasta with sauce and the frozen pasta
without sauce has determined the amount of sauce held.
Table 2 summarises the results of CHS test about the
twenty samples analysed.
This test has been applied even for cross-tests, and an
example of it is the following: the sauce of Mezze penne
ai formaggi (sauce a by rm 1) was added to the
corresponding pasta (pasta A), and to the pasta Mezze
penne ai formaggi by rm 2 (pasta B). The same crosstest has been applied even to the sauce of Mezze penne
ai formaggi by rm 2 (sauce b) with both abovementioned pasta samples (pastas A and B). This crosstest has been used for all other samples of sauces with
the same basic ingredients: tomato sauce, sh sauce and
mushrooms sauce. Even if the meat sauce considered is
the only one analysed, it was added to several kinds of
pasta with egg to evaluate its capacity to be held by the
pasta.
Spaghetti con le vongole veraci (Sample L-l), Tagliolini allo scoglio (sample S-s) by rm 6 and Tagliolini
con gamberi e zucchine (Sample T-t) by rm 1 have
been studied not only with the cross-test implemented
for the other products, but also after having separated
the large pieces of vegetables and shes present in the
sauces.
To evaluate the importance of the size of the main
components, the larger pieces of vegetable and sh
(812 mm) have been cut into small pieces (smaller than
3-mm diameter), and they have been added to the pasta.

This treatment has been allowed either with only pieces
of vegetable or shes, and with both pieces of them.
The cross test here described actually has widened the
number of the original samples that can be considered to
be 81 (Tables 3 and 4).
All samples were analysed in two replicates, and the
SD ranged between 0% and 3.8%.
Weight constancy
This factor, already previously experienced (Barbiroli &

Mazzaracchio, 1994), has been now applied to ready
pasta meals with some minor modications: the average
weight of 100 U of frozen pasta was calculated, with a
range corresponding to the SD. Table 5 reports on the
data regarding the weight constancy of the pasta
analysed.
Wholeness
Wholeness is considered when units of pasta have no

breaks or small crumbs (Barbiroli & Mazzaracchio,
1994): an amount corresponding to 100 g of units of
frozen pasta was measured and the number of whole
units found was expressed as percentage of number of
whole unit 100 U (Table 5).
Defects of the pasta
This test measures the imperfections in the product, such

as anomalous colouration (Barbiroli & Mazzaracchio,
1994). As some pasta samples presented irregular white
spots in limited points of the surface, these spots were
cut and weighed. The total amount of spots was
calculated as the percentage of spots weight in 100 g
of units present in the package (Table 5).
The purpose of setting food quality standardisation

systems can be achieved by upgrading the quality level
of the food industry, increasing the self-management
and quality control capabilities of the food enterprises,
providing safeguards to the consumer hygiene.
Previously, the standardisation of products had
already been implemented for dierent foodstus,
arriving at specic proposals (Barbiroli & Mazzaracchio, 1994). To apply an analogous quality system for
standardisation of frozen ready-to-eat meals, given their
complexity, some meaningful parameters have been
investigated and measured, and ve useful quality
factors have been chosen: DP, CHS, weight constancy,
wholeness of the units and defects of the pasta.
Frozen pasta moisture is the rst factor that has been
considered. In section Moisture content, a method to
test the water content for the cooked pasta is described,
and the range of the moisture content has been detected
between 50% and 65%.
1649
1650
Table 3 Cross-tests regarding the capacity to hold the sauce (CHS) of the analysed samples
Pasta
Firm
Cheese Sauce
Mezze Penne
1
Mezze Penne
1
Mezze Penne
2
Mezze Penne
2
Tomato sauce with mozzarella
Mezze Penne
4
Mezze Penne
4
Mezze Penne
4
Penne
1
Penne
1
Penne
1
Fusilli
2
Fusilli
2
Fusilli
2
Tomato sauce with ricotta
Pennette
3
Pennette
3
Pennette
3
Caserecce
2
Caserecce
2
Caserecce
2
Paccheri
1
Paccheri
1
Paccheri
1
Fish sauce
Linguine
2
Linguine
2
Spaghetti
3
Spaghetti
3
Meat sauce
Pappardelle
1
Pappardelle
2
Tagliatelle
1
Taglierini
3
Tagliatelle
5
Fish sauce
Pappardelle
1
Pappardelle
1
Pappardelle
2
Pappardelle
2
Tagliatelle
1
Tagliatelle
1
Taglierini
3
Taglierini
3
Tagliatelle
5
Tagliatelle
5
Tagliolini
6
Tagliolini
6
Tagliolini
1
Tagliolini
1
Mushrooms sauce
Pappardelle
1
Pappardelle
1
Pappardelle
2
Pappardelle
2
Tagliatelle
1
Sauce
Firm
Cross-test
CHS d.s.
Ai
Ai
Ai
Ai
1
2
1
2
A-a
A-b
B-a
B-b
25.2
42.1
27.0
40.6
0.3
1.3
0.2
1.7
Con pomodoro e mozzarella

Tricolore
Alla Sorrentina
Tricolore
Alla Sorrentina
Tricolore
Alla Sorrentina
4
1
2
4
1
2
4
1
2
D-d
D-e
D-f
E-d
E-e
E-f
F-d
F-e
F-f
34.5
29.1
33.3
33.1
29.2
36.9
27.5
31.8
27.9
1.5
1.2
0.0
1.1
1.5
2.0
0.4
2.2
1.1
Con melanzane, pomodorini e ricotta salata

Alla Siciliana
Caserecci
Alla Siciliana
Caserecci
Alla Siciliana
Caserecci
3
2
1
3
2
1
3
2
1
G-g
G-h
G-i
H-g
H-h
H-i
I-g
I-h
I-i
45.5
41.9
24.0
48.1
42.0
19.4
34.8
34.3
13.6
2.3
1.5
0.6
1.6
2.3
1.0
1.9
0.5
0.1
Allo
Con
Allo
Con
2
3
2
3
K-k
K-j
J-k
J-j
33.2
49.9
23.8
41.8
1.2
1.3
1.2
0.5
1
1
1
1
1
N-n
O-n
P-n
Q-n
R-n
41.0
55.3
41.2
41.1
48.6
1.4
2.4
2.6
1.5
1.8
3
5
3
5
3
5
3
5
3
5
1
1
6
1
N-q
N-r
O-q
O-r
P-q
P-r
Q-q
Q-r
R-q
R-r
S-t
S-l
T-s
T-l
38.4
31.5
47.6
40.7
43.9
28.7
35.6
35.1
49.6
43.4
40.6
14.1
42.0
27.1
1.0
0.5
0.9
2.6
0.8
0.6
2.1
1.8
2.0
0.3
0.4
0.2
0.6
1.1
1
2
1
2
1
N-p
N-o
O-p
O-o
P-p
43.7
25.2
54.5
51.5
40.0
3.1
1.6
3.8
3.0
0.2
Al
Al
Al
Al
Al
formaggi
formaggi
formaggi
formaggi
scoglio
calamari, polpo e vongole
scoglio
calamari, polpo e vongole
cinghiale
cinghiale
cinghiale
cinghiale
cinghiale
Con gamberetti, robiola e

Al Salmone
Al Salmone
Al Salmone
Al Salmone
Al Salmone
Con gamberi e zucchine
Con le vongole
Allo scoglio
Con le vongole
Ai
Ai
Ai
Ai
Ai
scorzetta di limone
scorzetta di limone
scorzetta di limone
scorzetta di limone
scorzetta di limone
funghi
funghi
funghi
funghi
funghi
Table 3 (Continued)
Pasta
Tagliatelle
Taglierini
Taglierini
Tagliatelle
Tagliatelle
Firm
Sauce
Firm
Cross-test
CHS d.s.
1
3
3
5
5
Ai
Ai
Ai
Ai
Ai
2
1
2
1
2
P-o
Q-p
Q-o
R-p
R-o
37.4
44.1
36.3
59.8
55.7
funghi
funghi
funghi
funghi
funghi
When pasta samples had moisture content under

50% 0.5% and presented dried areas in limited
points of the surface (Pastas A, O and R), it is possible
that the freezing rate has caused some freezer burns,
as described after as defects. Some pasta samples
(pastas B, D and E) showed a considerable amount of
colour defects even if their moisture content is up to
54%. On the other side, the pasta samples with a
moisture content up to 65% 0.5% had absorbed too
much water during cooking and could loose the
texture. All pasta samples analysed had a moisture
content under 65%.
Therefore, the moisture content may be considered as
fundamental for other quality factors of these foodstus, but it cannot be considered a quality factor itself.
Consequently, it is opportune to set lower (50%) and
upper (65%) limits for the moisture content to consider
the pasta within the standards.
As regards the DP, the higher grade was assigned to
samples which did not release material from the pasta
surface after warming, and the data obtained ranged
between 0.70% and 0.99%. The medium grade, from
1.00% to 1.20%, showed a moderate release of colloidal
layer formed during cooking, whereas the lower grade
has the highest release of material from the pasta
surface, 1.201.50%.
The CHS was standardised with a xed package ratio
of 50% pasta and 50% sauce. Three ranges have
been dened to evaluate the CHS: high CHS
ranged between 50.0% and 40.1%; medium CHS
ranged between 40.0% and 30.1% and low CHS ranged
between 30.0% and 20.0%.
To better understand the behaviour of the pasta with
the sauce, the cross tests applied have given meaningful
results. Tables 3 and 4 report on all CHS data obtained.
Pasta without egg: mezze penne (pastas A and B) by
rms 1 and 2 were avoured with cheese sauces (sauces
a and b), as described in the section Capacity to hold
the sauce (Fig. 1a) and linguine (pasta K) and spaghetti (pasta J), respectively, by rms 2 and 3 (Fig. 1b)
were avoured with sh sauces (sauces k and j). Both
mezze penne samples, avoured with the same sauce
(cheese sauce a or b), had similar CHS data, while
comparing the CHS data of the two cheese sauces,
mezze penne samples showed a dierent behaviour:
1.2
1.4
1.1
0.5
1.2
both mezze penne (pastas A and B) by rm 1 and 2 had

a CHS higher when avoured with sauce b. We must
highlight that these two sauces were completely homogeneous, without whole pieces. This result could be
explained by considering some dierent ingredients in
the sauce composition (the amount of vegetable fats and
cheese), responsible of this behaviour.
Changing the pasta size (linguine or spaghetti)

avoured with sh sauces (sauce k and j), CHS data
varied between pasta sample and sauce sample used
(Fig. 1b). In both experiments, linguine was able to
hold the sauces better than spaghetti. In this case,
sauces were not homogeneous and the size of whole
pieces of sh could inuence on the CHS data.
As regards tomato sauces, three products (mezze
penne pasta D; fusilli pasta F; penne pasta E)
had mozzarella cheese as a complementary ingredient of
the sauce, and another three samples (pennette pasta
G; caserecce pasta H; paccheri pasta I) had ricotta
cheese. Samples with mozzarella cheese were crossed
each other, as samples with ricotta cheese. Figure 1c
reports on CHS data of these products. The results show
that for all kinds of pasta tested, the capacity to hold the
dierent tomato sauces with mozzarella cheese is very
similar. Mozzarella cheese pieces had melted during
heating, and the sauce had become more homogeneous.
This change probably led to minor dierences among
the dierent kinds of pasta and the sauces used.
On the contrary, products with tomato sauce and
ricotta cheese showed a clear dierence between samples: paccheri (pasta I) presented the lower CHS data
with all sauces used and the corresponding sauce (sauce
i) was not able to be hold by the three pasta samples.
Paccheri is made by rm 1, which generally added
sunower oil on the pre-cooked pasta. Its CHS could be
inuenced not only by the presence of vegetable oil on
the surface of the pasta, but also from the size and the
shape of the paccheri, bigger and smoother than
pennette (pasta G) and Caserecce (pasta H). The
composition of the sauce i is quite dierent from the
other two because it had no pieces of aubergine that
could increase the weight of the sauce held by the pasta.
The presence of whole pieces seems to be an important
component to obtain higher CHS data. To conrm this
observation, a further product, Spaghetti con le vongole
1651
1652
Pasta
Sauce (fishes sauces)
Cross-test
CHS d.s.
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Spaghetti L
Tagliolini T
Tagliolini T
Tagliolini T
Tagliolini T
Tagliolini T
Tagliolini T
Tagliolini S
Tagliolini S
Tagliolini S
Tagliolini S
Tagliolini S
Tagliolini S
Allo scoglio (s)

Allo scoglio cut in small pieces (s)
whole fishes (s)
only fishes cut in small pieces (s)
whole vegetable (s)
only vegetable cut in small pieces (s)
Con gamberi e zucchine (t)
Con gamberi e zucchine cut in small pieces (t)
whole fishes (t)
only fishes cut in small pieces (t)
whole vegetable (t)
only vegetable cut in small pieces (t)
Con le vongole (l)
Con le vongole cut in small pieces (l)
Con gamberi e zucchine (t)
Con gamberi e zucchine cut in small pieces (t)
whole fishes (t)
only fishes cut in small pieces (t)
whole vegetable (t)
only vegetable cut in small pieces (t)
Allo scoglio (s)
Allo scoglio cut in small pieces (s)
whole fishes (s)
only fishes cut in small pieces (s)
whole vegetable (s)
only vegetable cut in small pieces (s)
L-s
L-s
L-s
L-s
L-s
L-s
L-t
L-t
L-t
L-t
L-t
L-t
L-l
L-l
T-t
T-t
T-t
T-t
T-t
T-t
S-s
S-s
S-s
S-s
S-s
S-s
35.1
23.5
33.6
30.2
31.5
20.8
37.9
34.9
25.7
29.4
46.6
32.6
14.6
15.9
49.4
41.0
39.7
35.9
51.1
47.6
34.1
39.5
41.2
36.0
35.5
30.4
veraci (L-l by rm 1), was avoured with its sauce

(sauce l with clams) and with sauces containing pieces
of sh and vegetables (sauce s and t). In Fig. 2a,b,
these cross-tests are reported: in Fig. 2a sauce l was less
Table 5 Quality characteristics of the ready pasta meals analysed

Defects (%)
Samples
Weight
constancy (%)
Wholeness (%)
Minor
defects
Serious
defects
A
B
C
D
E
F
G
H
I
J
K
M
N
O
P
Q
R
70.8
68.6
63.2
78.6
68.0
82.6
71.0
68.6
60.7
69.2
72.7
73.3
68.7
66.7
76.5
83.3
80.0
100
100
100
92
100
85
100
97
100
94
96
91
87
64
88
99
87
20
43
1
56
39
0
1
9
7
0
0
0
0
0
29
0
20
28
20
0
6
19
0
0
0
0
0
0
0
0
100
23
0
10
0.9
0.4
0.1
1.4
0.4
1.9
1.8
0.8
0.2
0.9
0.6
1.7
0.3
1.2
0.2
0.5
0.1
0.3
0.3
0.4
0.9
0.0
1.2
0.4
0.6
0.5
Table 4 Cross-tests regarding the capacity to

hold the sh sauces (CHS) of the analysed
samples added to pasta sample in different
way, as original sauce cut in small pieces,
whole shes, only shes cut in small pieces,
whole vegetable and only vegetable cut in
small pieces
held (it had no vegetables) rather than the other two

sauces; when spaghetti sample was avoured only with
whole pieces of sh (Fig. 2b), sauce s presented a
higher CHS data rather than the CHS of sauce t. This
dierence disappeared when shes were cut into small
pieces. On the contrary, sauce s is held in a minor
amount than sauce t when vegetables were added to the
pasta, both in whole and small pieces. In sauce s,
chopped tomatoes were present while in sauce t
chopped vegetable marrows were present. These last
ones had moisture content higher than tomatoes (91.4%
vs. 87.6%). During heating, the pasta absorbed water
from the sauce (data not reported), swelling and
increasing its CHS. From these experiments, it is
possible to observe that CHS test has been inuenced
above all by the type of ingredient and their size.
Pasta with egg: a rst cross test was carried out with a
meat sauce (ragu` di cinghiale, sauce n) and CHS data
are reported in Fig. 3a. Pappardelle (pasta O) by rm 2
and Tagliatelle (pasta R) by rm 5 had the highest CHS
data. These results were conrmed in other cross tests
for which dierent kinds of sauces were used (Fig. 3b,c).
In both cases, it is possible to observe that dierences
between two similar sauces, with sh and with mushrooms, used to avour the same type of pasta, were
present but not high. The exception to this behaviour
was Tagliatelle (pasta P) by rm 1 avoured with sh
(a)
(b)
Cheese sauce
Fish sauce
60
50
42.14
40.62
25.22
40
26.99
41.76
33.19
23.83
30
20
20
10
10
Sauce "a"
50
40
Figure 1 Comparison among the capacity to

hold the sauce data of different kind of pasta
without egg, avoured with type (a) (cheese
sauce), (b) (sh sauces) and (c) (tomato sauces
with mozzarella cheese or ricotta cheese).
30
Sauce "b"
Sauce "k"
Fusilli F
Penne E
Mezze penne D
60
50
36.89
34.46
31.79
33.30
33.12
29.1729.14
27.90
27.45
40
Paccheri I
Caserecce H
48.12 45.45
Pennette G 41.98 41.85
34.28
34.83
30
20
23.96
19.43
20 13.62
10
10
Sauce "e"
Sauce "f"
(a)
Sauce "j"
Tomato sauce with ricotta cheese
Tomato sauce with mozzarella
(c)
60
49.94
50
40
30
Linguine K
Spaghetti J
60
Mezze Penne A
Mezze Penne B
Sauce "d"
Sauce "i"
Sauce "h"
Sauce "g"
Sauce "s" whole
60
Sauce "s" cut in small pieces

Sauce "l" whole
50
Sauce "l" cut in small pieces

40
35.05
37.94
34.92
Sauce "t" whole
30
23.52
20
Sauce "t" cut in small pieces

15.94
14.63
10
0
Pasta L
(b)
60
50
40
46.64
Spaghetti L + sauce "s"
Spaghetti L + sauce "t"
37.94
35.05
34.92
33.57
30.20 29.44
Figure 2 Comparison among the capacity to hold

the sauce data of spaghetti (pasta L) avoured
with different sh sauces: (a) spaghetti avoured
with whole pieces or cut in small pieces of three
different sh sauces; (b) spaghetti avoured with
the two sh sauces having vegetables: data show
differences among the capacity to hold the whole
sauces compared to cut in small piece sauces or
the sauce where vegetables and sh have been
separated in whole or cut pieces.
30
23.52
25.68
20.84
20
10
Sauce "s" and "t"
sauces (sauce q and r) and Pappardelle (pasta N) by

rm 1 avoured with mushrooms sauces (sauce o and
p) that showed much more dierences in the CHSs.
32.61
31.50
Sauces in
small pieces
Whole fishes
Fishes in Whole vegetables Vegetables

small pieces
in small pieces
As regards the last two products, Tagliolini allo

scoglio (sample S-s by rm 6) and Tagliolini con
gamberi e zucchine (sample T-t by rm 1), cross-tests
1653
1654
(a)
60
Meat sauce
55.25
48.63
50
41.18
41.00
41.09
Pasta P
Pasta N
Pasta Q
40
30
20
10
0
Pasta O
Pasta R
Fish sauce
(b)
60
50
49.56
43.36
47.56
40.73
Sauce "q"
Sauce "r"
43.87
40
28.70
38.44
35.63 35.07
31.54
30
20
10
0
Pasta R
Pasta O
Pasta P
Pasta N
Pasta Q
Mushrooms sauce
(c) 59.78
60
50
40
30
20
10
0
55.71 54.50
51.50
40.00
37.39
Sauce "p"
Sauce "o"
43.68
44.14
36.33
25.16
Pasta R
Pasta O
Pasta P
Pasta N
Pasta Q
(d)
60
Tagliolini S + sauce "s"

49.41
39.45 40.97
40
51.13
Tagliolini T + sauce "t"
50
41.16
47.55
39.72
36.01 35.92
34.07
35.46
30.37
30
20
10
0
Sauce "s" and "t"
Sauces in
small pieces
Whole fishes
Fish in
small pieces
Whole vegetable
with the whole pieces and the cut pieces of sh or

vegetable, were done to reconrm what it was observed
with the pasta without egg, as Spaghetti con le vongole
veraci (sample L-l by rm 1). In Fig. 3d, these crosstests were reported and it is possible to observe that the
presence of vegetable marrows inuenced on the CHS of
both pasta samples, while the other combination showed
no remarkable dierences. Moreover, Tagliolini (pasta
T) held sauces better than Tagliolini (pasta S). The
dierences between the two Tagliolini samples did not
depend on the shape, but on the moisture content:
Tagliolini (pasta T) had 48.4% of moisture, while
Vegetable
in small pieces
Figure 3 Comparison among the capacity to hold

the sauce data of different kinds of pasta with egg
avoured with: (a) meat sauce; (b) two different
sh sauces; (c) two different mushrooms sauces;
(d) two different sh sauces with vegetables. In
this last experiment, pasta has been avoured
with whole sauces, cut in small pieces sauces, only
whole or cut sh pieces and only whole or cut
vegetables pieces.
Tagliolini (pasta S) had 61.7%. This result reconrms

that the pasta with less moisture led to absorb water
from the sauce, above all if the sauce has a high moisture
content (as vegetable marrows in sauce t), and this
behaviour increases the CHS.
As regards weight constancy, products with almost
75% of the units within the average weight (SD)
should be considered as having high constancy; products with 6575% of the units within the average weight
(SD) should be considered as having medium constancy; products with 5565% of the units within the
average weight (SD) should be considered as having
Table 6 Quality standard grades of the ready pasta meals

Grade A
Grade B
Grade C
Quality factor
Value
Score
Value
Score
Value
Score
DP
CHS
Wholeness (%)
Weight constancy (%)
Defects (no. in 100 U)
0.700.99
50.040.1
>96
>75.0
m.d.*20
s.d.**10
3
3
3
3
3
1.001.20
40.030.1
9296
65.175.0
m.d.*>20
s.d.**10
2
2
2
2
2
1.201.50
30.020.0
8892
55.065.0
m.d.*20
s.d.**<10
1
1
1
1
1
*m.d., minor defects; **s.d., serious defects.
low constancy, and products with <55% of the units

cannot be included in this standard.
As regards wholeness, products with almost 96% of
the units within the average weight (SD) should be
considered as having high wholeness; products with
9296% of the units within the average weight (SD)
should be considered as having medium wholeness;
products with 8892% of the units within the average
weight (SD) should be considered as having low
wholeness, and products with <88% of the units
cannot be included in this standard.
Defects observed in this type of products were
essentially the presence of white spots in limited points
of the surface. This anomalous colouration is probably
because of freezer burns: they appear when moisture
leaves the surface of the food to the storage atmosphere
and produces areas of a visible damage. Such areas have
a lighter colour (white spot) because of microscopic
cavities, previously occupied by ice crystals, which alter
the wavelength of the reected light. Colour imperfections, as freezer burns, are directly correlated to the loss
of moisture (Fellows, 1997).
It is important to highlight that whereas pasta with
freezer burns is healthy, the quality is lower.
Colour defects may be classied as
Minor defects: a unit affected by anomalous colouration, including a weight from 20% to 40% (w w).
Serious defects: a unit affected by anomalous colouration including a weight >40% (w w).
Low defectiveness was assigned when minor defects
were 20% and serious defects were 10% of total units
examined. Medium defectiveness was assigned when
minor defects were >20% and serious defects were
10%. Serious defectiveness was assigned when minor
defects were 20% and serious defects were >10%.
Table 5 shows defects values of samples analysed.
By using the ve quality factors before examined and
discussed, we can propose three quality grades for these
foodstus that are reported in Table 6. Grade A:
products with a total score of 1512. Grade B: products
with a total score of 118. Grade C: products with a
total score of 75.
This proposal considers the ve quality factors with

the same importance, but it is also possible to dierentiate them by giving each one a dierent weight on the
nal score, even if it is dicult to establish the rationale
for a dierentiation. The experience suggests to consider
the DP and the CHS as the two most meaningful factors
for assessing the global quality of these foodstus,
because they depend on some other factors such as the
moisture content, the content and type of fat, both in the
pasta and in the sauce.
References
A.O.A.C. (2000). Ofcial Methods of Analysis of the Association of
Ofcial Analytical Chemists, 17th edn. Lane, Alabama USA:
A.O.A.C., (Ocial Method 926.07, methods line 32.5.02).
Barbiroli, G. & Mazzaracchio, P. (1994). Classication and
standardisation of bakery products and our confectionery in
relation to quality and technological progress. Food Control, 5,
3338.
De Kock, S., Minnaar, A., Berry, D. & Taylor, J.R.N. (1995). The
eect of freezing rate on the quality of cellular and non-cellular parcooked starchy convenience foods. Lebensmittel-Wissenschaft undTechnologie, 28, 8795.
Farley, H.A. & Reed, Z. (2005). An integrated sensory study of
selected chilled lasagne ready meals. Food Service Technology, 5, 35
45.
Fellows, P. (1997). Food Processing Technology-Principles and Practice. London: Woodhead Publishing Ltd.
Food Safety Authority of Ireland (2001). Guidelines for the Microbiological Quality of Some Ready-To-Eat Foods Sampled at the Point
of Sale. Dublin: FSAI, Guidance note No. 3.
Gilbert, R.J., de Louvois, J., Donovan, T. et al. (2000). Guidelines for
the microbiological quality of some ready-to-eat food sampled at the
point of sale. Communicable Disease And Public Health, 3, 163167.
Gonzales, J.J., McCarthy, K.L. & McCarthy, M.J. (2000). Textural
and structural changes in lasagna after cooking. Journal of Texture
Studies, 31, 93108.
Gormley, R., Walshe, T., Hussey, K. & Butler, F. (2002). The eect of
uctating vs. constant frozen storage tmperature regimes on some
quality parameters of selected food products. Lebensmittel-Wissenschaft und-Technologie, 35, 190200.
Kindt, M., Lercker, G., Mazzaracchio, P. & Barbiroli, G. (2006).
Eect of lipids in the quality of commercial frozen ready pasta
meals. Food Control, 17, 847855.
Olivera, D.F. & Salvadori, V.O. (2006). Textural characterisation of
lasagna made from organic whole wheat. International Journal Food
Science Technology, 41, 6369.
1655
1656
Redmond, G.A., Gormley, T.R. & Butler, F. (2004). The eect of

short- and long-term freeze-chilling on the quality of cooked green
beans and carrots. Innovative Food Science & Emerging Technologies, 5, 6572.
Redmond, G.A., Gormley, T.R. & Butler, F. (2005). Eect of shortand long-term frozen storage with MAP on the quality of freeze-
chilled lasagne. Lebensmittel-Wissenschaft und-Technologie, 38,

8187.
Yookyung, K., Mukti, S. & Sandra, K. (2005). Incidence of Staphylococcus aureus in ready-to-eat meals from military cafeterias in
Ankara, Turkey. Food Control, 16, 531534.
Original article
Effect of gaseous ozone on microbial inactivation and sensory
of flaked red peppers
Meltem Y. Akbas1* & Murat Ozdemir2
1 Department of Biology, Gebze Institute of Technology, PO Box 141, 41400 Gebze, Kocaeli, Turkey
2 Department of Chemical Engineering, Section of Food Technology, Gebze Institute of Technology, PO Box 141, 41400 Gebze, Kocaeli, Turkey
(Received 14 August 2007; Accepted in revised form 3 January 2008)
Summary
Flaked red peppers inoculated with Escherichia coli, Bacillus cereus and B. cereus spores were exposed to
gaseous ozone at 20 C and 70% relative humidity (RH). Ozone concentrations of 0.1, 0.5 and 1.0 ppm up to
360 min were used to reduce E. coli and B. cereus, whereas 1.0, 5.0, 7.0 and 9.0 ppm ozone concentrations for
360 min were used to treat B. cereus spores. When aked red peppers were treated with 1.0 ppm ozone
concentration for 360 min, B. cereus and E. coli counts were decreased by 1.5 and 2.0 log numbers,
respectively. Bacillus cereus spores were reduced by 1.5 log numbers at ozone concentrations of 7.0 ppm or
above for 360 min. There were slight changes in avour, appearance and overall palatability of aked red
peppers treated with ozone between 5.0 and 9.0 ppm. Ozone concentration (1.0 ppm) for 360 min can be
used to decrease E. coli and B. cereus, whereas ozone concentrations 5.0 ppm can be used to reduce
B. cereus spores.
Keywords
Bacillus cereus, decontamination, Escherichia coli, aked red pepper, ozone, sensory, spore.
Introduction
Red pepper (Capsicum annum L.) is one of the most

important spices used as a natural avouring and
colouring agent all over the world. Red pepper grown
in Kahramanmaras region of Turkey is processed into
aked red pepper or red pepper powder. These products
are among the important export commodities of Turkey.
Spices including aked red peppers are prone to insect
infestation and microbial contamination during sun
drying, storage and transportation. Fresh red peppers
are either dried under sun or dehydrated with hot air.
When red peppers are dried under sun, they are
susceptible to bacterial contamination and spores.
Studies have shown that spices available at the retail
stores may contain a variety of bacteria and fungi
(Garcia et al., 2001). High microbial loads (up to
108 CFU g)1) may be found in spices, as shown for
black pepper and paprika (Baxter & Holzapfel, 1982;
McKnee, 1995). Such contamination is signicant
because spices are often added to foods that undergo
no further heat treatment; therefore, the surviving
pathogens have the potential to cross-contaminate other
foods. Bacteria such as Bacillus spp., E. freundii, E. coli,
e-mail: akbasm@gyte.edu.tr
Klebsiella spp., Serratia spp., Staphylococcus spp. and

Streptococcus spp. have been identied on peppers
(Freire & Oord, 2002).
Use of ethylene oxide to decontaminate spices have
recently been banned in many countries, including
European countries and Japan because of the formation
of possible toxic residues and potential health hazards
for workers in the fumigation plant (Uijl, 1992). Ionising
radiation is another method used for the decontamination of spices. The most signicant disadvantage of
ionising radiation is on the sensory properties of red
peppers. For instance, although an irradiation dose of
7 kGy reduced the population of mesophilic bacteria
and fungi eectively, it induced some changes on odour
of red pepper powder (Lee et al., 2004). Furthermore,
irradiating the packaging material induced the formation of 1,3-di-tert-butylbenzene, which migrated to red
pepper powder.
In 2001, ozone has been approved as an antimicrobial
agent in foods in the USA (FDA, 2001). There are many
applications of ozone in the food industry regarding
decontamination of poultry, fruit and vegetables, and
dried foods such as cereals, grains, black pepper and
pistachios (Naitoh et al., 1987, 1988; Zagon et al., 1992;
Zhao & Cranston, 1995; Kim et al., 1999; Liangji, 1999;
Khadre et al., 2001; Al-Haddad et al., 2005; Akbas &
Ozdemir, 2006a). Khadre & Yousef (2001) compared
doi:10.1111/j.1365-2621.2008.01722.x
1657
1658
Decontamination of flaked red peppers M. Y. Akbas and M. Ozdemir
the eectiveness of ozone and hydrogen peroxide against

Bacillus spores, and showed that ozone was more
eective against Bacillus spores than hydrogen peroxide.
Akbas & Ozdemir (2006a) studied the eect of dierent
ozone treatments for decontamination of pistachios
contaminated with B. cereus and E. coli. They found
that at 1.0 ppm ozone concentration for 360 min of
ozonation, B. cereus counts in pistachio kernels and
shelled pistachios were reduced by 3.0 log numbers,
whereas E. coli counts were decreased by 3.5 log
numbers.
The use of dierent ozone treatments on the microbial
population and sensory characteristics of aked red
peppers has not been studied. In this work, E. coli,
B. cereus and B. cereus spores inoculated to aked red
peppers were selected as model organisms to represent
the behaviour of Gram-negative and Gram-negative
bacteria, respectively. The organisms selected were
based on two important criteria. The rst criterion is
whether these organisms are the most commonly found
organisms in aked red peppers. The second criterion is
which organisms aked red peppers are the most
frequently contaminated with. Therefore, the objectives
of this study were to determine the eect of varying
ozone concentrations and exposure times for reducing
E. coli, B. cereus and B. cereus spores, and sensory
quality of aked red peppers.
Materials and ozonation system
Flaked red peppers were obtained from a local company

(Calli Co., Istanbul, Turkey). Ozonation treatments
were performed in a gas-tight plexiglass chamber
(30 30 50 cm) at a temperature of 20 0.5 C.
The relative humidity (RH) inside the chamber was
maintained at 70 3% with a saturated KCl solution
and a fan (Tuthill Pump Corp., Concord, CA, USA)
capable of operating at variable speeds. Air speeds
inside the chamber were measured using an anemometer
(Davis Instruments, Hayward, CA, USA). The fan
speed was set to 182 m min)1 to prevent stagnant air
layer formation over the surface of the product and to
achieve uniform RH throughout the chamber. RH and
temperature inside the chamber were monitored with a
digital hygrometer and thermometer (Model HD 50;
KIMO Instruments, Bordeaux, France).
Low ozone concentrations (0.11.0 ppm) in the
chamber were attained by a low-level ozone generator
(OG-200; Opal, Ankara, Turkey), whereas high ozone
concentrations (1.09.0 ppm) were generated by a highlevel ozone generator (Opal). Ozone concentration in
the chamber was continuously monitored and controlled
by a digital ozone switch (Model OS-3; Eco Sensors,
Inc., Santa Fe, NM, USA). The ozone switch has a
heated metal oxide semiconductor sensor, which acts

like a thermostat to control the ozone generator. The
sensor works by heating a small platinum substrate to
200 C. At this temperature, the substrate is very
sensitive to ozone. It sends a proportional signal to the
electronics and ozone concentration is displayed in ppm.
Microbiological inoculation
Prior to ozonation treatments, microbial cell and spore

suspensions were prepared. Although E. coli, B. cereus
and B. cereus spores were selected as model organisms in
this work, using a mix of microorganisms could provide
more information to determine ecacy of ozone.
Vegetative cells of E. coli ATCC 8739 and B. cereus
ATCC 6633 were obtained by scraping overnight
cultures from nutrient agar plates and they were diluted
with sterile NaCl solution (0.9%, w v) for immediate
use. For induction of spore formation, B. cereus ATTC
6633 has been grown on nutrient agar (Difco, Becton
Dickinson, Detroit, MI, USA) containing 500 ppm
Bacto dextrose (Difco) and 3 ppm manganese sulphate
for 7 days at 30 C. After incubation at 30 C for
7 days, sporulation was determined by the Schaeer
Fulton method (Schaeer & Fulton, 1933). Spores were
harvested, washed four times by centrifugation at
4000 r.p.m. for 20 min and re-suspended in sterile NaCl
solution (0.9%, w v). Tween 80 (0.05%, v v) (Sigma, St
Louis, MO, USA) was added to spore suspensions in
order to minimise clumping of spores and improve the
accuracy of spore counts. The spore counts were
enumerated by direct microscopic counting on a haemocytometer. For verication of the number of spores in
the spore suspension, the spore suspension was rst
heat-shocked at 80 C for 10 min to kill vegetative cells
and then quickly cooled to 5 C (Lopez-Pedemonte
et al., 2003). Next, the number of viability of spores was
estimated by spread plating of 0.1-ml spore suspension
on nutrient agar at 30 C for 24 h. The spore suspension
was stored at 4 C until used.
Flaked red peppers (100 g) were sprinkle inoculated
with microbial cell or spore suspensions in sterile bags at
a level of c. 107 microorganisms per gram. Samples were
mixed in sterile plastic bags and allowed to dry for 1 h at
25 C prior to ozonation to facilitate adherence. Inoculated and non-inoculated aked red peppers (100 g)
were divided into four separate Petri plates, and the
plates were placed into the ozonation chamber after
the desired ozone concentrations were reached inside the
chamber.
Twenty-ve gram of non-inoculated, inoculated and

ozonised, and inoculated and non-ozonised aked red
pepper samples were diluted with 225 mL of sterile
A sensory panel was done 24 h after the ozonation

treatments. Ten trained and experienced individuals
participated in the sensory evaluation of ozone and nonozone treated aked red pepper samples. The sensory
evaluation was done on the basis of ve-point hedonic
scale in an air-conditioned room at 20 C. The panellists
were trained on the use of hedonic scale and what they
need to consider during the evaluations. The highest
rating is 5 for like very much, 4 for like, 3 for neither
like nor dislike, 2 for dislike and the lowest is 1 for
dislike very much. Samples were placed into white
plates, labelled with three-digit random numbers and
given to the panellists in randomised order. Panellists
conducted sensory evaluations in individual booths
under white illumination. Cups with water were provided for panellists to use during the test to minimise
any residual eect between samples. Sensory attributes
evaluated by the panellists included taste, avour,
appearance and overall palatability (Kader et al., 1982).
All trials were replicated three times. One-way analysis

of variance (anova) was performed with spss (SPSS Inc.,
Ozone concentrations of 0.1, 0.5 and 1.0 ppm for

360 min were used to inactivate E. coli (Fig. 1). The
number of E. coli was reduced less than 1.0 log number
(a) 9
8
Log (cfu g1)
7
6
5
4
3
2
1
0
0
60
120
180
Time (min)
240
300
360
60
120
180
Time (min)
240
300
360
60
120
180
Time (min)
240
300
360
(b) 9
8
7
Log (cfu g1)
Sensory analysis
version 11.5, Chicago, IL, USA) to analyse data.

Signicant dierence (P < 0.05) between treated and
non-treated samples was determined with Tukeys HSD
range test.
6
5
4
3
2
1
0
(c) 9
8
7
Log (cfu g1)
peptone water (0.1%, w v) in stomaching bags (Seward

Medical Co., London, UK) and homogenised with a
stomacher (Model 400; Seward Medical Co.) at a
medium speed for 2 min. Further decimal dilutions
were prepared in peptone water solution and they were
plated out on the appropriate selective media for E. coli
and B. cereus. Escherichia coli was enumerated by
the pour plate technique using TBX (Triptone Bile
X-glucuronide; Merck, Darmstadt, Germany) agar after
the plates were incubated at 44 C for 24 h. Bacillus
cereus was enumerated by the surface spread plating
technique using MYP (Mannitol Egg Yolk Polymixin;
Merck) agar after the plates were incubated at 30 C for
24 h (Mossel et al., 1967). Results were expressed as log
CFU g)1 dry weight.
For identication, suspected colonies of E. coli and
B. cereus on selective agar plates were transferred to
nutrient agar slants, respectively. Escherichia coli was
identied by gram staining, catalase reaction, oxidase
reaction and biochemical tests using API 20E strips
(bioMerieux, Inc., Hazelwood, MO, USA). Bacillus
cereus was identied by gram staining and biochemical
tests including the VogesProskauer (VP) reaction,
degradation of gelatin, starch hydrolysis, indole production, catalase reaction and glucose fermentation as
outlined in the Bacteriological Analytical Manual
(Rhodehamel & Harmon, 1995). These tests showed
that there were no E. coli and B. cereus in noninoculated aked red pepper samples.
6
5
4
3
2
1
0
Figure 1 Comparative effects of ozone concentrations (a) 0.1, (b) 0.5

and (c) 1.0 ppm on population of E. coli in aked red peppers
(, ozonated; s, control).
1659
(a) 9
8
Log (cfu g1)
7
6
5
4
3
2
1
0
60
60
60
120
180
Time (min)
240
300
360
(b) 9
8
7
Log (cfu g1)
(P < 0.05) at 0.1 ppm ozone concentration for

360 min. At 0.5 ppm ozone concentration, E. coli counts
were reduced signicantly (P < 0.05) about 1.7 log
numbers at the end of 360 min of ozone exposure. When
the concentration of ozone was increased from 0.5 to
1.0 ppm for 360 min, the number of E. coli in aked red
pepper samples was signicantly (P < 0.05) reduced 2.0
log numbers. Bacillus cereus counts were not decreased
signicantly at 0.1 and 0.5 ppm ozone concentrations
for 360 min, but increasing the ozone concentration
from 0.5 to 1.0 ppm caused B. cereus population to
reduce signicantly (P < 0.05) about 1.5 log numbers
(Fig. 2). At all ozone concentrations regardless of the
exposure time, the reduction in the number of E. coli
was higher than that of B. cereus. Restaino et al. (1995)
also reported that Gram-negative bacteria were more
sensitive to ozone in pure water than Gram-positive
bacteria.
Similar reductions in B. cereus and E. coli counts
were also observed when the pistachios were treated
with gaseous ozone of 0.1, 0.5 and 1.0 ppm (Akbas &
Ozdemir, 2006a). For instance, at 1.0 ppm ozone
concentration for 360 min of ozonation, B. cereus
and E. coli counts were decreased by c. 2.0 log
numbers in ground pistachios. When the aked red
peppers were treated with 1.0 ppm ozone concentration
for 360 min, B. cereus and E. coli counts were
decreased by 1.5 and 2.0 log numbers, respectively.
The reductions in B. cereus and E. coli counts were
similar for the aked red peppers and ground pistachios. On the contrary, in pistachio kernels and shelled
pistachios, the reduction in B. cereus and E. coli counts
was higher than the ground pistachios when the
pistachios were subjected to 1.0 ppm ozone concentration for 360 min (Akbas & Ozdemir, 2006a). This
showed that surface area was a critical factor for the
eectiveness of the ozone treatment. The reason why
the microbial reduction in aked red peppers and
ground pistachios is less than the pistachio kernels and
shelled pistachios is that the inoculated bacteria could
possibly remain protective in aked red peppers and
ground pistachios. As a result, the oxidative power of
ozone decreases because of the reduced penetration
and the interaction of ozone with the organic materials.
Our results were consistent with the ndings reported
by Naitoh et al. (1989) and Zagon et al. (1992) who
found that surface area was an important factor during
the gaseous ozone treatment of cereal our and ground
pepper.
Ozone concentrations of 1.0, 5.0, 7.0 and 9.0 ppm
were used to reduce B. cereus spores (Fig. 3). The spore
counts of B. cereus were not decreased signicantly
(P 0.05) after the aked red peppers were treated with
1.0 ppm ozone concentration for 360 min. Higher ozone
concentrations (5.09.0 ppm) were then used to obtain
signicant reductions (P < 0.05) in the number of
6
5
4
3
2
1
0
120
180
Time (min)
240
300
360
(c) 9
8
7
Log (cfu g1)
1660
6
5
4
3
2
1
0
120
180
Time (min)
240
300
360
Figure 2 Comparative effects of ozone concentrations (a) 0.1, (b) 0.5

and (c) 1.0 ppm on population of B. cereus in aked red peppers
(, ozonated; s, control).
B. cereus spores. Bacillus cereus spores were decreased

signicantly (P < 0.05) about 1 log number at 5.0 ppm
ozone concentration, whereas the spore counts were
reduced substantially (P < 0.05) 1.5 log numbers at
7.0 ppm ozone concentration for 360 min of ozonation.
However, increasing the concentration of ozone from
7.0 to 9.0 ppm for 360 min, no signicant (P 0.05)
reduction in the spore count was obtained. When
compared with vegetative cells, spores showed greater
resistance to ozone treatments. These results were
consistent with the ndings reported by Kim & Yousef
(2000) and Akbas & Ozdemir (2006a). The greater
Log (cfu g1)
(b) 9
Log (cfu g1)
(a) 9
5
4
3
5
4
3
60
120
180
240
Time (min)
300
360
Log (cfu g1)
(d) 9
Log (cfu g1)
(c) 9
5
4
3
60
60
120
180
240
Time (min)
300
360
5
4
3
1
0
60
120
180
240
Time (min)
300
360
120
180
240
Time (min)
300
360
Figure 3 Comparative effects of ozone concentrations (a) 1.0, (b) 5.0, (c) 7.0 and (d) 9.0 ppm on population of B. cereus spores in aked red
peppers (, ozonated; s, control).
Table 1 Comparative eects of dierent ozone treatments on sensory

attributes of aked red peppers* (1 = dislike very much; 5 = like very
much)
Organoleptic properties
Ozone
concentration
(ppm)
Time
(min)
Taste
Flavour
Appearance
Overall
palatability
0 (Control)
0.1
1.0
5.0
7.0
9.0
0
360
360
360
360
360
5.0a
5.0a
5.0a
4.9a
4.8a
4.8a
5.0a
5.0a
5.0a
4.1b
4.0b
4.0b
5.0a
5.0a
5.0a
4.5b
4.0b
4.0b
5.0a
5.0a
5.0a
4.5b
4.5b
4.5b
*Mean values in the same column are not significantly different (P 0.05).
Values are the mean scores of ten individual panellists.
resistance of spores to ozone is expected to occur

because of the protection of protoplasm by a thick
cortex, multilayered spore coat and exosporangium,
whereas the protoplasm of a vegetative cell is protected
only by a cell wall (Broadwater et al., 1973).
The eect of dierent ozone treatments on organoleptic properties of aked red peppers is shown in
Table 1. No signicant (P 0.05) changes were found
between taste, avour, appearance and overall palatability scores of samples treated with ozone between 0.1
and 1.0 ppm. Slight, but signicant (P < 0.05) changes
were observed in avour, appearance and overall
palatability of aked red pepper samples ozonated
between 5.0 and 9.0 ppm. Our results are consistent
with previous studies that ozone had a negative impact
on organoleptic properties of some food commodities
1661
1662
such as grains (Naitoh et al., 1988) and ground spices

(Zagon et al., 1992) because of high ozone concentrations (>9.0 ppm) and longer exposure times (>6 h).
Akbas & Ozdemir (2006b) also found that ground
pistachios are more susceptible to changes in sensory
quality than pistachio kernels when they were treated
with ozone concentrations higher than 5.0 ppm.
Conclusions
Reductions in the number of E. coli and B. cereus

occurred at low ozone concentrations (1.0 ppm),
whereas decrease of B. cereus spores in aked red
peppers required higher ozone concentrations
(5.0 ppm). There were no signicant (P 0.05)
changes between taste, avour, appearance and overall
palatability scores of samples treated with ozone
1.0 ppm and non-treated samples. Whole red peppers, often heavily contaminated with microorganisms
during harvesting and drying, can be treated with
ozone before processing to reduce the vegetative cell
load, but the treatment is less eective on the spore
load. Results show that ozone treatment is an
alternative for the controlling of microbial agents in
spices when the contamination level is low. If appropriate hygienic measures are not taken from harvest to
processing and the microbial levels on raw red peppers
are high (107108 CFU g)1), additional decontamination methods with or without ozone treatment should
be considered.
Acknowledgments
Financial supports provided by the Government Planning Organization of Turkey and Gebze Institute of
Technology are greatly acknowledged.
References
Akbas, M.Y. & Ozdemir, M. (2006a). Eectiveness of ozone for
inactivation of E. coli and B. cereus in pistachios. International
Journal of Food Science & Technology, 41, 513519.
Akbas, M.Y. & Ozdemir, M. (2006b). Eect of dierent ozone
treatments on aatoxin degradation and physicochemical properties
of pistachios. Journal of the Science of Food and Agriculture, 86,
20992104.
Al-Haddad, K.S.H., Al-Qassemi, R.A.S. & Robinson, R.K. (2005).
The use of gaseous ozone and gas packaging to control populations
of Salmonella infantis and Pseudomonas aeruginosa on the skin of
chicken portions. Food Control, 16, 405410.
Baxter, R. & Holzapfel, W.H. (1982). A microbial investigation of
selected spices, herbs and additives in South Africa. Journal of Food
Science, 47, 570578.
Broadwater, W.T., Hoehn, R.C. & King, P.H. (1973). Sensitivity of
three selected bacterial species to ozone. Applied Microbiology, 26,
391393.
FDA (United States Food and Drug Administration) (2001). Secondary direct food additives permitted in food for human consumption,
nal rule. Federal Register, 66, 3382933830.
Freire, F.C.O & Oord, L. (2002). Bacterial and yeast counts in

Brazilian commodities and spices. Brazilian Journal of Microbiology,
33, 145148.
Garcia, S., Iracheta, F., Galvan, F. & Heredia, N. (2001). Microbiological survey of retail herbs and spices from Mexican markets.
Kader, A.A., Heintz, C.M., Labavitch, J.M. & Rae, H.L. (1982).
Studies related to the description and evaluation of pistachio nut
quality. Journal of the American Society for Horticultural Science,
107, 812816.
Khadre, M.A. & Yousef, A.E. (2001). Sporicidal action of ozone and
hydrogen peroxide: a comparative study. International Journal of
Food Microbiology, 71, 131138.
Khadre, M.A., Yousef, A.E. & Kim, J.G. (2001). Microbiological
aspects of ozone applications in food: a review. Journal of Food
Science, 66, 12421252.
Kim, J.G. & Yousef, A.E. (2000). Inactivation kinetics of food borne
spoilage and pathogenic bacteria by ozone. Journal of Food Science,
65, 521528.
Kim, J.G., Yousef, A.E. & Dave, S.A. (1999). Application of ozone for
enhancing the microbiological safety and quality of foods: a review.
Lee, J.H., Lee, K.T., Kim, M.R. & Sung, T.H. (2004). Eect of
gamma-irradiation on color, pungency, and volatiles of Korean red
pepper powder. Journal of Food Science, 69, 585592.
Liangji, X. (1999). Use of ozone to improve the safety of fresh fruit and
vegetables. Food Technology, 53, 5861.
Lopez-Pedemonte, T.J., Roig-Sagues, A.X., Trujillo, A.J., Capellas,
M. & Guamis, B. (2003). Inactivation of spores of Bacillus cereus in
cheese by high hydrostatic pressure with the addition of nisin or
lysozyme. Journal of Dairy Science, 86, 30753081.
McKnee, L.H. (1995). Microbial contamination of spices and herbs: a
review. Lebensmittel-Wissenschaft und Technologie, 28, 111.
Mossel, D.A.A., Koopmann, M.J. & Jongerious, E. (1967). Enumeration of Bacillus cereus in foods. Applied Microbiology, 15, 650653.
Naitoh, S., Okada, Y. & Sakai, T. (1987). Studies on utilization of
ozone in food preservation: III. Microbicidal properties of ozone on
cereal grains, cereal grain powders, peas, beans, and whole spices.
Journal of Japanese Society of Food Science and Technology, 34,
788793.
Naitoh, S., Okada, Y. & Sakai, T. (1988). Studies on utilization of
ozone in food preservation: V. Changes in microora of ozone
treated cereals, grains, peas, beans, and spices during storage.
Journal of Japanese Society of Food Science and Technology, 35, 69
77.
Naitoh, S., Sawada, Y. & Yamaguchi, N. (1989). Studies on utilization
of ozone in food preservation: eect of ozone treatment on storage
of packaged namamen Japanese row noodle. Journal of Antibacterial
and Antifungal Agents, 17, 517526.
Restaino, L., Frampton, E.W., Hemphill, J.B. & Palnikar, P. (1995).
Ecacy of ozonated water against food-related microorganisms.
Applied and Environmental Microbiology, 61, 34713475.
Rhodehamel, E.J. & Harmon, S.M. (1995). Bacillus cereus. In:
Bacteriological Analytical Manual, 8th edn. Pp. 14.0114.08. Gaithersburg: Association of Ocial Analytical Chemists International.
Schaeer, A.B. & Fulton, M.D. (1933). A simplied method of
staining endospores. Science, 77, 194.
Uijl, C.H. (1992). Beating the bugs. International Food Ingredients, 3,
911.
Zagon, J., Dehne, L.I., Wirz, J., Linke, B. & Boegl, K.W. (1992).
Ozone treatment for removal of microorganisms from spices: an
alternative to ethylene oxide fumigation or irradiation? Results of a
practical study. Bundesgeundheitsblatt, 35, 2023.
Zhao, J. & Cranston, P.M. (1995). Microbial decontamination of
black pepper by ozone and the eect of the treatment on volatile oil
constituents of the spice. Journal of the Science of Food and
Original article
The effect of pectin concentration on viscoelastic and sensory
properties of processed cheese
Ivana Macku,1 Frantisek Bunka,1* Vladimr Pavlnek,2 Petra Lecianova1 & Jan Hrabe1
1 Department of Food Engineering, Tomas Bata University in Zl n, nam. T.G. Masaryka 275, Zl n 762 72, Czech Republic
2 Polymer Centre, Tomas Bata University in Zl n, nam. T.G. Masaryka 275, Zl n 762 72, Czech Republic
(Received 8 October 2007; Accepted in revised form 4 February 2008)
Summary
The eect of pectin addition on viscoelastic properties of model processed cheeses with 40% w w dry matter
and 50% w w fat in dry matter after 42 days of storage at temperature 6 2 C has been investigated using
dynamic oscillation rheometry (plateplate geometry; frequency range 0.150.0 Hz; temperature 20 C). The
role of pectin concentration (0.2%, 0.4%, 0.6% and 0.8% w w) has been studied. Also, the sensory
evaluation of samples has been made to assess cheese appearance, rigidity, spreadability and avour. All
samples with the pectin addition were more rigid and less spreadable compared with processed cheeses
without pectin. With the increasing concentration of pectin the storage (G) and loss (G) moduli rose at the
whole tested frequency range (0.150.0 Hz). Growing pectin content resulted in the decrease in loss tangent
(the nature of gel was changed to more elastic material). The dependence of processed cheese rigidity on
pectin concentration (in range 00.8% w w) was not linear. The appearance and avour were not worse by
pectin addition.
Keywords
Pectin concentration, processed cheese, rheology, sensory analysis, viscoelastic properties.
Introduction
Many recent studies have focused on dairy products

containing various hydrocolloids such as carrageenan,
locust bean gum, xanthan, modied starch or pectin
(Sanchez et al., 2000; Tuinier et al., 2002; Laurent &
Boulenguer, 2003; Verbeken et al., 2004; Spagnuolo
et al., 2005; Vega et al., 2005). These hydrocolloids are
widely used in food industry as stabilising, thickening or
gelling agent whose role is to improve texture (e.g.
enhance viscosity, bind water and or create stable gel
systems) of nal products (Syrbe et al., 1998).
Processed cheese is a dairy product formed by
blending shredded natural cheeses of dierent types
and degrees of maturity with emulsifying agents, and by
heating the blend under a partial vacuum with constant
agitation until a homogenous mass is obtained. Water,
butter, milk protein ingredients (e.g. skim milk powder,
whey powder, coprecipitates, casein, caseinates), rework, vegetables and spices, muscle food ingredients,
avours, colours, salt, hydrocolloids and preservatives
(e.g. nisin, potassium sorbate, calcium sodium propionate) may be added to the blend (Guinee et al., 2004).
There are many factors that inuence the consistency of
*Correspondent: E-mail: bunka@ft.utb.cz
doi:10.1111/j.1365-2621.2008.01734.x
processed cheese, e.g. dry matter content, fat content,

maturity of raw material, amount and types of emulsifying agents used, pH, presence and concentration of
ions (especially calcium, sodium and potassium), processing protocol, using of emulsiers (e.g. mono- and
diacylglycerols), lactose and other sugars concentration,
etc. Also, hydrocolloids (including pectin) are often used
in processed cheese manufacture and signicantly aect
texture of nal products (Marchesseau et al., 1997; Muir
et al., 1997; Swenson et al., 2000; Guinee et al., 2004;
Lee et al., 2004; Mizuno & Lucey, 2005; Shirashoji
et al., 2006; Lu et al., 2007).
Pectin is predominantly linear polysaccharide consisting of a-d-galacturonic acid residues with a small
fraction of rhamnose and side chains created by other
sugars, e.g. galactose and arabinose (Leroux et al., 2003;
Lootens et al., 2003; Tsoga et al., 2004). Pectins can be
divided according to their degree of esterication (DE
the percentage of the carboxyl group on a-d-galacturonic acid esteried with methanol) into high-methoxyl
(HMP; DE > 50%) and low-methoxyl pectin (LMP;
DE < 50%) (Norziah et al., 2001; Tsoga et al., 2004).
The DE strongly aects gelling mechanisms and functional properties of pectin. Whereas LMPs create threedimensional network through calcium bridges between
two dierent pectin chains, HMP gels are formed via
1663
1664
Pectin concentration and processed cheese I. Macku et al.
hydrophobic interactions and hydrogen bonds between

methyl groups of dierent pectin chains (Norziah et al.,
2001; Sato et al., 2004). LMP gels are created in the
presence of calcium and in the wide range of pH
(practically 2.57.0). Rigidity of such gels is considerably inuenced by the amount of calcium present. The
gel strength increases with growing calcium concentration (El-Nawawi & Heikal, 1995; Tibbits et al., 1998;
Lootens et al., 2003). HMP creates stable gels at
pH 3.5 and in the presence of high sugar content
(>55%) (Tsoga et al., 2004; Lofgren & Hermansson,
2007). Hence, HMP serves usually as protein dispersion
stabiliser in products with reduced pH (yoghurt, acidic
milk beverages and milk juice blend). On the other
hand, LMP is preferred for its gelling thickening abilities in either acidic or less acidic products such as nonacid milk desserts (May, 2000; Laurent & Boulenguer,
2003; Matia-Merino & Singh, 2007). The stability and
quality of proteinpectin systems depend on a large
number of factors, e.g. pH, concentration of hydrocolloid, protein and other present substances, ionic environment, actual temperature and thermal history of
system, etc. (Syrbe et al., 1998).
Caseins are major proteins occurring in bovine milk.
Chemically, they are phosphoproteins and include four
main fractions: aS1-, aS2-, b- and j-casein (Lucey, 2002).
These proteins exist in milk in the form of spherical
casein micelles with a diameter of 50300 nm (Schorsch
et al., 1999; Dickinson, 2006). These associations are
highly stable in their natural environment due to outer
hairy layer which is created by the glycomacropeptide
segments of j-casein protruding from the micelle surface
(Schorsch et al., 1999; Spagnuolo et al., 2005). These
stable colloidal particles (micelles) are disrupted either
by enzyme action (e.g. rennin) during natural cheese
production (raw material for processed cheese manufacture) or by acidication in the production of acidic
dairy products such as yogurts (Dickinson, 2006).
Interactions between casein micelles and LMP or
HMP have been widely studied by a number of authors
(Pereyra et al., 1997; Syrbe et al., 1998; Maroziene & de
Kruif, 2000; Tuinier et al., 2002). These interactions can
be attractive or repulsive especially depending on pH.
Attractive interactions appear at pH < 5 and lead to
the adsorption of pectin onto the casein micelle surface
and are of electrostatic nature (Maroziene & de Kruif,
2000; Tuinier et al., 2002). At a neutral pH, when both
biopolymers carry negative charge, pectin is nonadsorbing and repulsive interactions occur (Tuinier
et al., 2002). In available literature, however, there have
been found neither information about interactions
between pectin and non-micellar casein nor papers
dealing with the eect of dierent pectin addition on
consistency of processed cheese.
The objective of this paper was to study the viscoelastic and sensory properties of model processed cheese
with 40% w w dry matter and 50% w w fat in dry

matter as a function of non-esteried pectin concentration. This study could help manufacturers and researchers in understanding the eect of dierent pectin
concentrations on processed cheese quality.
Sample preparation
Model processed cheese (40% w w dry matter and 50%

w w fat in dry matter) was prepared for the analysis of
inuence of non-esteried citrus pectin concentration on
their viscoelastic and sensory properties. Non-esteried
citrus pectin was purchased from Sigma Aldrich, Inc., St
Louis, MO, USA (specication of producer: >90%
w w dry matter and >74% w w of galacturonic acid).
This polysaccharide was applied into the model processed cheese in the dierent target concentrations of
0.2%, 0.4%, 0.6% and 0.8% w w. Ingredients used for
processed cheese manufacture were: (1) Eidamsky Blok
Dutch-type cheese with 30% w w fat in dry matter
and 50% w w dry matter; (2) butter; (3) deionised water
[for keeping the ionic environment at a constant level;
the same condition used, e.g. Shirashoji et al. (2006)]; (4)
commercial emulsifying agents (sodium salts of phosphates and polyphosphates JOHA, Benckiser-Knapsack, Ladenburg, Germany). The highest reached
temperature during production ranged from 92 to
93 C and total cooking time was 1012 min. These
processing parameters and the ingredients were chosen
to imitate the industrial conditions (Piska & Stetina,
2004).
Processed cheese was made using a 2-L capacity
Vorwerk Thermomix TM 31 blender cooker (Vorwerk
& Co. Thermomix; GmbH, Wuppertal, Germany). Lee
et al. (2004) used similar type of equipment in their
study. First, shredded natural cheese and butter (pieces
approximately 2 2 2 cm) were prepared in adequate
amount, placed into the cooker and minced at the
temperature 50 C for 1 min. Further, water and emulsifying agents (2.0% w w) were added to the mixture.
Pectin in powder form was then applied to the blend in
concentrations of 0.2%, 0.4%, 0.6% and 0.8% w w and
the mixture was subjected to the heat treatment of up to
92 C at constant agitation (this target temperature was
kept 1 min). After the thermal processing and packaging
into standard 100 g plastic doses with sealable lids, the
prepared samples were cooled and stored in a refrigerator at 6 2 C. The total amount of manufactured
processed cheese was 800 g per batch.
Three groups of model processed cheeses (i = I, II,
III), prepared using the same processed cheese formulation and technology, were manufactured separately on
dierent days. There were ve batches in each group
of model processed cheeses: control sample without
addition of pectin and processed cheeses with 0.2%,

0.4%, 0.6% and 0.8% w w of pectin. Evaluation of all
samples was performed after 42 days of storage at
6 2 C.
All natural cheese blocks were obtained from the same
producer. Three groups of model processed cheese
slightly diered in the maturity of natural cheese
(between 8 and 10 weeks) which caused dierent initial
rigidity of control samples in all tested groups [see
storage (G) and loss (G) moduli in Table 1]. This
design of experiment is consistent with industrial
practice, where manufacturers process natural cheeses
with dierent degree of maturity. Hence, all ndings
about the inuence of pectin addition on the rheology
and sensory attributes of model processed cheeses were
referred to the control sample of specic group.
eect of pectin concentration on processed cheese

texture was evaluated also by means of the loss tangent
(tan d = G G) and the complex modulus G*.
q

1
G G0 2 G00 2
For statistical evaluation of data the reference frequency of 1 Hz was chosen, which is widely used as
described by Piska & Stetina (2004), Bennett et al.
(2006) and Bunka et al. (2007).
The eect of pectin addition on viscoelastic properties
of processed cheese was analysed using Winters critical
gel theory (Winter & Chambon, 1986). According to
Gabriele et al. (2001), who applied this theory to food,
the complex modulus can be expressed by:
G x AF x1=z
where (AF) is the gel strength and (z) is the number of

structure units interacting with one another in a threedimensional network (the interaction factor).
Chemical analysis
Each sample was characterised by dry matter content

and pH. Dry matter content was determined according
to ISO Standard No. 5534 (2004). pH of model
processed cheese was measured three times by pH meter
with glass electrode.
Sensory analysis
Model processed cheese was evaluated by a panel of

2024 selected assessors (employees and students of
Department of Food Engineering) trained according to
ISO Standard No. 8586-1 (1993). This evaluation was
performed within the sensory laboratory equipped with
sensory booths in accordance with ISO Standard No.
8589 (1988). Samples were coded and served at the room
temperature (22 2 C). Seven-point hedonic scale
was used for assessment of cheese appearance, rigidity,
spreadability and avour.
Rheological measurement
Rheological measurements were made in dynamic

oscillation mode using a controlled stress Bohlin
rheometer (Bohlin GEMINI, Malvern Instruments
Ltd, Malvern, UK) with parallel plates geometry
(40 mm diameter). The processed cheese sample was
carefully put between the plates (at the temperature of
20 C) and the exposed edge was coated with low
viscosity silicone oil to avoid sample humidity loss.
Measurements of the storage (G) and loss (G) moduli
were performed in the linear viscoelastic region (the
amplitude of shear stress 40 Pa) and in the frequency
sweep mode at the frequency range 0.150.0 Hz. The
Statistical data analysis
Results of basic chemical analysis, dynamic oscillation

rheometry and sensory analysis were statistically evaluated by means of non-parametric analysis of variance
Table 1 The values of storage modulus (G), loss modulus (G) and loss tangent (tan d) for the reference frequency of 1 Hz in tested processed
cheeses (Groups IIII)
Group of processed cheese*
I
Amount of added
pectin (% w w)
None
0.2
0.4
0.6
0.8
II
G [Pa]
2511
5367
8862
12 400
14 135
G [Pa]
112
367b
26c
226d
92e
1901
2837
4023
5184
5746
Tan d [)]
a
67
175b
48c
145d
49e
0.757
0.529
0.454
0.418
0.407
III
G [Pa]
4494
10 245
10 890
14 230
14 370
G [Pa]
457
141b
67c
128d
99d
2812
5449
5530
5891
6567
Tan d [)]
a
184
66b
21c
123d
108e
0.626
0.532
0.489
0.414
0.457
G [Pa]
2863
7378
9018
11 205
14 480
G [Pa]
20
139b
107c
244d
322e
2209
4236
4699
5386
6272
Tan d [)]
a
30
32b
17c
141d
185e
0.771
0.574
0.521
0.481
0.433
*Storage (G) and loss (G) moduli are presented by mean standard deviation; tan d = G G. Mean values having the same superscript letter in each
column are not significantly different (P 0.05).
1665
(KruskalWallis test) or Wilcoxon test (Agresti, 1984).

Dierences among comparisons had to achieve
P < 0.05 to show signicance. The dependency of
complex modulus values (G*) on the pectin concentration were also evaluated using regression analysis (least
squares method). Linear (y = a1 + a2x), quadratic
(y = a1 + a2x + a3x2) and cubic (y = a1 + a2x
+ a3x2 + a4x3) models were chosen for tting [where
y represented values of the complex modulus (G*), x
expressed the pectin concentration and a1, a2, a3, a4 were
regress coecients for tested model]. Parameters (AF)
and (z) in eqn 2 were calculated by least squares method.
Chemical analysis
Viscoelastic and sensory properties of model processed

cheeses with pectin concentrations of 0.2%, 0.4%, 0.6%
and 0.8% w w and control samples without pectin were
analysed. Results obtained by basic chemical analysis
showed insignicant dierences (P 0.05) in pH and the
dry matter content among samples within the group.
The dry matter content in Group I ranged from 41.86%
to 42.90% w w, in Group II from 42.57% to 43.72%
w w and in Group III from 42.39% to 42.95% w w.
Measured values of pH in tested groups were as follows:
5.845.96 (Group I), 5.645.72 (Group II) and 5.915.94
(Group III). The processed cheese consistency and the
eciency of pectin are substantially inuenced by the
dry matter content and pH of system (Marchesseau
et al., 1997; Sharma et al., 1998; Lee et al., 2004).
Hence, practically similar values of the dry matter
content and pH allowed analysing of the inuence of
dierent pectin concentration on viscoelastic and sensory properties of model processed cheese in three tested
groups.
Rheological analysis
The small amplitude oscillatory shear test (determination of storage G and loss G moduli) is nowadays
useful and usual tool for evaluation of food gels
(Gabriele et al., 2001). This technique is widely used
for studying of viscoelastic properties of dairy products
including processed cheeses as demonstrated, e.g. works
of Lucey (2002), Lee et al. (2004), Shirashoji et al.
(2006), Lu et al. (2007). Table 1 reects that the
addition of pectin in all tested concentrations (0.2%,
0.4%, 0.6% and 0.8% w w) caused signicant growth in
the storage (G) and loss (G) moduli of model processed
cheeses compared to control samples; this increase was
observed in all used groups (P < 0.05). The similar
results described also Swenson et al. (2000) who
observed the increase of processed cheese rmness due
to hydrocolloid addition (gelatin, carrageenan, locust
bean gum and guar gum) compared with control

samples. The storage (G) and loss (G) moduli values
of control samples in this paper are in agreement with
data published by Lee & Klostermeyer (2001), Piska &
Stetina (2004) and Bunka et al. (2007).
Figure 1 shows the dependence of the storage (G) and
loss (G) moduli on frequency for products in Group I.
According to this gure, with the increasing concentration of pectin the storage (G) and loss (G) moduli of
model processed cheeses rose at the whole tested
frequency range (0.150.0 Hz). The similar increasing
trends were obtained as well for samples within Group
II and Group III (data are not graphically presented).
Statistical analysis of storage (G) and loss (G) moduli
values at the reference frequency 1 Hz (see Table 1)
proved signicant rise of these moduli in most batches
as a result of increasing pectin concentration
(P < 0.05). Generally, the higher was pectin concentration, the higher strength of gel was obtained. Table 1
also shows that the growing pectin concentration caused
the decrease in the loss tangent (tan d) of most tested
batches. The decline of loss tangent (tan d) pointed that
the rise in gel rigidity was caused by the increase of gel
elasticity. These results indicate that the nature of gel is
changed to rmer and more elastic material due to
increasing amount of pectin.
Table 2 presents the values of the gel strength (AF)
and the interaction factor (z) calculated from the
dependency of the complex modulus (G*) on angular
frequency (x) using eqn 2. The growing pectin concentration caused signicant rising (P < 0.05) of the gel
Storage G and loss G moduli (Pa)
1666
104
103
0.1
10
Frequency (Hz)
Figure 1 Dependence of the storage G (lled symbols) and loss G
(open symbols) moduli on frequency for products in Group I; pectin
amount 0% w w (r)), 0.2% w w ( 4), 0.4% w w (ds), 0.6% w w
(.5), 0.8% w w ( h).
Table 2 The values of gel strength (AF) and interaction factor (z) in tested processed cheeses (Groups IIII)
Group of processed cheese*
I
Amount of added
pectin (% w w)
None
0.2
0.4
0.6
0.8
II
AF [Pa s1 z]
3136
5990
9565
13 192
14 997
AF [Pa s1 z]
z
a
142
385b
64c
282d
128e
2.67
3.29
3.85
4.10
4.26
III
0.05
0.03b
0.03c
0.04d
0.01e
5246
11 330
11 885
15 132
15 497
AF [Pa s1 z]
z
a
462
861b
996b
405c
213c
2.95
3.55
3.67
3.83
3.87
0.13
0.10b
0.19b
0.14c
0.15c
3595
8371
9956
12 150
15 442
z
a
51
129b
102c
578d
633e
2.64
3.23
3.50
3.71
4.03
0.08a
0.05b
0.03c
0.06d
0.04e
*Gel strength (AF) and interaction factor (z) are presented by mean standard deviation. Mean values having the same superscript letter in each column
are not significantly different (P 0.05).
strength of model processed cheese (expressed as AF).

Similar behaviour, i.e. growth of gel rigidity due to
increasing pectin amount in samples, was also observed
by El-Nawawi & Heikal (1995) who found out that the
strength of LMP gels (prepared from water, pectin,
sucrose and calcium ions) increased with rising pectin
concentration.
The increasing pectin addition promoted the enhancement (P < 0.05) of the interaction factor (z) in tested
batches of processed cheese (see Table 2). According to
Gabriele et al. (2001) and Mart nez-Ruvalcaba et al.
(2007), the higher z-value means the higher number of
interactions (degree of connectivity) in three-dimen-
sional network. Thus, the growth of gel rigidity (gel

strength) and the increase of gel elasticity due to pectin
addition (described above) could be explained by rising
number of interactions in matrix of processed cheese.
Values of complex modulus (G*) at the reference
frequency 1 Hz (calculated according to eqn 1) also
showed the rising trend (increasing rigidity of system)
with the growing amount of added pectin (see Table 3).
Regression analysis of the dependence of the complex
modulus values (G*) on the pectin concentration for
three dierent models was carried out. Linear, quadratic
and cubic models were chosen. Further, the coecients
of determination (r2) for all three groups of the model
Table 3 The values of mean of complex modulus (G*) for the reference frequency of 1 Hz in tested processed cheeses (Groups IIII) and the values
of regress coecients (ai) and coecient of determination (r2) for chosen models (linear, quadratic and cubic) expressing changes in G* depending
on amount of added pectin*
Group of processed cheese
Amount of added
pectin (% w w)
G* [Pa]
None
0.2
0.4
0.6
0.8
Y = a 1 + a 2x
y = a1 + a2x + a3x2
y = a1 + a2x + a3x2 + a4x3
II
a
3150 146
6071 407b
9732 44c
13 445 262d
15 265 106e
a1 = 3238.0
a2 = 15 758.0
r2 = 0.9851
a1 = 2709.4
a2 = 19 723.0
a3 = )4625.2
r2 = 0.9900
a1 = 3185.4
a2 = 9302.9
a3 = 19 371.0
a4 = )27 622.0
r2 = 0.9981
III
a
5302 486
11 605 912b
12 125 1082b
15 400 411c
15 800 285c
a1 = 7087.9
a2 = 12 396.0
r2 = 0.8519
a1 = 5794.8
a2 = 25 327.0
a3 = )16164.0
r2 = 0.9329
a1 = 5503.9
a2 = 35 750.0
a3 = )52 521.0
a4 = 30 297.0
r2 = 0.9446
3616 34a
8508 136b
10 170 99c
12 435 544d
15 780 651e
a1 = 4450.6
a2 = 14 128.0
r2 = 0.9657
a1 = 4094.8
a2 = 17 686.0
a3 = )4447.3
r2 = 0.9710
a1 = 3663.9
a2 = 33 126.0
a3 = )58 310.0
a4 = 44 885.0
r2 = 0.9935
*In the regress equation y expressed values of complex modulus (G*) and x expressed amount of added pectin.
The complex modulus (G*) is presented by mean standard deviation. Mean values having the same superscript letter in each column are not
significantly different (P 0.05).
1667
1668
processed cheese and for all three calculated models

were expressed (see Table 3). The comparison of these
coecients proved that quadratic or cubic models
describe the dependence of G* values on the pectin
concentration more accurately and at a closer tting
than the linear model did. Results of Wade (1957) and
Sharma et al. (1998) indicated that the curve expressing
the dependence of gel strength on the increasing pectin
concentration can be non-linear.
Sensory analysis
Results obtained by the sensory analysis showed that

appearance and avour of most model processed cheese
were not signicantly changed with the increasing pectin
concentration (P 0.05) (see Table 4), i.e. neither
appearance nor avour was worsen due to used pectin
additions. Model processed cheeses with 0.2% w w of
pectin in Groups I and II were even evaluated as
signicantly better in appearance and avour
(P < 0.05) compared with the other samples in the
group. This fact may be explained by better shine of
these cheeses in comparison with control samples and
products with higher pectin concentrations.
All model processed cheeses with pectin additions
were signicantly more rigid and less spreadable compared with control samples (P < 0.05). Generally,
Table 4 also shows that rigidity increased and spreadTable 4 Results (expressed as median) of the sensory analysis of tested
processed cheese in Groups IIII*

Sensory evaluation (median)
Group of Amount of
processed added pectin
Appearance Rigidity Spreadability Flavour
cheese
(% w w)
I
II
III
None
0.2
0.4
0.6
0.8
None
0.2
0.4
0.6
0.8
None
0.2
0.4
0.6
0.8
2a
1b
2a
2a
2a
2a
1b
2a
2a
2a
2a
2a
2a
2a
2a
5a
4b
3c
2d
2d
4a
3b
3b
2c
2c
5a
4b
3c
2d
2d
5a
4b
3c
2d
2d
4a
3b
3b
2c
2c
5a
4b
4b
3c
3c
2a
2a
2a
3a
3a
3a
2a
2a
3a
3a
3a
2a
3a
3a
3a
*Used hedonic scales: Appearance: 1 excellent to 7 very poor,

unacceptable. Rigidity: 1 very stiff to 7 very soft. Spreadability: 1
spreadability is impossible to 4 optimal spreadability to 7 almost liquid
sample. Flavour: 1 excellent to 7 very poor, unacceptable.
Median values having the same superscript letter in each column are
not significantly different (P 0.05); each of groups was evaluated
separately.
ability decreased with the growing amount of pectin

(P < 0.05).
Selected assessors noted signicant rise in rigidity and
reduction in spreadability of model processed cheeses
with the growing pectin concentration in range of
00.6% w w within Group I (P < 0.05). The increase
in rmness of product observed by dynamic oscillation
rheometry (see Fig. 1, Tables 13) was also distinguishable by selected assessors.
The rise in rigidity and the decline in spreadability
with the growing pectin concentration (in range of
00.6% w w) were obtained also in model processed
cheeses within Group II and Group III (P < 0.05).
Nevertheless, there were not detected signicant dierences in rigidity and spreadability between samples with
0.2% and 0.4% w w of pectin in Group II. This nding
corresponded with results obtained by dynamic oscillation rheometry [see values of complex modulus (G*) in
Table 3]. Model processed cheeses with 0.2% and 0.4%
w w of pectin in Group III did not signicantly dier in
spreadability (P 0.05).
In Groups I and III dynamic oscillation rheometry
(see Tables 13) detected signicant growth of rmness
in model processed cheese with 0.8% w w of pectin
compared with product with 0.6% w w (P < 0.05). On
the other hand, sensory analysis (see Table 4) did not
indicate any dierences in rigidity or spreadability
between processed cheese with 0.6% and 0.8% w w of
pectin in Groups I and III (P 0.05). This fact can be
explained by formation of very strong gels whose
dierences in rigidity and spreadability are sub-threshold for selected assessors in term of dierence threshold.
Problems with dierence thresholds for higher intensities of textural attributes were reported also by Hough
et al. (1996) who studied correlations between sensory
and instrumental methods for cheese assessment. Hence,
for higher amounts of pectin in processed cheeses
(>0.6% w w) it can be assumed that dierences in
rigidity and spreadability between close pectin concentrations will not be unambiguously determined only by
sensory analysis.
Conclusion
The inuence of pectin addition on viscoelastic and

sensory properties of model processed cheese with 40%
w w dry matter and 50% w w fat in dry matter was
investigated. All used pectin concentrations (0.2%, 0.4%,
0.6% and 0.8% w w) caused signicant increase in
rigidity and decrease in spreadability of products compared with control samples in all tested groups. The eect
of pectin on the viscoelastic properties of model processed cheese depended on used polysaccharide concentration. In general, the rigidity of products increased with
the growing pectin concentration. However, the dependence of rmness on pectin amount was not linear.
Sensory analysis detected signicant increase in rigidity and spreadability in most products with rising
concentration of pectin in range of 00.6% w w. The
dierences in rigidity and spreadability between model
processed cheeses containing 0.6% and 0.8% w w of
pectin were sub-threshold for selected assessors and
were not detected by sensory analysis. Hence, the
combination of instrumental methods and sensory
analysis is useful for the quality control of processed
cheeses. Appearance and avour of model processed
cheese were not negatively modied by pectin addition.
The pectin usage for desired modication of texture
properties of processed cheeses is possible (in appropriate amount) without risks of worsening appearance and avour which are very important for
consumer.
Acknowledgment
This work was kindly supported by a project of Czech

Ministry of Education, Youth and Sports (Grant No.
MSM 7088352101).
References
Agresti, A. (1984). Analysis of Ordinal Categorical Data. New York:
John Wiley & Sons.
Bennett, R.J., Trivedi, D., Hemar, Y., Reid, D.C.W., Illingworth, D. &
Lee, S.K. (2006). The eect of starch addition on the rheological and
microstructural properties of model processed cheese. The Australian Journal of Dairy Technology, 61, 157159.
Bunka, F., Pavl nek, V., Hrabe, J., Rop, O., Janis , R. & Krejc , J.
(2007). Eect of 1-monoglycerides on viscoelastic properties of
processed cheeses. International Journal of Food Properties, 10, 819
828.
Dickinson, E. (2006). Structure formation in casein-based gels, foams,
and emulsions. Colloids and Surfaces A: Physicochemical and
Engineering Aspects, 288, 311.
El-Nawawi, S.A. & Heikal, Y.A. (1995). Factors aecting the
production of low-ester pectin gels. Carbohydrate Polymers, 26,
189193.
Gabriele, D., de Cindio, B. & DAntona, P. (2001). A weak gel model
for foods. Rheologica Acta, 40, 120127.
Guinee, T.P., Caric, M. & Kalab, M. (2004). Pasteurized processed
cheese and substitute imitation cheese products. in: Cheese: Chemistry, Physics and Microbiology. Volume 2: Major Cheese Groups
(edited by P.F. Fox, P.L.H. McSweeney & T.P. Cogan). Pp. 349
394. London and New York: Elsevier Applied Science.
Hough, G., Califano, A.N., Bertola, N.C. et al. (1996). Partial least
squares correlations between sensory and instrumental measurements of avour and texture for Reggianito grating cheese. Food
Quality and Preference, 7, 4753.
ISO Standard No. 5534 (2004). Cheese and Processed Cheese
Determination of the Total Solids Content (Reference Method).
Geneva: International Organization for Standardization.
ISO Standard No. 8586-1 (1993). Sensory Analysis General Guidance
for the Selection, Training and Monitoring of Assessors Part 1:
Selected Assessors. Geneva: International Organization for Standardization.
ISO Standard No. 8589 (1988). Sensory Analysis General Guidance
for the Design of Test Rooms. Geneva: International Organization
for Standardization.
Laurent, M.A. & Boulenguer, P. (2003). Stabilization mechanism of

acid dairy drinks (ADD) induced by pectin. Food Hydrocolloids, 17,
445454.
Lee, S.K. & Klostermeyer, H. (2001). The eect of pH on the
rheological properties of reduced-fat model processed cheese
spreads. Lebensmittel Wissenschaft und Technologie, 34, 288292.
Lee, S.K., Anema, S. & Klostermeyer, H. (2004). The inuence of
moisture content on the rheological properties of processed cheese
spreads. International Journal of Food Science and Technology, 39,
763771.
Leroux, J., Langendor, V., Schick, G., Vaishnav, V. & Mazoyer, J.
(2003). Emulsion stabilizing properties of pectin. Food Hydrocolloids, 17, 455462.
Lofgren, C. & Hermansson, A.M. (2007). Synergistic rheological
behaviour of mixed HM LM pectin gels. Food Hydrocolloids, 21,
480486.
Lootens, D., Capel, F., Durand, D., Nicolai, T., Boulenguer, P. &
Langendor, V. (2003). Inuence of pH, Ca concentration, temperature and amidation on the gelation of low methoxyl pectin. Food
Lu, Y., Shirashoji, N. & Lucey, J.A. (2007). Rheological, textural and
melting properties of commercial samples of some of the dierent
types of pasteurized processed cheese. International Journal of Dairy
Lucey, J.A. (2002). Formation and physical properties of milk protein
gels. Journal of Dairy Science, 85, 281294.
Marchesseau, S., Gastaldi, E., Lagaude, A. & Cuq, J.L. (1997).
Inuence of pH on protein interactions and microstructure of
process cheese. Journal of Dairy Science, 80, 14831489.
Maroziene, A. & de Kruif, C.G. (2000). Interaction of pectin and
casein micelles. Food Hydrocolloids, 14, 391394.
Mart nez-Ruvalcaba, A., Chornet, E. & Rodrigue, D. (2007). Viscoelastic properties of dispersed chitosan xanthan hydrogels. Carbohydrate Polymers, 67, 586595.
Matia-Merino, L. & Singh, H. (2007). Acid-induced gelation of milk
protein concentrates with added pectin: eect of casein micelle
dissociation. Food Hydrocolloids, 21, 765775.
May, C.D. (2000). Pectins. In: Handbook of Hydrocolloids (edited by
G.O. Phillips & P.A. Williams). Pp. 169188. Cambridge and Boca
Raton: Woodhead Publishing Limited and CRC Press LLC.
Mizuno, R. & Lucey, J.A. (2005). Eects of emulsifying salts on the
turbidity and calcium-phosphate-protein interactions in casein
micelles. Journal of Dairy Science, 88, 30703078.
Muir, D.D., Williams, S.A.R., Tamime, A.Y. & Shenana, M.E. (1997).
Comparison of the sensory proles of regular and reduced-fat
commercial processed cheese spreads. International Journal of Food
Norziah, M.H., Kong, S.S., Karim, A.A. & Seow, C.C. (2001). Pectin
sucroseCa2+ interactions: eects on rheological properties. Food
Pereyra, R., Schmidt, K.A. & Wicker, L. (1997). Interaction and
stabilization of acidied casein dispersions with low and high
methoxyl pectins. Journal of Agricultural and Food Chemistry, 45,
34483451.
Piska, I. & Stetina, J. (2004). Inuence of cheese ripening and rate of
cooling of the processed cheese mixture on rheological properties of
processed cheese. Journal of Food Engineering, 61, 551555.
Sanchez, C., Zuniga-Lopez, R., Schmitt, C., Despond, S. & Hardy, J.
(2000). Microstructure of acid-induced skim milk-locust bean gumxanthan gels. International Dairy Journal, 10, 199212.
Sato, Y., Kawabuchi, S., Irimoto, Y. & Miyawaki, O. (2004). Eect of
water activity and solvent-ordering on intermolecular interaction of
high-methoxyl pectins in various sugar solutions. Food Hydrocolloids, 18, 527534.
Schorsch, C., Jones, M.G. & Norton, I.T. (1999). Thermodynamic
incompatibility and microstructure of milk protein locust bean
gum sucrose systems. Food Hydrocolloids, 13, 8999.
1669
1670
Sharma, S.K., Liptay, A. & Maguer, M.L. (1998). Molecular characterization, physico-chemical and functional properties of tomato
fruit pectin. Food Research International, 30, 543547.
Shirashoji, N., Jaeggi, J.J. & Lucey, J.A. (2006). Eect of trisodium
citrate concentration and cooking time on the physicochemical
properties of pasteurized process cheese. Journal of Dairy Science,
89, 1528.
Spagnuolo, P.A., Dalgleish, D.G., Go, H.D. & Morris, E.R. (2005).
Kappa-carrageenan interactions in systems containing casein
micelles and polysaccharide stabilizers. Food Hydrocolloids, 19,
371377.
Swenson, B.J., Wendor, W.L. & Lindsay, R.C. (2000). Eects of
ingredients on the functionality of fat-free process cheese spreads.
Syrbe, A., Bauer, W.J. & Klostermeyer, H. (1998). Polymer science
concepts in dairy systems an overview of milk protein and food
hydrocolloid interaction. International Dairy Journal, 8, 179193.
Tibbits, C.W., MacDougall, A.J. & Ring, S.G. (1998). Calcium
binding and swelling behaviour of a high methoxyl pectin gel.
Carbohydrate Research, 310, 101107.
Tsoga, A., Richardson, R.K. & Morris, E.R. (2004). Role of cosolutes
in gelation of high-methoxy pectin. Part 1. Comparison of sugars and
polyols. Food Hydrocolloids, 18, 907919.
Tuinier, R., Rolin, C. & de Kruif, C.G. (2002). Electrosorption of
pectin onto casein micelles. Biomacromolecules, 3, 632638.
Vega, C., Dalgleish, D.G. & Go, H.D. (2005). Eect of j-carrageenan
addition to dairy emulsions containing sodium caseinate and locust
bean gum. Food Hydrocolloids, 19, 187195.
Verbeken, D., Thas, O. & Dewettinck, K. (2004). Textural properties
of gelled dairy desserts containing j-carrageenan and starch. Food
Wade, P. (1957). Relationship of the strength of high solids pectin gels
to the concentration and jelly-forming capacity of the pectin present.
Nature, 180, 10671068.
Winter, H.H. & Chambon, F. (1986). Analysis of linear viscoelasticity
of a crosslinking polymer at the gel point. Journal of Rheology, 30,
367382.
Original article
Sensory shelf life of butterhead lettuce leaves in active and passive
modified atmosphere packages
Gaston Ares,1,2* Claudia Lareo2 & Patricia Lema2
1 Seccion Evaluacion Sensorial, Departamento de Ciencia y Tecnolog a de Alimentos, Facultad de Qu mica, Universidad de la Republica, Gral,
Flores 2124, CP 11800, Montevideo, Uruguay
2 Instituto de Ingenier a Qu mica, Facultad de Ingenier a, Universidad de la Republica, Julio Herrera y Reissig 565, C.P. 11300, Montevideo,
Uruguay
(Received 3 December 2007; Accepted in revised form 22 January 2008)
Summary
The objective of the present work was to study the inuence of both passive and active modied atmosphere
packaging on sensory shelf life of butterhead lettuce leaves, stored at 5 and 10 C. Results showed that the
ecacy of active modied atmosphere with respect to passive modied atmosphere depended on the storage
temperature considered. When stored at 10 C, despite being subjected to dierent package atmosphere
composition, lettuce leaves in active and passive modied atmosphere showed the same sensory deterioration
rate and sensory shelf life. On the contrary, when stored at 5 C, lettuce leaves in active modied atmosphere
packages showed lower deterioration rate and higher sensory shelf life than those in passive modied
atmosphere. Sensory shelf life of lettuce in modied atmosphere packages was estimated using criteria
determined according to consumers rejection to purchase percentage, which consisted on an improvement
over more arbitrary criteria presented in most studies.
Keywords
Butterhead lettuce, minimally processed vegetables, modied atmosphere, shelf life.
Introduction
In recent years, minimally processed vegetables have

gained great acceptance by consumers (Zhou et al.,
2004), which could be mainly attributed to the desire of
consumers for fresh products with reduced preparation
time (Odumeru et al., 2002). Lettuce is one of the most
important ready-to-use products (Odumeru et al., 2002;
Zhou et al., 2004).
Colour and appearance are critical quality aspects for
shoppers when selecting fresh fruits and vegetables
(Ragaert et al., 2004). Therefore, the shelf life of lettuce
can be dened as the length of time for which lettuce can
maintain an appearance that appeals to the consumer
(Zhou et al., 2004).
Modied atmosphere packaging of fresh produce
relies on modication of the atmosphere inside the
package, achieved by the natural interplay between the
respiration or the product and the transfer of gases
through the package, which leads to an atmosphere
richer in CO2 and poorer in O2. This atmosphere can
potentially reduce respiration rate, decay and physiological changes (Fonseca et al., 2002). Both active
and passive modied atmosphere packaging can be
*Correspondent: E-mail: gares@fq.edu.uy
employed. Active modication occurs by the displacement of air in the package using a desired mixture of
gases. On the contrary, passive modication occurs
when the product is packaged under air using a selected
lm type, and therefore the atmosphere develops only as
a consequence of the product respiration and the
diusion of gases through the lm (Fonseca et al.,
2002; Farber et al., 2003).
Modied atmosphere packages should be carefully
designed as a system incorrectly designed may be
ineective or even shorten the shelf life of the product.
Normally, in modied atmosphere packages, the concentration of O2 is kept low (15%) to reduce the
respiration rate of fruits and vegetables, which prolongs
the self life of the products (Fonseca et al., 2002).
However, at extremely low O2 levels (<1%), anaerobic
respiration can occur, resulting in tissue destruction and
the production of substances that contribute to oavours and o-odours, as well as the potential for
growth of foodborne pathogens such as Clostridium
botulinum (Farber et al., 2003). The inhibition of CO2
on respiration rate depends on the type and developmental stage of the commodity, CO2 concentrations and
time of exposure (Fonseca et al., 2002). In modied
atmosphere packages, the excessive accumulation of
CO2 can cause physiological injuries to the product,
such as browning (Fonseca et al., 2002).
doi:10.1111/j.1365-2621.2008.01736.x
1671
1672
Sensory shelf life of butterhead lettuce leaves G. Ares et al.
The objective of the present work was to study the

inuence of both passive and active modied atmosphere packaging on sensory shelf life of butterhead
lettuce leaves, stored at 5 and 10 C.
Lettuce and storage conditions
Butterhead lettuces (Lactuca sativa L., cv Wang) were

obtained from a local farm near Montevideo (Uruguay).
Within 24 h after harvest, lettuces were transported to
the School of Engineering in Montevideo, Uruguay.
Outer damaged and yellowed leaves, roots and stems
were removed. The remaining leaves were washed with
cold water, treated with chlorinated water [200 parts per
million (ppm) total chlorine] for 10 min, and then rinsed
with water. Lettuce were dried by centrifugation in a
kitchen manual basket type centrifuge for 1 min.
Approximately 70 5 g whole lettuce leaves were
selected at random and placed in 30 40 cm bags of
common polypropylene (PP) (40 lm thickness). Bags
were lled with air (passive modied atmosphere) and
an initial gas mixture of 5% O2 and 2.5% CO2 (active
modied atmosphere), and sealed using a Supervac
GK105 1 packaging machine (Supervac, Wien, Austria). The composition of the gas mixture was selected
according to the published data (Varoquaux et al., 1996;
Beaudry, 1999). Packaged samples were stored at
5 0.5 C and 10 0.5 C. Samples stored at 5 C
were evaluated after 0, 14, 21, 28, 35, 42 and 49 days,
whereas samples stored at 10 C were evaluated at days
0, 7, 10, 14, 17 and 21.
A preliminary test was performed to identify those

defects most likely to appear because of prolonged
storage. Four samples of lettuce with dierent deterioration times (0, 7, 10 and 17 days) at 15 C were
presented to the assessors, who wrote down the
descriptors that made the appearance of those samples
dierent. Through open discussion with the panel
leader, the assessors agreed on the appearance descriptors that best dierentiated the stored samples from the
fresh one. These descriptors were wilting appearance,
presence of dark and necrotic stains on the leaf surface
and presence of browning on the midribs. Once the
descriptors were selected, assessors were trained by
measuring samples stored at 5 C with dierent storage
times using 10-cm unstructured intensity scales.
At each storage time and for each storage condition, a
leaf of lettuce, randomly selected, was presented to the
assessors in a closed odourless plastic container, labelled
with three-digit random numbers, at room temperature.
Assessors had to evaluate the following attributes:
wilting appearance, presence of dark and necrotic stains
on the leaf surface, presence of browning on the midribs,
presence of browning on the cut surfaces and o-odour.
For scoring, 10-cm unstructured scales anchored with
nil and high were used.
A balanced complete block design was carried out for
duplicate evaluation of the samples. Testing was carried
out in a sensory laboratory that was designed in
accordance with ISO 8589: 1988. Evaluations were
performed under articial daylight-type illumination,
temperature control (between 22 and 24 C) and air
circulation.
Consumer panel
Package atmosphere composition
Changes in package atmosphere composition were

evaluated by using a gas analyzer. O2 and CO2 concentrations were determined pumping 15 mL gas sample
directly from the package through a needle and a silicon
tube, and analysed using a Servomex 1450C Food
Package Analyzer (Servomex, Crowborough, Sussex,
UK).
Weight loss
Weight loss was determined by weighting the content of

the packages before and after the storage period and
expressed as the percentage of weight loss with respect
to the initial weight.
At each storage time, a consumer study was carried out

using forty people who consumed lettuce. Their ages
ranged between eighteen and fty, and they were
approximately 50% female and 50% male. Consumers
were recruited among students and workers from the
School of Chemistry in Montevideo, Uruguay.
Each consumer received one leaf of lettuce, for each
storage condition, in a closed odourless plastic container, labelled with three-digit random numbers. For
each sample, consumers had to evaluate its appearance
and respond yes or no to the question Imagine you
are in a supermarket, you want to buy a minimallyprocessed lettuce, and you nd a package of lettuce with
leaves like this, would you normally buy it?.
Data analysis
Trained sensory panel
Samples were evaluated by a panel of six assessors, who

had previous experience in the evaluation of fresh
vegetables.
Analysis of variance
For package atmosphere composition, respiration rate,

weight loss and instrumental data at each temperature, an analysis of variance (anova) was performed
sensory attributes as explanatory variable (Gambaro

et al., 2006). The following equation was used:
considering type of modied atmosphere, storage time

and their interaction as sources of variation. An anova
for each temperature was performed on the sensory data
obtained, considering assessor, type of modied atmosphere, storage time and the interaction type of modied
atmosphere storage time as sources of variation. When
the eects were signicant, mean ratings for each
condition and storage time were calculated. Honestly
signicant dierences were calculated using Tukeys test.
Dierences were considered signicant when P 0.05.

Logistic : percentofrejection a
1 ecXd
; 2
where X is each of the evaluated sensory attributes, and

a, b, c and d are the regression constants.
All statistical analyses were carried out using Genstat
Discovery Edition 2 (VSN International, Oxford, UK).
Sensory attribute development modelling
In order to model the development of the evaluated

sensory attributes as a function of storage time, the
following equation that assumes rst-order reaction rate
was used:
X X0 expkT t;
Atmosphere composition within packages
Changes in O2 and CO2 concentrations inside the

packages are shown in Fig. 1. For each storage temperature, anova showed that the eects of the type of
modied atmosphere, storage time and their interaction
were highly signicant (P < 0.001).
As shown in Fig. 1, for passive modied atmosphere
packages, O2 concentration signicantly decreased
(P < 0.05) with increasing storage time, while CO2
concentration increased (P < 0.05). This trend could be
attributed to lettuce respiration, which consumes O2 and
produces CO2. O2 and CO2 lm permeability did not
compensate the change in the concentration of these
gases.
On the contrary, for active and passive modied
atmosphere packages, O2 concentration signicantly
where X is the value for the sensory attribute X at time t;

X0, the value for the sensory attribute X at time t = 0;
kT, the reaction rate constant at storage temperature T;
t, the storage time. The equations parameters were
estimated using non-linear regression.
Regression analysis
A logistic regression was carried out considering percent

of rejection (calculated as the proportion of consumers
who answered no when asked if they would normally
buy a package of lettuce containing leaves like the ones
they saw) as dependent variable and the evaluated
(a)
Passive 5 C
Passive 10 C
Active 5 C
Active 10 C
Passive 5 C
Active 5 C
(b)
6
Passive 10 C
Active 10 C
CO2 concentration (%)
Concentracin de O2 (%)
20
15
10
5
0
10
20
30
40
Storage time (days)
50
0
0
10
20
30
Storage time (days)
40
50
Figure 1 (a) O2 and (b) CO2 concentrations inside active and passive modied atmosphere packages during storage at 5 and 10 C. Vertical bars
represent Tukeys honestly signicant dierences.
1673
1674
increased (P < 0.05) with storage time, which could be

explained considering that in the studied conditions,
because of low O2 concentrations, O2 consumption rate
was lower than O2 permeation rate through the lm.
Besides, CO2 concentration signicantly increased
(P < 0.05), suggesting that CO2 production rate was
higher than lm permeability.
Oxygen concentrations did not fall bellow 5% and
CO2 maximum concentrations were lower than 6%,
during the entire tested storage time, for all the
evaluated storage conditions. Under this atmospheric
composition, fermentative metabolism in the plant
tissue is not supposed to occur (Farber et al., 2003).
According to Varoquaux et al. (1996), for butterhead
lettuce physiological disorders, such as brown and
necrotic stains, might occur if CO2 concentration inside
the packages is higher than 10%, and therefore in the
studied storage conditions physiological damage due to
CO2 was not expected.
As shown in Fig. 2, increasing storage temperature
caused a decrease in O2 concentration and a signicant
increase in CO2 concentration inside the packages,
mainly because of an increased respiration rate.
On the contrary, passive modied atmosphere packages showed higher O2 concentrations than those with
active modied atmosphere, which could be attributed
to higher initial O2 concentration. CO2 concentration
was higher in passive modied atmosphere packages
than in those with active modied atmosphere. These
results could be explained considering that lettuce leaves
in passive atmosphere packages showed a higher respi-
Figure 2 Weight loss for lettuce in active and passive modied

atmosphere packages during storage at 5 and 10 C. Vertical bars
represent Tukeys honestly signicant differences.
ration rate than those stored in active modied atmosphere packages, because of higher O2 concentration
inside the packages.
Weight loss
Analysis of variance showed that type of modied

atmosphere and storage time had a highly signicant
eect (P < 0.001) on weight loss. Weight loss signicantly increased (P < 0.05) with storage time, as shown
in Fig. 2. Weight loss increased with increasing storage
temperature, which could be attributed to higher respiration and higher dehydration. For all the studied
storage temperatures, by the end of the tested time,
weight loss reached values higher than 20%, suggesting
an important deterioration in lettuce.
No signicant dierences were found in the weight
loss of lettuce leaves packaged in active and passive
modied atmosphere at both the storage temperatures.
This lack of dierence suggests that dierences in
package atmosphere composition, and therefore in
respiration rate, did not aect weight loss of lettuces
leaves. Thus, in the studied storage conditions, weight
loss might have been more aected by supercial
dehydration than by respiration rate.
Sensory analysis
For all the storage conditions, the evaluated sensory

attributes were highly signicantly (P < 0.001) aected
by storage time, supporting the validity of the chosen
descriptors as indicators of lettuce deterioration. As
lettuce leaves deteriorated, wilting appearance increased,
midribs developed browning, brown and necrotic stains
areas and developed o-odour. The development of the
evaluated sensory attributes is shown in Fig. 3.
As shown in Fig. 3, o-odour only appeared after
21 days of storage at 10 C and after 42 and 49 days of
storage at 5 C in passive and active modied atmosphere packages, respectively, being appearance attributes the defects that rst occurred. Therefore, the
development of o-odour might not limit the sensory
shelf life of lettuce leaves in modied atmosphere
packages, being the sensory shelf life of this product
determined by changes in its appearance. O-odour
development might be due to fermentative metabolism
of the plant tissue, caused by low O2 concentrations
(Ballantyne et al., 1988; Beaudry, 1999, 2000), or by
spoilage micro-organisms growth (Carlin et al., 1989;
Jacxsens et al., 2003). In this case, as O2 concentrations
were higher than 5% during the entire storage period,
o-odour development might be attributed to the
growth of spoilage micro-organisms at the end of the
tested storage time.
Lettuce leaves stored at 10 C showed signicantly
higher wilting appearance, presence of dark and necrotic
Figure 3 Development of the evaluated sensory attributes as a function of storage time for lettuce leaves in active and passive modied atmosphere
packages during storage at 5 and 10 C: (a) off-odour, (b) wilting appearance, (c) brown and necrotic stains on the leaf surface, (d) browning
on the midribs. Vertical bars represent Tukeys honestly signicant differences.
stains on the surface and browning on the midribs, than

those stored at 5 C. These results indicate an increase in
sensory deterioration rate with increasing storage temperature that might be attributed to higher respiration
rate, dehydration and metabolical activity (Tavarini
et al., 2007).
Throughout the evaluated storage period, no significant dierences (P > 0.05) were found in the wilting
appearance, presence of stains on the leaf surface or
browning on the midribs for lettuces packaged in active
and passive modied atmosphere and stored at 10 C.
The dierence in atmosphere composition of packages
with active and passive modied atmosphere suggests a
dierent respiration rate for lettuce leaves in active and
passive modied atmosphere. However, this dierence
did not aect deterioration rate. Therefore, at 10 C,
active modied atmosphere was as ecient as passive

modied atmosphere in preserving the quality of lettuce
leaves.
On the contrary, when packages were stored at 5 C,
lettuce leaves in passive modied atmosphere showed
signicantly higher o-odour, wilting appearance, presence of stains on the leaf surface or browning on the
midribs than those in active modied atmosphere,
indicating a higher deterioration rate. These results
suggest that active modied atmosphere slowed down
sensory deterioration rate when compared with passive
modied atmosphere, probably because of lower respiration rate.
Thus, the inuence of active modied atmosphere
with respect to passive modied atmosphere depended
on the storage temperature considered. Active modied
1675
1676
atmosphere yielded better results when packages were

stored at 5 C, being eective in retarding the deterioration of lettuce leaves. Therefore, for the evaluated
lm, the use of active modied atmosphere in replacement of passive modied atmosphere would only be
recommended if a storage temperature of 5 C is used.
A relatively good t was obtained for all the evaluated
sensory attributes considering rst-order reaction rate,
in agreement with the published data for cut lettuce
(Vankerschaver et al., 1996; Piagentini et al., 2005).
Models parameters, as presented in eqn 1, are shown in
Table 1.
Consumer study
All the evaluated attributes were signicantly correlated

to each other (R2 higher than 0.85, P < 0.05, data not
shown), indicating that they were all aected in a similar
way by lettuce deterioration. As the presence of stains
on the leaf surface and browning on the midribs were
the attributes that signicantly increased rst, they were
selected for further analysis.
A good t was obtained for logistic regression of
consumers rejection to purchase vs. presence of stains
on the surface and browning on the midribs (R2 = 0.73
and 0.83, respectively), suggesting that these attributes
were considered by consumers when deciding to purchase a package of lettuce leaves. The equations
constants are shown in Table 2.
These regressions could be used to estimate the
attribute intensity that corresponds to a certain consumer rejection percentage. Twenty-ve percent is one
of the most used consumers rejection percentages to
estimate sensory shelf life (Gambaro et al., 2006;
Gimenez et al., 2007). Therefore, a leaf of lettuce
showing an intensity higher than 2.0 for presence of
stains or an intensity higher than 1.3 for browning on
the midribs would be rejected by more than 25% of the
consumers.
Table 2 Models parameters (cf. eqn 2) for the logistic regression of

consumer rejection percentage vs. presence of dark and necrotic stains
on the leaf surface and browning on the midribs
Sensory attribute
R2
Presence of dark and necrotic stains

Browning on the midribs
9.1
5.6
71.1
74.6
49.9
13.6
2.0
1.4
0.73
0.83
In most studies, the sensory shelf life of minimally

processed vegetables has been estimated considering the
time necessary to reach an arbitrary limit of a certain
sensory attribute, as evaluated by a trained sensory panel.
This limit has been considered as a score of 50% of the
scale used to evaluate that attribute (Li et al., 2001; Zhou
et al., 2004; Piagentini et al., 2005). In this case, scores of
less than 25% of the scale corresponded to consumer
rejection percentages higher than 25%. Therefore, more
conservative criteria than the usually employed would
be necessary for shelf life estimation, in order to assure the
products quality at the end of its shelf life. Moreover,
these results showed the importance of performing
consumer studies in order to establish proper criteria
to estimate the shelf life of fresh fruits and vegetables.
Shelf life estimation
Sensory shelf life could be determined as the time

required for a certain sensory attribute to reach a
predetermined value (Gambaro et al., 2006). As the
development of the evaluated sensory attributes with
storage time was modelled (Table 1), shelf life of lettuce
leaves in each of the evaluated storage conditions was
estimated as the time necessary to reach attribute
intensities that corresponded to a 25% consumer
rejection to purchase. Estimated shelf lives are shown
in Table 3. As expected, sensory shelf life decreased with
increasing storage temperature. Regarding active and
passive modied atmosphere, results depended on the
Storage condition
Sensory attribute
Off-odour
Wilting appearance
Presence of dark stains
Browning on the midribs
Parameter
X0
kT
R2
X0
kT
R2
X0
kT
R2
X0
kT
R2
Passive 5 C
2.0
0.16
0.99
1.8
0.26
0.96
5.2
0.10
0.99
1.1
0.13
0.98
10
)3
10)5
10)2
10)2
Active 5 C
2.4
0.19
0.90
3.2
0.19
0.90
6.0
0.28
0.95
5.0
0.29
0.94
10
)4
10)4
10)6
10)6
Passive 10 C
)5
1.0 10
0.70
0.95
2.7 10)1
0.073
0.74
2.7 10)1
0.11
0.87
3.4 10)1
0.09
0.88
Active 10 C
3.7
0.56
0.99
1.0
0.30
0.98
1.0
0.20
0.84
2.3
0.12
0.92
10)5
Table 1 Models parameters (cf. eqn 1) for the

development of sensory attributes as a function of storage time, for lettuces leaves in
passive and active modied atmosphere
packages, stored at 5 and 10 C
10)2
10)1
10)1
Table 3 Shelf lives (days) for lettuce in active and passive modied
atmosphere packages stored at 5 and 10 C, estimated considering
attribute intensity corresponding to 25% consumers rejection to
purchase
Shelf life (days) considering the following sensory
attributes
Storage
condition
5 C
Passive
Active
10 C
Passive
Active
Dark and necrotic stains

on the leaf surface
Browning on the
midribs
37
46
36
43
17
15
14
14
storage temperature being considered. If lettuce leaves

were stored at 10 C, no dierences were found between
the shelf lives of lettuce packaged in passive and active
modied atmosphere. On the contrary, if lettuce leaves
were stored at 5 C, the use of active modied atmosphere increased the shelf life of lettuce leaves in
approximately 20% with respect to passive modied
atmosphere, in agreement with results for sensory
deterioration rate. The use of the evaluated active
modied atmosphere would only be recommended at
low storage temperatures.
Conclusions
Results showed that the ecacy of active modied

atmosphere with respect to passive modied atmosphere
depended on the storage temperature considered. When
stored at 10 C, despite being subjected to dierent
package atmosphere composition, lettuce leaves in
active and passive modied atmosphere showed the
same sensory deterioration rate and sensory shelf life.
On the contrary, when stored at 5 C, lettuce leaves in
active modied atmosphere packages showed lower
deterioration rate and higher sensory shelf life than
those in passive modied atmosphere. These results
suggest the importance of studying the inuence of
gaseous atmosphere at dierent storage temperatures,
instead of studying both the factors separately.
Scores of less than 25% of the scale for presence of
dark and necrotic stains on the leaf surface and
browning of the midribs corresponded to 25% consumer rejection to purchase. Therefore, criteria commonly used to determine the sensory shelf life of
minimally processed vegetables might not be strict
enough as to assure the products quality at the end of
its shelf life. Results showed the importance of
performing consumer studies in order to establish
proper criteria to estimate the shelf life of fresh fruits
and vegetables.
Acknowledgments
The authors are indebted to PDT (Programa de

Desarrollo Tecnologico, Uruguay) for nancial support,
and grateful to Ines Mart nez for her extremely valuable
help.
References
Ballantyne, R., Stark, J. & Selman, D. (1988). Modied atmosphere
packaging of shredded lettuce. International Journal of Food Science
Beaudry, R. (1999). Eect of O2 and CO2 partial pressure on selected
phenomena aecting fruit and vegetable quality. Postharvest Biology
Beaudry, R. (2000). Responses of horticultural commodities to low
oxygen: limits to the expanded use of modied atmosphere
packaging. Horticulture Technology, 10, 491500.
Carlin, F., Nguyen-the, C., Cudennec, P. & Reich, M. (1989).
Microbiological spoilage of fresh ready-to-use carrots. Sciences des
Aliments, 9, 371386.
Farber, J.N., Harris, L.J., Parish, M.E. et al. (2003). Microbiological
safety of controlled and modied atmosphere packaging of fresh and
fresh-cut produce. Comprehensive Reviews in Food Science and Food
Safety, 2, 142160.
Fonseca, S.C., Oliveira, A.R. & Brecht, J.K. (2002). Modelling
respiration rate of fresh fruits and vegetables for modied atmosphere packages: a review. Journal of Food Engineering, 52, 99119.
Gambaro, A., Ares, G. & Gimenez, A. (2006). Shelf life estimation of
apple baby food. Journal of Sensory Studies, 21, 101111.
Gimenez, A., Varela, P., Salvador, A., Ares, G., Fiszman, S. &
Garitta, L. (2007). Shelf life estimation of brown bread: A consumer
approach. Food Quality and Preference, 18, 196204.
Jacxsens, L., Devlieghere, F., Ragaert, P., Vanneste, E. & Debevere, J.
(2003). Relation between microbiological quality, metabolite production and sensory quality of equilibrium modied atmosphere
packaged fresh-cut produce. International Journal of Food Microbiology, 83, 263280.
Li, Y., Brackett, R.E., Shewfelt, R.L. & Beuchat, L.R. (2001). Changes
in appearance and natural microora on iceberg lettuce treated in
warm, chlorinated water and then stored at refrigeration temperature. Food Microbiology, 18, 299308.
Odumeru, J.A., Boulter, J., Knight, K., Lu, X. & McKellar, R. (2002).
Assessing of a thermalchemical process to extend the shelf life of
ready-to-use lettuce. Journal of Food Quality, 26, 197209.
Piagentini, A.M., Mendez, J.C., Guemes, D.R. & Pirovani, M.E.
(2005). Modelling changes of sensory attributes for individual and
mixed fresh-cut leafy vegetables. Postharvest Biology and Technology, 38, 202212.
Ragaert, P., Verbeke, W., Devlieghere, F. & Debevere, J. (2004).
Consumer perception and choice of minimally processed vegetables
and packaged fruits. Food Quality and Preference, 15, 259270.
Tavarini, S., DeglInnocenti, E., Pardossi, A. & Guidi, L. (2007).
Biochemical aspects in two minimally processed lettuces upon
storage. International Journal of Food Science and Technology, 42,
214219.
Vankerschaver, K., Willocx, F., Smout, C., Hendrichkx, M. &
Tobback, P. (1996). Modeling and prediction of visual shelf life of
minimally processed endive. Journal of Food Science, 61, 1094
1098.
Varoquaux, P., Mazollier, J. & Albagnac, G. (1996). The inuence of
raw material characteristics on the storage life of fresh-cut butterhead lettuce. Postharvest Biology and Technology, 9, 127139.
Zhou, T., Harrison, A.D., McKellar, R. et al. (2004). Determination
of acceptability and shelf life of ready-to-use lettuce by digital image
analysis. Food Research International, 37, 875881.
1677
1678
Original article
Effect of soluble CO2 stabilisation and vacuum packaging in the
shelf life of farmed sea bream and sea bass fillets
Rogerio Mendes* & Amparo Goncalves
Department of Technological Innovation and Upgrading of Fish Products, National Research Institute on Agriculture and Fisheries
INIAP IPIMAR, Avenida Bras lia, 1449-006, Lisboa, Portugal
(Received 18 December 2007; Accepted in revised form 22 January 2008)
Summary
The objective of this study was to determine the dierences of sensory, microbiological and chemical quality
in vacuum-packaged llets of sea bream and sea bass previously submitted to soluble gas solubilisation
(SGS) with 100% CO2, at 2 bar for 30 and 60 min and stored at chilled temperature for 15 days. Apart from
pH value that showed a regular increase during chilled storage, the other chemical index [total volatile bases
nitrogen (TVB-N), trimethylamine nitrogen (TMA-N) and thiobarbituric acid reactive substances (TBARs)]
had showed to be poor indicators of changes in quality of products. Final TVB-N values ranged from 16.0 to
17.4 mg N per 100 g and from 17.3 to 19.4 mg N per 100 g in sea bream and sea bass, respectively. Sensory
evaluation resulted as the most reliable parameter of quality decay. The results show that SGS treatment
kept the initial quality of llets for longer time, which was particularly visible on the sea bream llets, thus
contributing to an extension in 23 days of the shelf life. SGS had also a positive eect in the delay of
microbial growth.
Keywords
Chill storage, quality changes, sea bass, sea bream, shelf life, soluble gas stabilisation, vacuum packaging.
Introduction
Increasing consumer demand for fresh sh is one of the

major contemporary trends in sh consumption. The
growing need for animal protein for human consumption and mainly these new consumer trends have
contributed signicantly for the raise of the importance
of aquacultured species. On account of this, sea bass and
sea bream sh farming in the Mediterranean has
increased from 54 000 ton (1996) to 200 000 ton
(2006) in a time span of a decade (INFOFISH, 2007).
Sea bream and sea bass are economically important sh
species in many southern European countries mainly on
account of being easily farmed and also because of their
desirable aroma and nutritional quality. These species
are generally marketed wholesale, gutted or ungutted for
immediate consumption at retail stores.
Demand for fresh convenient products that oer
healthy, tasty and fast meal solutions stresses, however,
the need for new ways of marketing these aquaculture
species. A good alternative could be a lleted product in a
convenient package, more easy to handle, e.g. in supermarkets and also ready to cook. According to the EEC
*Correspondent: Fax: +351 213015948;
e-mail: rogerio@ipimar.pt
Council Directive of 22 July 1991, modied atmospherepacked and vacuum-packed sh products are considered
as fresh products. Consequently, CO2 can be used for
preservation of fresh sh products, when a shelf life
extension of only a few days is found to be sucient.
Modied atmospheres packaging (MAP) together
with refrigeration has become increasingly popular as
a method of food preservation ensuring quality maintenance with minimum losses (Ashie et al., 1996). To
increase shelf life of products, active packaging has
emerged within MAP as a new technological development in the eld of food packaging and is used whenever
the package requires a function other than providing an
inert barrier to external conditions. Generally, this
technique consists of the addition inside the package
of mechanisms to delay the product deterioration, such
as gas scavengers or emitters.
Soluble gas stabilisation (SGS) is a relatively recent
methodology of active packaging that has been proposed by Sivertsvik (2000) to extend the shelf life of
packaged sh.
The action of CO2 on the bacterial growth, namely, its
preservative eect caused by the absorption of CO2
from the package atmosphere to the product, is well
known from the principles in which the MAP techniques
are based upon. The application of SGS using elevated
doi:10.1111/j.1365-2621.2008.01737.x
Sensory, microbiological and chemical quality of sea bream and sea bass R. Mendes and A. Goncalves
pressures is done before packaging and the CO2 is

dissolved in the sh at low temperature (0 C) and
pressures 2 bar, according to the following equation:
CO2 g H2 Ol $ HCO
3 H $ CO3 2H
The dissolvement of CO2 into sh esh follows

Henrys Law, and at normal pH, most of the CO2 will
exist as dissolved CO2 (aq) and only about 1%
dissociate according to the equation shown. The solubility of CO2 hence increases with increased partial
pressures of CO2 and therefore also with increased
overall pressure. However, the diusion process follows
Ficks 2nd Law and elevated pressure will not increase
diusional speed, but the overall CO2 absorption rate
increases because of the increased solubility at elevated
pressures.
The use of this technique would favour a CO2
supersaturated product that could then be packaged
with a smaller g p (gas/product) ratio or even vacuum
packaged. The same eect as normal MAP would
therefore be obtained but with reduced packaging
volume. Furthermore, using vacuum packaging, the
deleterious eects that occur sometimes after a prolonged exposure of products to CO2 (Sivertsvik et al.,
2004; Masniyom et al., 2005; Goulas & Kontominas,
2007) could be avoided. In addition, packaging conditions that reduce the amount of oxygen present in the
package (e.g. vacuum packaging) are known to extend
the shelf life of product by inhibiting the growth of
aerobic spoilage bacteria (FDA, 2001).
In previous work, the eect of soluble CO2 stabilisation treatment in the shelf life of farmed sea bream
and sea bass llets packaged in air was studied (Mendes
& Goncalves, 2008). Despite SGS treatment presented a
positive eect on the microbiological quality of the
llets, no signicant extension of sensory shelf life was
visible as a function of SGS treatment. Therefore, the
objective of this work was to evaluate the eect of
soluble CO2 stabilisation and vacuum packaging on the
quality and shelf life of llets of farmed sea bream
(Sparus aurata) and European sea bass (Dicentrarchus
labrax) by monitoring microbiological, sensorial and
chemical changes throughout refrigerated storage at
2 C.
and sea bass, respectively. Upon arrival, the sh were

lleted and the llets (with skin) were washed by hand
and drained. The average weights of llets were
107.8 15.8 and 193.2 22.8 g, respectively, for sea
bream and sea bass.
Soluble gas solubilisation treatment and vacuum packaging
Sets of four llets from four dierent shes were placed

skin down in a polypropylene tray (22 cm 17 cm
2.5 cm) together with a moisture absorber (size
20 cm 15 cm 0.2 cm). For the SGS treatment, the
trays were placed inside a constant volume and gas
impermeable steel chamber ( 50 cm; length 74 cm) at a
temperature of 1.0 0.5 C. The cylindrical chamber
was ushed with 100% CO2 for removal of air and after
closure of purging valve, it was kept lled with CO2 by
allowing a continuous supply of CO2 at 2 bar for 30 and
60 min. The ratio of gas volume to product volume
was around 5. The llets volume was estimated by
dividing the llet mass with muscle density, 1.05 g cm)3
(Lowndes, 1955). After SGS treatment, trays were
packaged in a polyamide polyethylene gas barrier bag
(Vaessen-Schoemaker, Portugal) with transmission rates
of 25.0 for O2, 61.0 for CO2 and 8.8 for N2, cc m)2 per
24 h, at 75% RH and 23C. Packages were immediately
submitted to partial vacuum (40 mbar) and sealed on a
Multivac A 300 52 machine (Multivac, Germany). All
packages were stored at 1 0.5 C, together with the
control lot, composed by llets without the SGS treatment and therefore vacuum packaged directly.
Storage and sampling
During the chilled storage, samples were taken for

sensory evaluation, physical chemical and microbial
analysis after 0, 2, 6, 9, 13 and 15 days. Two packages
from each batch (control, SGS 30 and SGS 60) were
collected, and from each package, two llets were taken
for sensory assessment and the other two llets were
used to prepare a composite sample for microbial
analysis, pH value and chemical determinations.
pH value determination
The pH was measured directly on sh mince, using a

surface electrode (Metrohm 744, Metrhom, Switzerland).
Raw material
Immature farmed sea bream (S. aurata) and sea bass (D.
labrax) raised in tanks (Tavira, Portugal) were fasted for
48 h prior to slaughter by immersion in an ice and seawater slurry (3:1). The sh was kept in ice and
transported to laboratory within 24 h. Average weight
was 345.2 36.3 and 453.1 52.1 g for sea bream
Moisture, crude protein, ash and fat were determined on

day 0 by the procedures described in AOAC (1998).
Moisture was determined by the air drying method, at
105 2 C in air oven (Method 950.46). Crude protein
was determined by Kjeldahl method and the percentage
of protein was calculated as total nitrogen 6.25
1679
1680
(Method 981.10). Ash content was determined by

incineration of sh muscle (previously dried) in a
furnace at 500 25 C until to constant weight
(Method 942.05). Total fat was extracted with diethyl
ether solvent in a Soxleth extractor (Method 991.36).
Sample was reuxed for 7 h at adjusted temperature.
Fat was determined by weight after drying to constant
weight in a 105 2 C air oven.
texture (rmness and succulence) of sh esh was also

evaluated using the category scale (Meilgaard et al.,
1999) of ve points (1 soft dry; 5 rm succulent).
Average texture score of 2 was considered as the limit of
acceptability of llets.
Total volatile bases nitrogen (TVB-N), trimethylamine
nitrogen (TMA-N)
Volatile compounds were determined in extracts from

25 g of sh mince homogenised with 50 ml of 10%
trichloracetic acid (TCA) by a microdiusion method
(Cobb et al., 1973).
Thiobarbituric acid reactive substances (TBARs)
According to Vyncke (1970), lipid oxidation was

followed by determining the TBARs. These substances
include the malonaldehyde (MDA) which is an oxidative
secondary compound.
For microbial counts (total viable counts TVC), 10 g

of sh muscle (a pool of two llets from each package)
was homogenised, diluted ninefold in tryptone salt
solution (0.85%) and poured onto Plate Count Agar
(Merck, Darmstadt, Germany). The plates were incubated at 30C for 72 h, under anaerobic conditions.
Sensory evaluation
Sensory evaluation was carried out in a special room

equipped wit individual booths by a sensory panel
composed by four experienced assessors. The llets were
presented to the panel in white-coded dishes. Each
panelist assessed one llet from each lot. The odour and
colour of raw llets were assessed using a category scale
(Meilgaard et al., 1999) of ve points (1 putrid
odour evident discoloration; 5 typical fresh
odour characteristic colour). A score of 3 points was
considered as limit of acceptability of raw llets.
After the raw assessment, the llets were wrapped
with perforated aluminium foil and cooked for 10 min
at 100 C in a saturated steam oven (Rational CombiMaster CM6 Cross Kuchentechnik CmbH, Landsberg
a. Lech, Germany). The odour and avour were
assessed using the Torry scale (Howgate, 1985), slightly
modied, and their scores range from 10 to 3: 10 very
fresh sh, with weak odour sweet characteristic avour;
5 slightly sour rancid odour, trace of o-avours; 3
putrid or spoiled sh. Average score 5.5 for odour and
avour is usually used as limit of acceptability. The
The proximate composition of llets of both species at

day 0 is presented in Table 1. The results are in
accordance with those reported by other authors
(Tejada & Huidobro, 2002; Bandarra et al., 2004;
Goulas & Kontominas, 2007; Grigorakis, 2007). The
fat content in sea bass is higher than the 1.7% value
reported by Ozogul et al. (2005) in wild specimens, thus
reecting, namely, the specic nutrition and the reduced
activity of the farmed sh.
Estimation of dissolved CO2 in the fish fillets
Under the conditions of our experiments (2 bar 100%

CO2 = 200 kPa CO2; applied for 30 and 60 min), the
concentration of CO2 as a function of time and llet
thickness can be estimated by the equation proposed by
Crank (1975):

1
X
4Ct1
1n
2n 1px
CO2
t1
cos
CCO2 t; x CCO2
2x0
p n0 2n 1
"
#
2n 12 p2 D
exp
t
4x20
t1
where CCO
is the amount dissolved CO2 at equilibrium
)1 2
(mg kg ), x is the diusional position, x0 is the llet
thickness 2 = 0.5 cm, t is the time (seconds) and D is
the diusion coecient of raw sh llets = 1.68
10)5 cm2 s)1 (calculated by Sivertsvik et al., 2004).
The
application
of
Henrys
Law:
t1
t1
HCO2 ;prod CCO
allows the estimation of the
PCO
2
2
amount of dissolved CO2 at equilibrium around
t1
is the amount dissolved CO2
4.43 g kg)1, where CCO
2
Table 1 Proximate composition of the esh of cultured sea bream and
sea bass
Proximate composition (g per100 g flesh)
Sea bream
Sea bass
Protein
Fat
Moisture
Ash
19.0
5.3
74.2
1.21
23.9
10.7
64.0
1.20
(0.6)
(0.4)
(0.2)
(0.04)
(0.6)
(0.4)
(0.2)
(0.04)
Results outcome from the analysis of composite samples. Values in

parenthesis are the analytical precision (analyses were done in
duplicate).
at equilibrium (mg kg)1), HCO2 ;fish is the Henrys constant for raw sh llets = 45.1 Pa ppm)1 (calculated by
t1
is the partial pressure
Sivertsvik et al., 2004) and PCO
2
of CO2 = 200 kPa.
Therefore, assuming a uniform geometry of the llet
and a thickness of 1 cm, the amount of CO2 in the centre
of llets (x = x0) after 30 and 60 min may be estimated
as 0.37 and 1.34 g kg)1, respectively.
uum packaging reinforced the positive eect of SGS

treatment on the inhibition of microbial growth. Lower
microbial counts in SGS samples can be attributed to
the inhibitory eect of CO2 after dissolution in both the
aqueous and fatty phase of sh (Masniyom et al., 2002;
Sivertsvik et al., 2002; Ozogul et al., 2004). Genigeorgis
(1985) suggested that the anti-microbial activity of CO2
was a result of the gas being absorbed onto the surface
of the food forming carbonic acid, subsequent ionisation
of the carbonic acid and a reduction in pH. However,
the same author reports this decrease in pH as minimal
and probably would not be sucient to cause any
signicant bacteriostatic eect. The bacteriostatic eect
of CO2 has been referred as resulting from several
factors, namely, the alteration of cell membranes
including eects on nutrient uptake and absorption,
inhibition of enzymes or decreases in rate of enzyme
reactions, changes in the physicochemical properties of
proteins and intracellular pH changes (Farber, 1991;
Sivertsvik et al., 2002). Sivertsvik (2000) also showed
that SGS treatment and vacuum packaging had a
positive eect on TVC of chilled salmon llets. Values
attained in these conditions were lower than 6 log cfu g)1 after 16 days of storage.
Taking into account that vacuum packaging creates

extremely reduced oxygen conditions, TVC in anaerobic
conditions were determined in sea bream and sea bass
llets (Fig. 1). According to Huss (1995), total microbiological counts reect mainly, the initial contamination characteristic of the capture area, the type of shing
as well as handling conditions not being therefore
possible to relate this with the freshness condition.
TVC showed relatively low values in all samples at the
beginning of the storage (<2.0 log cfu g)1), thus indicating that the initial quality of the sh was good and
that all the sh handling (lleting, SGS treatment and
vacuum packaging) was done correctly. Changes during
chilled storage were characterised by a considerably long
lag phase, 13 days in sea bream llets and 9 days in sea
bass llets. In sea bream, SGS treatment had a considerable bacteriostatic eect on bacteria growth, whereas
in sea bass, the eect of CO2 treatment was only visible
after 13 days. At the end of storage, bacterial counts in
sea bream and sea bass llets followed the order 60 min
SGS <30 min SGS < control. Final counts were
between 1.7 and 2.8 log cfu g)1 in sea bream and
between 3.4 and 4.5 log cfu g)1 in sea bass. Mendes &
Goncalves (in press) in a previous study with SGS
treatment and air packaging of the same sh species
showed the existence of considerably higher TVC
(variation range: 2.89.3 log cfu g)1 for sea bream and
2.88.6 log cfu g)1 in sea bass) and a similar order of
increase during chilled storage, i.e. 60 min
SGS < 30 min SGS < control, thus showing that vac-
pH value
In Fig. 2, the changes in pH during chilled storage of sea

bream and sea bass llets are presented. Values of pH
increased progressively during the storage period,
although, in a short range of variation. In sea bream
llets, pH changed from 5.96 to 6.20, whereas in sea
bass, pH ranged from 6.09 to 6.43. Debevere & Boskou
(1996) report that CO2 solubilisation in the muscle has a
stabilisation eect on pH, thus counteracting the
increasing eect that the accumulation of ammonium
and other volatile nitrogenous components originated
from the bacterial metabolism have on muscle pH
(Andersen et al., 1997). No signicant changes were also
accounted for the SGS treatment. Furthermore and
contrary to what would be expected during the rst days
Sea bream
SGS treated and vacuum-packaged llets of

farmed sea bream and sea bass, during chilled
storage at 1.0 0.5 C. Values outcome
from composite sample analysis. Standard
deviation of analytical duplicates was 0.66
(log cfu g)1).
Sea bass
5.0
4.0
4.0
Log cfu g1
)1
Figure 1 Changes in TVCs (log cfu g ) of
Log cfu g1
5.0
3.0
2.0
3.0
2.0
1.0
1.0
0.0
0.0
0
Control
8 10 12 14 16
Days
30 min
60 min
Control
8 10 12 14 16
Days
30 min
60 min
1681
1682
pH
Sea bream
6.45
6.45
6.35
6.35
6.25
6.25
6.15
6.15
6.05
6.05
pH
Sea bass
5.95
5.95
0
10 12 14 16
Days
30 min
Control
60 min
30
25
25
20
20
15
15
10
10
TVB-N
(mg 100 g1)
10 12 14 16
Days
30 min
Control
TVB-N
Sea bream
(mg 100 g1)
30
60 min
Sea bass
0
0
Control
8 10 12 14 16
Days
30 min
60 min
Control
Figure 2 Changes in pH values of SGS

treated and vacuum-packaged llets of
deviation of analytical duplicates was 0.03.
8 10 12 14 16
Days
30 min
60 min
of chilled storage, increased CO2 application induced

higher absolute pH values in llets of both species.
According to Sivertsvik et al. (2002), in sh that is
packaged in a rich CO2 atmosphere, a certain amount of
CO2 dissolves into the product leading to the production
of carbonic acid, subsequent ionisation of the carbonic
acid and reduction in pH. This decrease in pH was,
however, reported by Genigeorgis (1985) as being
minimal which seems to be in accordance with the
non-existence of important dierences in the majority of
the samples between control and SGS treated sh. In a
previous study with the same sh species packaged in
air, Mendes & Goncalves (in press) reported the same
behaviour and added that taking into account that the
pH was measured not at the sh surface but in a
homogenate of the muscle, one possible explanation
would be that although pH change can occur at the sh
surface, carbonic acid production does not diuse to the
interior of the muscle or that diusion is insucient to
change signicantly the overall muscle pH. Nevertheless,
the lower microbiological contamination and lower
concentrations of volatile basic nitrogen compounds
derived from bacterial metabolism may be accounted in
the present study for the lower pH values determined.
Figure 3 Changes in TVB-N values of SGS

deviation of analytical duplicates was
0.7 mg N per 100 g.
Total volatile bases nitrogen
TVB-N values for vacuum-packaged sea bream and sea

bass llets are presented in Fig. 3. TVB-N is a product
of bacterial spoilage and endogenous enzymes action,
and its content is often used as an index to assess the
keeping quality and shelf life of products (Connell, 1995;
EEC, 1995). O-avouring compounds such as ammonia, monoethylamine, dimethylamine and TMA are
responsible for the levels of this index (Connell, 1995;
Debevere & Boskou, 1996; Ruiz-Capillas & Moral,
2005). The initial TVB-N value of 19.0 and 21.9 mg N
per 100 g, respectively, in sea bream and sea bass llets,
is of the same order of magnitude as the value of 14.2
and 17.6 mg N per 100 g reported in our previous work
(Mendes & Goncalves, in press). The initial values are
also similar to other values reported for fresh sea bream
and fresh sea bass. Namely for sea bream, 16 mg N per
100 g at 0 days (Goulas & Kontominas, 2007); 20 mg N
per 100 g after 1 day of storage at 2 1 C (Tejada &
Huidobro, 2002); 19 mg N per 100 g at 0 days (Goncalves et al., 2004) and for sea bass, 1922 mg N per
100 g at 0 days (Castro et al., 2006); 16 mg N per 100 g
at 0 days (Grigorakis et al., 2004); 22.7 mg N per 100 g
and 27.2 mg N per 100 g, respectively, in whole and

lleted sea bass stored 1 day in ice (Taliadourou et al.,
2003). Small dierences in initial TVB-N levels may be
related to non-protein nitrogen content which is dependent by his side of factors such as type of sh feeding,
season of catching, sh size and also environmental
factors (Connell, 1995). Ultimately, TVB-N content is
directly related to the initial microbial activity in the sh
tissue (Connell, 1995; Kyrana et al., 1997; Ozogul et al.,
2004).
No signicant changes were determined in both
species between the control batches and the SGS-treated
samples during the total period of chilled storage. No
major changes in TVB-N values were also visible during
chilled storage. In both species, levels below the limit
established by European regulation (EEC, 1995)
(35 mg N per 100 g of sh esh) were observed till the
end of storage. Final values ranged from 16.0 to
17.4 mg N per 100 g and from 17.3 and 19.4 mg N
per 100 g in sea bream and sea bass, respectively. These
values are lower than TVB-N values reported in the
previous work, in which SGS-treated llets were packaged in air (Mendes & Goncalves, in press) and are
possibly the reection of the lower microbial contamination in vacuum-packaged samples.
Banks et al. (1980) report that the dierences in TVBN amounts between sh packaged in air and in MAP
must have been caused by a smaller number of bacteria
and or their lower ability to act on the oxidative
deamination of non-protein nitrogen compounds. Another explanation advanced (Soccol et al., 2005) refers
that anaerobic conditions existent in CO2 MAP and also
in vacuum packaging may inhibit the oxidative deamination because of the lower O2 quantity. The type of
bacteria existing in the sh may depend on atmospheric
and pH changes caused by the absorption of CO2. These
bacteria show low capacity to produce ammonia when
compared with ordinary aerobic spoilage bacteria.
Considering the dierent deterioration mechanisms of
CO2 MAP and vacuum-packaged sh in comparison
with the one that occurs in ice-stored sh, TVB-N
TMA
(mg 100 g1)
1.5
Figure 4 Changes in TMA-N values of SGS

deviation of analytical duplicates was
0.7 mg N per 100 g.
production has been therefore considered as a poor

spoilage indicator (Banks et al., 1980; Kyrana et al.,
1997; Gimenez et al., 2002; Tejada & Huidobro, 2002).
Trimethylamine nitrogen
The TMA-N values of vacuum-packaged llets of sea

bream and sea bass are presented in Fig. 4. TMA is part
of the non-protein nitrogen fraction of the marine sh
esh and is formed from trimethylamine oxide (TMAO)
by bacterial enzymatic activity during spoilage. TMA is
mainly responsible for the disagreeable shy odour. It
has been used as a biochemical index to assess the
quality and shelf life of marine sh (Connell, 1995;
Debevere & Boskou, 1996; Gram & Huss, 1996).
Initial values of TMA-N in sea bream and sea bass
were 0.93 and 0.20 mg N per 100 g of esh, respectively.
Sea bream llets always showed values lower than the
initial ones, and concentrations were aected by SGS
treatment. The longer the time of CO2 application
during SGS treatment, the lower the amount of TMA-N
detected. Levels in sea bass remained unchanged during
the entire period of chilled storage. These values are the
result of low initial TMAO-N values found for this
cultured species, ranged between 3 and 13 mg N per
100 g sh esh (unpublished data). TMA-N levels in sea
bream were of the some order of magnitude of values
reported by Goulas & Kontominas (2007), 0.31 mg N
per 100 g, 1.51 mg N per 100 g and by Tejada &
Huidobro (2002), 0.40 mg N per 100 g. Considerably
higher values, 1.513.70 mg N per 100 g in sea bream
and 0.489.72 mg N per 100 g in sea bass, were reported
by Mendes & Goncalves (in press) in SGS-treated
samples packaged in air during a 15-day period of
chilled storage.
Farber (1965) reported in sole a value of 5 mg N per
100 g when the sh was spoiled. Lannelongue et al.
(1982) suggested 5 mg N per 100 g as the limit of
acceptability for swordsh also. Goulas & Kontominas
(2007) used 23 mg N per 100 g as limit of acceptability,
whereas Kyrana et al. (1997) and Tejada & Huidobro
Sea bream
TMA
(mg 100 g1)
1.5
1.0
1.0
0.5
0.5
0.0
Sea bass
0.0
Control
8 10 12 14 16
Days
30 min
60 min
Control
8 10 12 14 16
Days
30 min
60 min
1683
1684
(2002) used 1 mg N per 100 g. By using any of these

limits, samples of sea bream and sea bass were acceptable during the total period of chilled storage. The
inexistence of an association with the reduction of
sensorial quality hindered, however, the use of TMA-N
as a good-quality index in the present study.
Taking into account that the reduction of TMA-O is a
property of Gram-negative microorganisms (Lannelongue et al., 1982), the low rate of TMA production in the
vacuum-packaged llets was most likely as a result of a
reduction in growth of aerobic, Gram-negative bacteria,
including TMA-producing micro-organisms, by CO2
application during SGS and by low O2 atmosphere in
vacuum packaging.
The non-existence of a direct response with time of
TVB-N and TMA-N changes during the chilled storage
of sea bream and sea bass llets prevents the use of these
parameters as quality indicator for these species in the
packaging conditions studied. Similar results have also
been reported by other authors in Sparidae species
which is ice stored (Kyrana et al., 1997; Huidobro et al.,
2001; Cakli et al., 2006).
Thiobarbituric acid reactive substances
For the assessment of the degree of lipid oxidation,

determination of the changes in the TBA index has been
recommended (Connell, 1975; Nishimoto et al., 1985).
TBA values of sea bream and sea bass (data not shown)
were at the beginning 0.45 and 0.02 lmol g)1 fat,
respectively. At the end of the storage period of 15 days,
TBA values of sea bream llets ranged from 0.21 to
0.32 lmol g)1 fat and sea bass TBA values ranged from
0.04 to 0.08 lmol g)1 fat. Levels were always under the
optimal limit (12 lmol g)1 fat) proposed by Connell
(1975) for this parameter, and therefore, no signs of
oxidation spoilage were determined. Low levels of TBA
index were also reported by Grigorakis et al. (2004) in
sea bass after 15 days of ice storage (1.48 lmol kg)1
tissue equivalent to 0.033 lmol g)1 fat).
SGS treatment did not show any inuence in the TBA
values. In the previous study, with sea bream and sea
bass llets stored in air (Mendes & Goncalves, in press),
higher TBA values were determined namely in sea
bream llets (0.391.23 lmol g)1 fat), therefore showing
that vacuum packaging, by removing atmospheric
oxygen from the contact with lipids, has a positive
eect on the prevention of oxidation spoilage.
Sensory evaluation
The results of the sensory evaluation of vacuumpackaged sea bream and sea bass llets are presented
in Tables 2 and 3. As expected, the sensory scores of the
three batches from both species decreased with storage
time.
Table 2 Sensory changes during chilled storage (1.0 0.5 C) of

vacuum-packaged sea bream llets treated with SGS
Storage time (days)
Sensory parameter
Batches
13
15
Control
SGS 30
SGS 60
Control
SGS 30
SGS 60
4.8
4.8
4.8
4.1
4.1
4.1
3.8
4.0
4.0
3.3
3.3
3.1
3.3
3.3
3.5
2.9
2.5
2.9
3.3
3.3
3.3
2.6
2.8
3.1
2.5
3.3
3.0
2.3
2.4
2.6
2.8
2.8
3.0
2.6
2.3
2.0
Control
SGS 30
SGS 60
Control
SGS 30
SGS 60
Control
SGS 30
SGS 60
9.3
9.3
9.3
9.5
9.5
9.5
4.3
4.3
4.3
9.0
8.5
8.8
9.0
8.5
8.3
3.5
3.5
3.5
6.3
6.8
6.8
5.5
6.5
7.8
2.8
3.0
3.3
6.5
6.8
6.8
5.3
6.5
6.3
2.9
2.8
3.0
5.3
6.0
5.8
5.3
5.3
5.5
2.3
2.4
2.5
5.0
5.0
5.0
4.3
4.3
4.3
2.0
2.1
2.3
Raw
Odour
Colour
Cooked
Odourb
Flavourb
Texturec
Values correspond to the average of four panelists (each assessor

evaluated one fillet). Mostly, standard deviation was in the range of 0.5
1.0, except for texture which was 0.6.
SGS 30 and SGS 60, fillets pre-treated with SGS (2 bar 100% CO2, for 30
and 60 min, respectively); Control, fillets without SGS pre-treatment.
a
Category scale (15): 1, putrid odour evident discolouration; 5, typical
fresh odour characteristic colour.
b
Torry scale (103): 10, very fresh fish, with weak odour sweet characteristic flavour; 5, slightly sour rancid odour, trace of off-flavours; 3,
putrid or spoiled fish.
c
Category scale (15): 1, soft dry; 5, firm succulent.
As expected, the most sensitive sensory attribute for

quality evaluation of a sh product would be odour.
Based on odour scores, raw sea bream llets treated
with SGS received an odour score higher or similar
than the acceptability limit (score 3.0) up to 13 days,
whereas control sample was scored at the limit of
acceptability after 9 days and was rejected on the 13th
day. Sea bass llets presented a similar behaviour in all
samples and also an identical shelf life of 13 days.
Sivertsvik (2000) reported a positive eect of joint SGS
treatment and vacuum packaging on the preservation
of an acceptable odour during chilled storage also on
salmon llets.
Colour scores of sea bream and sea bass llets
decreased in all SGS treatments with increasing time
of chilled storage. Sea bass llets (Table 3) kept the
characteristic colour of fresh products until the 6th day
of storage. On the contrary, sea bream llets started to
loose the characteristic abdominal pink and dorsal grey
colours immediately after the 1st day of storage. SGS
had a little eect on the colour of llets during chilled
storage.
Table 3 Sensory changes during chilled storage (1.0 0.5 C) of

vacuum-packaged sea bass llets treated with SGS
Storage time (days)
Sensory parameter
Batches
13
15
Control
SGS 30
SGS 60
Control
SGS 30
SGS 60
4.5
4.5
4.5
4.0
4.0
4.0
4.3
4.0
4.0
4.0
4.3
4.0
3.8
3.8
3.8
4.0
3.8
3.8
3.3
3.0
3.5
3.0
3.3
3.5
3.3
3.5
3.3
3.5
3.0
3.3
2.5
2.8
2.5
3.0
2.5
2.5
Control
SGS 30
SGS 60
Control
SGS 30
SGS 60
Control
SGS 30
SGS 60
9.5
9.5
9.5
9.8
9.8
9.8
4.4
4.4
4.4
9.0
8.8
8.8
8.8
8.0
7.5
3.6
3.3
3.4
7.3
7.3
7.5
7.0
7.8
7.0
3.8
3.6
3.5
6.8
7.0
6.8
7.0
6.5
6.5
2.8
3.0
3.1
7.3
7.5
6.5
7.0
7.0
5.8
3.3
3.1
2.9
5.5
6.5
7.0
4.8
5.3
6.0
2.8
2.8
3.0
Raw
Odour
Colour
Cooked
Odourb
Flavourb
Texturec
Values correspond to the average of four panelists (each assessor

evaluated one fillet). Mostly, standard deviation was in the range of 0.5
1.0, except for texture which was 0.6.
SGS 30 and SGS 60, fillets pre-treated with SGS (2 bar 100% CO2, for 30
and 60 min, respectively); Control, fillets without SGS pre-treatment.
a
Category scale (15): 1, putrid odour evident discolouration; 5, typical
fresh odour characteristic colour.
b
Torry scale (103): 10, very fresh fish, with weak odour sweet characteristic flavour; 5, slightly sour rancid odour, trace of off-flavours; 3,
putrid or spoiled fish.
c
Category scale (15): 1, soft dry; 5, firm succulent.
Torrieri et al. (2006) reported also the development of

a yellowish colour in sea bass packaged in air, whereas
samples in MAP (50% and 70% CO2) were not so much
aected by chilled storage, thus keeping the characteristic colour for a longer period. Eectiveness of MAP
(50% CO2 50% N2) in comparison to air and vacuum
packaging was also shown by Gimenez et al. (2002) on
the maintenance of sea bream llets colour after 22 days
of chilled storage.
Cooked products were sensory rejected on odour
(score 5.5) after 13 (control) and 15 days (SGS
batches) in the case of sea bream llets (Table 2) and
were still acceptable at the end of the chilled storage in
the case of sea bass llets (Table 3). At the end of the
storage, sea bass llets were characterised as having
neutral odour, whereas control batch was characterised
as starting to develop acid and sour odours. Regarding
the same parameter, Mendes & Goncalves (in press)
showed on the same sh species which were packaged in
air and also SGS treated 2 days less of shelf life.
Removal of O2 from the packages by vacuum packaging
may explain the positive eect on the inhibition of
rancid odours.
According to avour, shelf life of sea bream llets was

extended 3 days (from day 6 to day 9) when treated for
30 min in CO2 and 7 days (from day 6 to day 13) when
CO2 was applied for 60 min.
In the case of sea bass llets, changes in avour were
less noticeable and shelf life extension was only visible in
the samples which were SGS treated for 60 min and
rejection was delayed 2 days, from day 13 to 15. Rancid
avours were not detected during chilled storage of
vacuum packaging llets. The use of CO2 and particularly the vacuum packaging had a positive eect on
avour, although this eect was particularly noticeable
on the sea bream llets chilled stored. Similar results
reported by Randell et al. (1999) observed the inhibition
of rancid taste and avour when O2 was removed from
the packages in lleted salmon. Sivertsvik (2000) also
reports that the removal of oxygen (vacuum packaging)
is more important than the level of CO2 in what
concerns odour and avour of salmon.
In relation to texture, the average value of succulence
and rmness was considered to better characterise this
parameter. Lower scores than the limit of acceptability
(2.0) were never determined during chilled storage of
sea bream and sea bass llets. Although considerable
dierent SGS-treated samples during extended time
(60 min) did not show higher absolute texture scores.
Torrieri et al. (2006) reported that texture of sea bass
packaged in MAP with 5070% CO2 remained unchanged during chilled storage, whereas control samples with O2 higher than 20% showed a decrease in
texture scores after the 7th day. Sivertsvik et al. (2004)
report that MAP may aect the sensory quality of
products, namely, succulence because of the reduction
of the water retention capacity caused by the decrease
of pH induced by solubilisation of CO2 in tissues. With
the SGS procedure and vacuum packaging, the reduction of pH was, however, not relevant, nor the
development of drip or the loss of succulence, which
may be accounted as advantages of this process in
relation to the use of high CO2 concentrations during
MAP chilled storage.
In sea bream, stored in ice or packed, with or without
O2 absorber, Goncalves et al. (2004) reported an
acceptability limit till 10 days. Gimenez et al. (2002)
also reported that sea bream llets packaged in vacuum
were acceptable for 22 days and kept better quality
during chilled storage than when in MAP with 2030%
O2 but worse than in MAP with 010% O2. Papadopoulos et al. (2003) reported a shelf life of 13 and 8 days
for ungutted and gutted sea bass stored in ice. In
ungutted sea bass, Poli et al. (2001) determined a shelf
life of 10 days.
Previous work on air packaging of SGS-treated sea
bream and sea bass llets (Mendes & Goncalves, in
press) showed rejection based on odour to occur 4 days
before it was seen in this study. Comparison with the
1685
1686
present data shows that vacuum packaging had also a

positive eect on the preservation of the characteristic
colour and all other sensory attributes of raw and
cooked sea bream and sea bass llets. Vacuum-packaged
products were still acceptable at the end of the chilled
storage, whereas air-packaged llets of both species were
rejected on account of the yellowish colour after 9 days.
Therefore, SGS together with vacuum packaging leads
also to an increased shelf life of the llets when
compared with SGS and chilled storage in packages
with air.
Conclusions
Apart from pH value that showed a regular increase

during chilled storage, the other chemical indices (TVBN, TMA-N and TBA) showed to be poor indicators of
changes in quality of products. Final TVB-N values
ranged from 16.0 to 17.4 mg N per 100 g and from 17.3
and 19.4 mg N per 100 g in sea bream and sea bass,
respectively.
SGS pre-treatment kept acceptable sensory quality of
llets for longer time, which was particularly visible on
the sea bream llets. Thus, based on a sensory criterion,
SGS lead to an extension of the shelf life of the chilled
llets in 23 days. At the end of storage, bacterial
counts followed the order 60 min SGS < 30 min
SGS < control in both species. Treatment of llets in
100% CO2 at 2 bar inhibited the microbial growth to
levels below the limit of acceptability during the total
period of storage. The low microbiological contamination and low TVB-N and TMA-N content derived from
bacterial metabolism may be as the result of the low pH
values determined.
Acknowledgments
The present study was nancially supported by EUQCA III-MARE FEDER: Project Quality and Innovation of Fishery Products. The authors would like to
thank the farm TIMAR (Tavira, Portugal) for the
supply of sh. They would also like to thank Alexandra
Gameiro, Carla Pestana and Soa Antas for their
technical assistance on the analyses.
References
Andersen, U.B., Thomasssen, M.S. & Rra, A.M.B. (1997). Texture
properties of farmed rainbow trout (Oncorhynchus mykiss): eects of
diet, muscle fat content and time of storage on ice. Journal of the
Science of Food and Agriculture, 74, 347353.
AOAC (1998). Ofcial Methods of Analysis of AOAC International,
16th edn, 4th Rev. Gaithersburg: Association of Ocial Analytical
Chemists International, Inc., CD-ROM.
Ashie, I.N.A., Smith, J.P. & Simpson, B.K. (1996). Spoilage and selflife extension of fresh sh and shellsh. Critical Reviews Food
Bandarra, N., Calhau, M.A., Oliveira, L. et al. (2004). Composicao e

valor nutricional dos produtos da pesca mais consumidos em
Portugal. Publicacoes Avulsas do IPIMAR, Lisboa, 11, 103.
Banks, H., Nickelson, R. & Finne, G. (1980). Shelf life studies on
carbon dioxide packaged nsh from the Gulf of Mexico. Journal of
Cakli, S., Kilinc, B., Cadun, A., Dincer, T. & Tolasa, S. (2006). Eects
of gutting and ungutting on microbiological, chemical, and sensory
properties of aquacultured sea bream Sparus aurata and sea bass
Dicentrarchus labrax stored in ice. Critical Reviews in Food Science
and Nutrition, 467, 5195279.
Castro, P., Padern, J.C.P., Cansino, M.J.C., Velasquez, E.S. & De
Larriva, R.M. (2006). Total volatile base nitrogen and its use to
assess freshness in European sea bass stored in ice. Food Control, 17,
245248.
Cobb, B.F. III, Alaniz, I. & Thompson, C.A. Jr (1973). Biochemical
and microbial studies on shrimp: volatile nitrogen and amino
nitrogen analysis. Journal of Food Science, 38, 431436.
Connell, J.J. (1975). Methods of assessing and selecting for quality. In
Control of Fish Quality. Editor: J.J. Connell. Pp. 107132. Surrey:
Fishing News Books Ltd.
Connell, J.J. (1995). Control of Fish Quality, 4th edn. Pp. 245. London:
Fishing News Books Limited.
Crank, J. (1975). The Mathematics of Diffusion, 2nd edn. Oxford:
Clarenden Press, Pp. 426.
Debevere, J. & Boskou, G. (1996). Eect of modied atmosphere
packaging on the TVB TMA-producing microora of cod llets.
EEC (1995). Decision 95 149 EC. Total volatile basic nitrogen TVBN limit values for certain categories of shery products and
specifying the analysis methods to be used. Ofcial Journal, L097,
8487.
Farber, L. (1965). Review of the Method for the Determination of
Spoilage of Fish. Pp. 5870. London: Fishing News.
Farber, J.M. (1991). Microbiological aspects of modied-atmosphere
packaging technology a review. Journal of Food Protection, 541,
5870.
FDA (2001). Clostridium botulinum toxin formation (a biological
hazard) Ch. 13. In Fish and Fishery Products Hazards and Controls
Guidance, 3rd edn. Pp. 83102. Washington: Food and Drug
Administration, Center for Food Safety and Applied Nutrition,
Oce of Seafood. http://www.cfsan.fda.gov/-comm/haccp4a.html
[accessed 2 Nov 2007].
Genigeorgis, C.A. (1985). Microbial and safety implications of the use
of modied atmospheres to extend the storage life of fresh meat and
sh. International Journal of Food Microbiology, 1, 237251.
Gimenez, B., Sanchez-Escalante, A., Torrescano, G., Roncales, P. &
Beltran, J.A. (2002). Dierent packaging conditions to improve shelf
life of lleted gilt-head sea bream (Sparus aurata). Journal of Aquatic
Food Product Technology, 11, 275286.
Goncalves, A., Mendes, R. & Nunes, M.L. (2004). Eect of oxygen
absorber on the shelf life of gilthead seabream Sparus aurata.
Journal of Aquatic Food Product Technology, 133, 4959.
Goulas, A.E. & Kontominas, M.G. (2007). Combined eect of light
salting, modied atmosphere packaging and oregano essential oil on
the shelf-life of sea bream Sparus aurata: biochemical and sensory
attributes. Food Chemistry, 100, 287296.
Gram, L. & Huss, H.H. (1996). Microbial spoilage of sh and sh
products. International Journal of Food Microbiology, 33, 121137.
Grigorakis, K. (2007). Compositional and organoleptic quality of
farmed and wild gilthead sea bream (Sparus aurata) and sea bass
(Dicentrarchus labrax) and factors aecting it: a review. Aquaculture, 272, 5575.
Grigorakis, K., Alexis, M., Gialamas, I. & Nikolopoulou, D. (2004).
Sensory, microbiological and chemical spoilage of cultured common
sea bass Dicentrarchus labrax stored in ice: a seasonal dierentiation. European Food Research Technology, 219, 584587.
Howgate, P. (1985). Approaches to the denition and measurement of

the storage life of chilled and frozen sh. In: Proceeding of
International Institute of Refrigeration Commissions C2 and D3
(Edited by Institut International du Froid). Pp. 4553. Aberdeen.
Huidobro, A., Mendes, R. & Nunes, M.L. (2001). Slaughtering of
gilthead seabream Sparus aurata in liquid ice: inuence on sh
quality. European Food Research and Technology, 213, 267272.
Huss, H.H. (1995). Quality and quality changes in fresh sh. FAO
Fisheries Technical Paper, 348. Rome: FAO, Pp. 195.
INFOFISH (2007). Seabass and Seabream Market Report May
2007. Technical and trade database. Available at http://www.infosh.org/marketreports/sas0507.html [Accessed on 14 December
2007].
Kyrana, V.R., Lougovois, V.P. & Valsamis, D.S. (1997). Assessment
of shelf-life of maricultured gilthead sea bream Sparus aurata stored
in ice. International Journal of Food Science and Technology, 32, 339
347.
Lannelongue, M., Finne, G., Hanna, M.O., Nickelson, R. & Vanderzant, C. (1982). Microbiological and chemical changes during
storage of swordsh (Xiphias gladius) steaks in retail packages
containing CO2 enriched atmospheres. Journal of Food Protection,
45, 11971203.
Lowndes, A.G. (1955). Density of shes. Some notes on the swimming
of sh to be correlated with density, sinking factor and the load
carried. Annals and Magazine of Natural History, 8, 241256.
Masniyom, P., Benjakul, S. & Visessanguan, W. (2002). Shelf life
extension of refrigerated sea bass slices under modied atmosphere
packaging. Journal of the Science of Food and Agriculture, 82, 873
880.
Masniyom, P., Benjakul, S. & Visessanguan, W. (2005). Collagen
changes in refrigerated sea bass muscle treated with pyrophosphate
and stored in modied- atmosphere packaging. European Food
Research and Technology, 220, 322325.
Meilgaard, M., Civille, G.V. & Carr, G. (1999). Sensory Evaluation
Techniques, 3rd edn. Pp. 281. London: CRC Press.
Mendes, R. & Goncalves, A. (2008). Eect of soluble CO2 stabilization
on the quality of llets from farmed gilthead sea bream (Sparus
aurata) and European sea boss (Dicentrarchus labrax). Journal of
Aquatic Food Product Technology, 17, In press.
Nishimoto, J., Suwetja, I.K. & Miki, H. (1985). Estimation of keeping
freshness period and practical storage life of mackerel muscle during
storage at low temperatures. Memoirs of the Faculty of Fisheries
Kagoshima University, 34, 8996.
Ozogul, F., Polat, A. & Ozogul, Y. (2004). The eect of modied
atmosphere packaging on chemical, sensory and microbiological
changes of sardines (Sardina pilchardus). Food Chemistry, 85, 4957.
Ozogul, F., Gokbulut, C., Ozyurt, G., Ozogul, Y. & Dural, M. (2005).
Quality assessment of gutted wild sea bass (Dicentrarchus labrax)
stored in ice, cling lm and aluminium foil. European Food Research

Papadopoulos, V., Chouliara, I., Badeka, A., Savvaidis, I.N. &
Kontominas, M.G. (2003). Eect of gutting on microbiological,
chemical, and sensory properties of aqua-cultured sea bass Dicentrarchus labrax stored in ice. Food Microbiology, 20, 411420.
Poli, B.M., Parisi, G., Zampacavallo, G., Mecatti, M., Lupi, P. &
Gualtieri, M. (2001). Quality outline of European sea bass Dicentrarchus labrax reared in Italy: shelf life, edible yield, nutritional and
dietetic traits. Aquaculture, 202, 303315.
Randell, K., Hattula, T., Skytta, E., Sivertsvik, M., Bergslien, H. &
Ahvenainen, R. (1999). Quality of lleted salmon in various retail
packages. Journal of Food Quality, 22, 485497.
Ruiz-Capillas, C. & Moral, A. (2005). Sensory and biochemical aspects
of quality of whole bigeye tuna Thunnus obesus during bulk storage
in controlled atmospheres. Food Chemistry, 893, 347354.
Sivertsvik, M. (2000). Use of soluble gas stabilisation to extend
shelf-life of sh. In: Proceedings of 29th WEFTA-Meeting (edited by
S.A. Georgakis). Pp. 7991. Leptocarya, Pieria, Greece, 1014
October 1999. Thessaloniki: Greek Society of Food Hygienists and
Technologists.
Sivertsvik, M., Jeksrud, W.K. & Rosnes, J.T. (2002). A review of
modied atmosphere packaging of sh and shery products
signicance of microbial growth, activities and safety. International
Sivertsvik, M., Rosnes, J.T. & Jeksrud, W.K. (2004). Solubility and
absorption rate of carbon dioxide into non-respiring foods. Part 2:
Raw sh llets. Journal of Food Engineering, 63, 451458.
Soccol, M.C.H., Oetterer, M., Gallo, C.R., Spoto, M.H.F. & Biato,
D.O. (2005). Eects of modied atmosphere and vacuum on the
shelf life of tilapia (Oreochromis niloticus) llets. Brazilian Journal of
Food Technology, 8, 715.
Taliadourou, D., Papadopoulos, V., Savvaidis, I.N. & Kontominas,
M.G. (2003). Microbiological, chemical and sensory changes of
whole and lleted ice stored Mediterranean aqua-cultured sea bass
Dicentrarchus labrax. Journal of the Science of Food and Agriculture,
83, 13731379.
Tejada, M. & Huidobro, A. (2002). Quality of farmed gilthead
seabream Sparus aurata during ice storage related to the slaughter
method and gutting. European Food Research and Technology, 215,
17.
Torrieri, E., Cavella, S., Villani, F. & Masi, P. (2006). Inuence of
modied atmosphere packaging on the chilled shelf life of gutted
farmed bass Dicentrarchus labrax. Journal of Food Engineering, 77,
10781086.
Vyncke, W. (1970). Determination of the ammonia content of sh as
an objective quality assessment method. Medelingen van de Faculteit
Landbouwwetenschappen, Rijkauniversteit Gent, 35, 10331046.
1687
1688
Original article
The effect of cutting and fish-orientation systems on the deheading
yield of carp
Andrzej Dowgiallo*
Department of Processing Technology, Sea Fisheries Institute, Kottataja 1 Str., 81-332, Gdynia, Poland
(Received 4 January 2008; Accepted in revised form 28 March 2008)
Summary
Applied research into carp-deheading yield indicated that the V-cut with two circular knives averaged 77.9%;
the V-cut with one cup-type knife 75.6%, and the straight cut at a 79 angle to the sh backbone 77.4%.
The yield averages for deheaded and gutted carp were 63.6%, 62.4% and 62.9%, respectively. Standard
analysis of variance demonstrated that there were no statistically signicant dierences between the mean
yields of these three deheading systems. Furthermore, the potential inuence of the sh-orientation system in
close connection with the cutting systems on the deheading yield was analysed. This indicated that, with the
same yields, the straight-cutting system simplies the precise orientation of the sh in relation to the cutting
knives.
Keywords
Carp, deheading, orientation, yield.
Introduction
A survey conducted in 2005 among participants of the

Tenth Polish Conference of Carp Farmers indicated that
there is a need for an inexpensive carp-deheading
machine that leaves the collar bones and pectoral ns
with the head, and produces as high a deheading yield as
possible.
A research and development project for constructing
such a machine was undertaken at the Sea Fisheries
Institute in Gdynia. The project contractor was the
Agency for the Restructuring and Modernisation of
Agriculture. As stipulated, the project began by determining the most ecient carp-deheading system.
A review of the available literature on carp processing
was done, and only one article on the topic of deheading
yield was found (Okoniewska & Okoniewski, 1969).
Based on their own research, these authors reported that
the mean carp-deheading yield was 62.3% with the Vcut (similar to that performed with two circular knives).
There is however a lack of data on the carp-deheading
yield with other systems. For this reason, the actual
deheading machine design was preceded by investigations to determine the yields.
Deheading cutting systems
Two dierent cutting systems are commonly used to

dehead sh; the straight cut, performed either perpen*Correspondent: E-mail: techmech@mir.gdynia.pl
dicular or at an angle to the sh backbone, and the more

complicated V-cut, used when higher machine costs are
compensated for by higher deheading yield. In both
systems, the collar bones remain either with the head or
the body. In general, a straight cut is made with one
circular knife, while a V-cut is made with two circular
knives. V-cut deheading can also be done with a single
cup-type knife (Dowgiallo & Dutkiewicz, 2007).
The cutting lines, determined before yield measurements, are presented in Fig. 1. The classical V-cut lines
(Fig. 1a) are determined by:
1 the line tangential to the gill cover, running close behind
the pectoral n root, and
2 the line tangential to the collar bone, running close
behind the pectoral n root.
The angle between the cutting lines is 122. The
supraoccipital bone (a median bone on the upper rear
end of the cranium Fig. 2) is considered to be a part of
the tissue near the upper part of the gill cover. For this
reason, leaving this tissue with the head decreases the
deheading yield signicantly, but it makes lleting sh
easy.
The V-cut lines for this type of deheading with one
cup-type circular knife (Fig. 1b) are determined by: (i)
the line perpendicular to the sh backbone that runs
close behind the root of the pectoral n up to the
backbone; and (ii) the line tangential from the sh
backbone to the gill cover. The angle between the
doi:10.1111/j.1365-2621.2008.01750.x
Orientation and deheading systems of carp A. Dowgiallo
Figure 1 Carp head and deheading cut lines:

(a) V-cut with two circular knives; (b) V-cut
with one cup-type circular knife; (c) straight
cut at an angle to the sh backbone.
The yield of deheading is expressed by mass ratios of

deheaded and whole carp. The mean yield of each
described deheading system was determined based on 12
measurements performed on carp from 330 to 535 mm
in length. All measurement series consisted of: (i)
weighing the whole carp; (ii) manually deheading in
one of described systems; (iii) weighing deheaded carp;
(iv) manually gutting deheaded carp; and (v) weighing
deheaded and gutted carp. The deheaded and gutted
carps are shown in Fig. 3.
Figure 2 Carp cranium. Arrow indicates the supraoccipital bone.
cutting lines is 152. The positioning of these cutting

lines results not only from the prevailing conditions, but
also from the physical conditions of deheading with one
knife (Dowgiallo & Dutkiewicz, 2007).
Figure 1c presents the cutting line for straight deheading. The line is tangential to the gill cover and runs
close behind the root of the pectoral n. This line runs at
a 79 angle to the sh backbone.
The mean values and 95% condence intervals of

yields obtained with the described deheading systems
are presented in Fig. 4. The values of the mass ratios
of deheaded and whole carp deheading yields are
aected by the gut content, which is why the
deheading and gutting yields were also determined.
The mean values and 95% condence intervals are
shown in Fig. 5.
Standard analysis of variance indicated that there is
no statistically signicant dierence between the dierent mean deheading yields at a 95% condence level.
There was also no statistically signicant dierence in
the mean yields of deheaded and gutted carp. This is
why the eciency of deheading is only negligibly
dependent on the cutting system, and the results
Figure 3 Deheaded and gutted carps: (a, b)

V-cut deheading; (c) straight cut deheading
at an angle to sh backbone.
1689
1690
Figure 4 Mean values and 95% condence intervals of deheading

yields: (A) V-cut deheading like with two circular knives; (B) V-cut
deheading like with one cup-type circular knife; (C) straight cut
deheading at an angle to sh backbone like with one circular knife.
dierence in stiness between the backbone and the

muscle tissues provides the proper shape of the cutting
line (Dowgiallo & Dutkiewicz, 2007). However, this
deheading system requires a relatively complicated
mechanism to modify sh position during deheading
(Dowgiallo & Dutkiewicz, 2007), and its application increases deheading costs and complicates the
operation.
In order to determine the possibility of applying a
single carp orientation operation during deheading with
the two remaining systems, an analysis of the external
sh anatomy was conducted. It was revealed that the
same position can be obtained consistently based on
constant elements of the sh body, including the same
sections of the dorsal or ventral carp contours and the
position of pectoral n roots in a single line a. The
orientation systems are shown in Fig. 6. The same
segments of the ventral contour in carp from 350 to
535 mm in length are about half as much as the dorsal
ones (c. 1.5b), which is why the carp orientation
system shown in Fig. 6b is more stable.
Its application in straight-cut deheading at an angle a
and V-cut deheading is shown in Fig. 7. This type of
orientation system is sucient with straight-cut deheading (Fig. 7a), whereas V-cut deheading with two circular
knives requires an additional sh-positioning correction
during processing. Fish smaller than the largest processed sh, which determines the zero position, must
shift Dh mm so that the distances of points O and O
Figure 5 Mean values and 95% condence intervals of deheading and

gutting yields: (A) V-cut deheading like with two circular knives; (B)
V-cut deheading like with one cup-type circular knife; (C) straight cut
deheading at an angle to sh backbone like with one circular knife.
obtained during manual processing are theoretical and

the maximal obtainable during mechanised processing.
The processing yield, regardless of the deheading
system, is aected by the proper orientation of sh in
relation to the deheading knife or knives. Departures
from the theoretical orientation cause either decreased
yield or incorrect deheading. This is why, with a lack of
signicant dierences between deheading yield, the
decisive factor in choosing a deheading system was the
systems sh-orientation component, which should be
simple and independent of sh size.
Fish-orientation systems
When deheading with a single cup-type knife, the shorientation system is very simple and the considerable
Figure 6 Same position of carp orientation based on dorsal (a) or

ventral (b) contours and pectoral n position (carp from 330 to
535 mm in length). a, line of pectoral n position; b and c, the same
segments of body contour.
Figure 7 Application of carp-orientation

system: (a) straight deheading cut; (b) V-cut
deheading with two knives. c, cutting line; Dh,
shift; 1, lengthwise retainer; 2, crosswise
retainer.
Figure 8 Example of measurements of chosen

carp morphometric parameters.
from line x are equal (Fig. 7b). In the present study

(total carp length in the range of 350535 mm), the
maximal shift was 22 mm (Fig. 7b).
Position correction requires additional measuring
mechanisms, either directly or indirectly, the value of
the required shift and sh displacement. While direct
measurements are very precise, they rely on visual
techniques, and this renders them relatively expensive.
The indirect measurement of the required shift can be
approximated with any sh body parameter that is
connected through a morphometric relationship (necessary condition), and is easily measurable (desirable
condition). Although indirect measurements based on
morphometric relationships are considerably cheaper,
they are burdened with prediction error. In order to
clarify this point, consider the relationship between
maximal thickness b of sh, which is a readily measurable morphometric parameter, and the corresponding
value of Dh. This relationship was obtained from

measurements conducted according to the diagram in
Fig. 8 on a sample size of n = 36.
Shifts in the value of Dh were calculated with pattern
Dhi = h - hi, and after statistical processing, the following morphometric relationship was obtained:
h 50:304 0:505bR2 88:38%
Figure 9 shows the regression line and 95% prediction
limits, which provide a valid statistical estimate of the
range of Dh values to be expected. This demonstrates
clearly that the dierence between the theoretical values
of Dh and new observations are about 5 mm. This will
result in either incorrect deheading or decreased yield.
Moreover, the application of a mechanism for correcting sh position during processing complicates the
design of the deheader and increases its cost. Correction
1691
1692
Figure 9 Regression line h = f(b) with observed values and 95%

prediction limits.
mechanism inertia can also decrease the throughput of

the deheading machine.
Conclusions
Considering the following aspects:

the yield differences between V-cut and straight deheading were negligible,
the efciency of the sh-orientation system,
It was determined that in the case of carp, straight-cut

deheading at an angle to the sh backbone is the most
advantageous.
Therefore, despite the satisfactory results of deheading with one cup-type knife (Dowgiallo & Dutkiewicz,
2007), a continuous process deheading machine for
straight cutting was designed. The main advantage of
such a machine is the constant position of the sh with
respect to the knife, which not only simplies its
construction, but most importantly eliminates losses in
deheading yield, which is the principle criterion for
evaluating its usefulness. In comparison with the V-cut
deheading machine with two circular knives, the only
drawback is the necessity of frequently sharpening the
deheading knife, which becomes dull faster than in the
two-knife system.
This simple and inexpensive machine has already been
through a successful trial at a sh-processing plant.
References
Dowgiallo, A. & Dutkiewicz, D. (2007). Possibilities of utilizing the
dierences of sh tissues stiness in the mechanization of cyprinid
deheading. Journal of Food Engineering, 83, 111115.
Okoniewska, Z. & Okoniewski, Z. (1969). Preliminary results of
investigations on the weight of commercially valuable body parts
and the proximal composition of herbivorous sh and carp (Wyniki
wstepnych badan ciez_ aru czes ci u_zytkowych i skadu chemicznego
ciaa ryb ros lino_zernych i karpia). Roczniki Nauk Rolniczych Polskiej
Akademii Nauk. Seria H Rybactwo. Tom, 91, 385401.
Original article
Effect of fortification of defatted soy flour on sensory and
rheological properties of wheat bread
Morteza Mashayekh,1* Mohammad Reza Mahmoodi2 & Mohammad Hasan Entezari3
1 Food Science and Technology Department, School of Nutrition and Food Technology, Shaheed Beheshti University of Medical Sciences,
Tehran, Iran
2 Department of Nutrition, School of Health, Kerman University of Medical Sciences, Kerman, Iran
3 Department of Nutrition, School of Health, Isfahan University of Medical Sciences, Isfahan, Iran
(Received 8 October 2007; Accepted in revised form 15 April 2008)
Summary
The objective of this study was to determine the eects of soy-fortied bread on the sensory and rheological
properties. Ground defatted soy our was blended with wheat our at 3%, 7% and 12%. The organoleptic
characteristics of soy-fortied wheat breads were carried out by taste panel. The eect of this fortication on
the rheological properties of the resulting dough was investigated using farinograph and extensograph for
quality assessment of the nal product. The ash and protein contents of 3% and 7% wheatsoy bread blends
increased compared with control. The results revealed that organoleptic characteristics score such as
bendability, appearance, avour and taste, crust texture and overall acceptability properties of bread
containing 3% defatted soy our was highest even though it is not signicantly dierent. Therefore, we
conclude that adding 3% or 7% defatted soy our actually gives as good a loaf of bread as the 100% wheat
bread with higher nutritional quality and acceptable consumer attitude with rheological and sensory
characteristics.
Keywords
Defatted soy-fortied bread, sensory, rheology, wheat our.
Introduction
One of the nutritional problems in most of the developing

world is protein-calorie malnutrition. The strategic use of
inexpensive high protein resources that complement the
amino acid pattern of cereal-staple foods is highly
recommended to upgrade the nutritional status of many
people. Thus, fortication of bread with soybean our
can dramatically improve their protein quality, with only
a slight increase in the production cost (El-Adawy, 1997).
It is an excellent source of bre, and a good source of
lecithin, vitamins, minerals and lignins, which may also be
of benet in the prevention of cardiovascular disease and
cancers (Ames et al., 1993; Rimm et al., 1993; Stampfer
et al., 1993; Giovannucci et al., 1995; Scheiber et al.,
2001; Castle & Thrasher, 2002; Hasler, 2002).
The increasing importance of various types of bakery
products in todays eating habits suggests that these
food products can serve as vehicles for important
nutrients while being readily accepted by consumers.
In recent years, considerable interest has been generated in the development and consumption of foods
*Correspondent: Fax: (98 21) 22 357 483 5;
e-mail: mortezamashayekh@yahoo.com; m.mashayekh@nnftri.ac.ir
doi:10.1111/j.1365-2621.2008.01755.x
enriched with various healthy ingredients. Adding soy

ingredients to baked products inuences textural and
sensory properties (Vittadini & Vodovotz, 2003). It is
well established in the literature that oil seed our
change wheat ours mixing and other properties.
Matthews et al. (1970) concluded that oil seed our
added to wheat our increased the farinograph water
absorption, reduced dough tolerance and produced
bread with lower loaf volume. The changes in the
physicochemical properties of bread containing soy
our were reported by Shiraldi et al. (1996) who found
a signicant reduction of bread staling caused by soy
our. Reports in the literature conclude that high
protein breads contain up to 1520% protein (Mohamed et al., 2006).
Several studies have indicated the possibility of
incorporating soybean (Rastogi & Singh, 1989; Dhingra
& Jood, 2004) into wheat our at various levels and the
rheological and baking properties have been reported.
Such composite our can be used for producing bread,
biscuits and other snacks.
However, information is scanty on the supplementation of wheat our with combinations of defatted soy
our for bread making and therefore the present
investigation was undertaken to study the eect of
1693
1694
Effect of fortification of defatted soy flour on wheat M. Mashayekh et al.
supplementation with 3%, 7% and 12% defatted soy

our on the functional, baking, organoleptic and
rheological properties of nal products and acceptable
bread.
content of control and soy fortied bread was measured

by micro-Kjeldahl method No. 14.026 (AOAC, 1995)
and was converted to protein value by using a factor of
5.7. Phytic acid analysis was performed according to the
method of Latta & Eskin (1980). Chemical composition
(ash, protein, fat and phytic acid) was measured as dry
basis.
Preparation of flour blends and bread
Ground defatted soy our purchased from oil extraction

and renery factory (Naz Isfahan, Isfahan, Iran) was
blended with wheat our at 3%, 7% and 12% levels for
10 min to ensure homogeneity using a mixer (Oghab
factory, Tehran, Iran). Wheat our alone was used as
control. The control bread (Taftoon) was bought from the
university bakery and the samples with defatted soy our
blends were baked in the same bakery with the same
procedure. The wheat were grown and harvested in Iran
homeland and mixed together. Taftoon is sour-dough at
bread, round in shape (about 30 cm diameter) with small
holes on its surface which is consumed in Iran. The bread
was prepared by mixing all ingredients such as our with
extraction level of 8284%, rened salt-free iodine (12%
w w), water (5560% w w) and bakery yeast as leavening
agent to optimum. The chemical composition of defatted
soy our and wheat our is given in Table 1. To prepare
fermented dough, two consecutive fermentations were
carried out using 3% inocula (called sour dough that
separated from a previous batch) from a previous
fermentation, to start fermentation of each subsequent
batch. After fermentation for 90 min, about 450 g balls
were made and sheeted in a round shape, and punctured
to prevent pung during oven baking, as well as for
decorative purposes, stuck to the wall of heated spherical
oven and baked for about 2 min at about 300 C (Faridi
et al., 1982; Institute of Standards and Industrial
Research of Iran, 1998; Ter-Sarkissian et al., 1974).
Organoleptic characteristics
A panel of 145 judges comprising students, sta and

faculty of Isfahan Medical Science University recruited
by advertisements evaluated the organoleptic characteristics of prepared bread. They assessed bendability,
appearance, avour and taste, crust texture and overall
acceptability. The samples (control and 3%, 7% and
12% defatted soy fortied bread blends) were served in
dishes labelled randomly with three-digit random numbers for all panellists. Each panellist received a rating
form scored on a 19 hedonic scale (9 being considered
excellent; 5, acceptable; and 1, extremely poor), as
suggested by Austin & Ram (1971). All panellists tasted
every sample in dishes and rated on rating form scored
sheet. Breads were sliced into small pieces (15 15 cm)
and were oered in distinct dishes at the same time.
Water was provided for rinsing purposes.
Farinograph testing
The dough mixing properties of the control and dierent

wheatsoy our blends were examined with the Brabender farinograph (Brabender, Duisburg, West Germany) according to the constant our weight procedure
(AACC, 1983; ICC, 1992). Three hundred grams our
was mixed at optimum water absorption and the
farinograph curve was centred on the 500 BU line.
The dough water absorption, mixing tolerance index
(MTI) and stability proles were calculated.
Chemical analysis
Samples were analysed for moisture, ash and ether

extract following standard American Association of
Cereal Chemists (AACC, 2000) methods. Nitrogen
Extensograph testing
Dough from the farinograph measurements was cut into

two parts of 150 g each and was passed through the
Table 1 Eect of blending on chemical composition of soy our, wheat bread and dierent wheatsoy bread
Blend
Moisture (%)
Ash (%)
Protein* (%)
Fat (%)
Phytic acid (mg 100 g)
Defatted soy flour

Control (WF: 100%)
WF: 97% + SF: 3%
WF:93% + SF: 7%
WF:93% + SF: 7% + 3% sugar
25.7
26.9
23.7
23.0
6.0
0.9
1.1
1.4
1.4
48.9
9.9
11.8
14.0
14.2
1.0
1.3
1.3
1.3
1.3
352
184
175
143
0.3
0.5
0.7
0.3
0.1
0.1
0.1
0.1
0.1
1.2
0.8
1.0
1.2
1.5
0.1
0.1
0.1
0.1
0.1
6.4
2.1
3.6
1.9
Values are averages of three repetitions (standard error). Chemical composition (ash, protein, fat and phytic acid) is measured as dry basis. WF, wheat
flour; SF, defatted soy flour.
*N 5.7 for wheat bread and all bread blends.
balling and moulder unit of a Brabender extensograph

(Brabender, Duisburg, West Germany). After 45 min
resting in the fermentation cabinet, the dough was
stretched. After this rst test, the balling and moulding
operations were repeated and the dough was tested
again after a further 45 min resting time. The same
procedure was repeated for a third time, following the
ocial procedure (AACC, 1983; ICC, 1992).
Statistical methods
Data were expressed as mean SE. Signicant dierences were determined at the P < 0.05 level. The data
were statistically analysed in a completely randomised
design by analysis of variance using the standard
methods (Panse & Sukhatme, 1961). When the overall
F is signicant and more than two groups are being
compared, post-hoc (Tukeys HSD) tests to determine
which pairs of means dier from each other and SPSS
software, version 11.0 were used (SPSS Inc., Chicago,
IL, USA).
The chemical composition of wheat bread as control

and dierent defatted soy-fortied breads is shown in
Table 1. The ash content of 3% and 7% wheatsoy
bread blends vs. control increased 15% and 33%,
respectively. As soy our increased in blends, protein
content of bread blends increased. Then the protein
content of 3% and 7% wheatsoy bread blends vs.
control increased 21.4% and 29.1%, respectively. Data
on the eect of fermentation on phytic acid contents of
the control and dierent wheatsoy bread indicated that
adding sugar to the blend decrease phytic acid content.
It is probable that adding sugar to the blend may speed
up fermentation and also increase degradation of phytic
acid. This reduction in phytic acid may be useful in
improving nutritional quality of soy with respect to
mineral bioavailability.
Some investigations exist on bread enrichment with
oil seed our and protein isolates. Khan & Lawhon
(1980) reported the eects of protein isolates from soy,

cotton seeds and peanuts on bread baking. The highest
amount of protein isolate added to produce bread with
acceptable volume was 4% for soy and 8% for cotton
seed and peanut. Sahni & Krishnamurthy (1975)
reported optimum levels of ground nuts (10%) and
soy our (10%) that can be added to produce acceptable
speciality Indian bread. In a study, wheat our was
fortied with 8% DSBM (defatted soybean meal), 12%
DSBM and 8% DSBM 4% DSM (defatted sesame
meal) to increase the protein content and amino acid
prole, and to upgrade the overall nutritional value of
the bread (Serna-Saldivar et al., 1988). As soy our
increased in our blends, protein and ash content of
bread blends increased, which may result in improving
nutritional quality of bread as shown in all other studies.
Fortication with soy our is advantageous due to the
increased nutritional value (higher mineral and protein
content).
Organoleptic characteristics
One hundred and forty-ve panellist evaluated the

samples. The test was conducted in a cafeteria with no
facility of sensory evaluation room and panellists were
seated one per table separately and rinsing water was
provided for palate cleansing after each sample testing.
Bendability, appearance, avour and taste, crust texture
and overall acceptability properties were evaluated.
Data are presented in Table 2.
Appearance score with respect to the control bread
decreased signicantly upon increasing the blending
level to 7% and 12% with soy our, whereas the score of
bread containing 3% defatted soy our was the most
acceptable.
Flavour and taste score signicantly decreased with
increasing soy our substitution levels. The sample
containing 12% of defatted soy our was rated poorest
in avour and taste, whereas the score of bread
containing 3% defatted soy our was highest. The
avour of 12% soy our-blended bread might be
aected by the avour of soy bean our.
Table 2 Eect of blending on sensory evaluation of dierent bread

Bread
Control (WF: 100%)
WF: 97% + SF: 3%
WF: 93% + SF: 7%
WF: 88% + SF: 12%
F value
P value
Bendability*
6.1 0.15
6.5 0.12
6.3 0.15
6.1 0.16
1.68
<0.1702
Appearance
6.4 0.12ac
6.5 0.12a
5.8 0.15b
5.9 0.15bc
7.1031
<0.0001
Flavour and taste

5.7 0.14ef
6.2 0.12e
5.7 0.14ef
5.2 0.17f
7.4175
<0.0001
Crust texture
5.6 0.14
5.9 0.14
5.8 0.15
5.4 0.15
2.3652
<0.0701
Overall acceptability
6.0 0.13gh
6.3 0.13g
6.0 0.14gh
5.7 0.15h
3.2492
<0.0216
WF, wheat flour; SF, defatted soy flour. Values are mean standard error of all independent determinations, scored on a 9-point scale. Mean values in
the same column followed by different letters are significantly different (P < 0.05).
*The property of being easily bent without breaking.
1695
1696
The results revealed that there are no statistically

signicant dierences for crust texture and bendability
of various breads.
Overall acceptability rating was the mean score of all
the organoleptic characteristics in the present study. The
results showed that the overall acceptability score of all
the supplemented bread at the 3% level was highest.
Overall acceptability score of 12% soy our is statistically signicantly lower than 3% soy our substitution
level.
The results from Table 2 show that the overall F is
signicant in appearance, avour and taste and overall
acceptability attributes. Therefore, post-hoc tests to
determine pair-wise dierences were used. In appearance
attribute, the signicant pair-wise dierences in the 7%
defatted soy bread scored signicantly lower than the
3% defatted soy and control bread. Another signicant
pair wise dierence in the 12% defatted soy bread
scored signicantly lower than 3% defatted soy bread.
In avour and taste and overall acceptability attributes,
the only signicant pair wise dierence is the 12%
scored lower than the 3% defatted soy bread.
The blending of wheat our with soy our at dierent
levels altered the organoleptic properties of dierent
blended breads. As the 12% defatted soy fortied bread
has statistically lower score in appearance, avour and
taste, crust texture and overall acceptability attributes,
for the rest of study it was omitted, and instead a blend
contains 7% defatted soy our with 3% sugar was
applied. The present study suggests that the blending of
wheat our with defatted soy our at dierent levels
altered the organoleptic properties of dierent blended
breads. Therefore, we conclude that there is no significant dierence between any of the samples of 3% and
7% added soy and the control, while 3% added soy has
the highest score in most of the attributes. Gayle et al.
(1986) and Hoseney (1994) suggested that the appearance of bread was an important sensory characteristic
on which the acceptability of bread depends. In another
study (Dhingra & Jood, 2004), addition of 10% soy
our (full fat and defatted), 15% barley plus full fat or
defatted soy our to wheat our produce acceptable
bread. Therefore, substitution of soy (full fat and

defatted) higher than 15% did not produce organoleptically acceptable bread. Organoleptic evaluation of
their study revealed that, as the level of soy our was
increased, the crust colour of the bread changed from
creamy white to dull brown. Appearance %score
decreased signicantly upon increasing the blending
levels at 15% and 20% with soy our. Flavour of bread
increased on increasing the level of soy our up to 10%
level. Rastogi & Singh (1989) revealed that the avour of
12% soy our blended bread might be aected by the
beanie avour of soybean our. In contrast, in another
study beanie avour ratings for bread containing up to
30% soy our were not signicantly dierent from the
control whole wheat bread (Shogren et al., 2003). In
addition, Dhingra & Jood (2004) showed that the crust
texture was related to the external appearance of the
bread top. Crust texture score also decreased with
increase in substitution of soy our in wheat our as
compared with control bread. Therefore, they concluded
that soy our could be added to bread our up to level
of 10% without any signicant change in organoleptic
characteristics. While, our study showed that addition of
3% defatted soy our produced the most acceptable
organoleptic bread. For sensory evaluation, in another
study, it was found that using either up to 5% soybean
our could replace the wheat our in cookies formula
without adversely aecting baking performance or
altering the physical characteristics of the end product
(Hegazy & Faheid, 1990) which complies with our
study.
Rheological properties
Farinograph testing of the control our and dierent

wheatsoy our blends were tested by the Brabender
farinograph. Our results showed that the wheat our
fortied with 12% soy absorbed more water than the
other fortied bread and the control. Water absorption
increased as the level of fortication increaseds because
the higher protein content resulted in a greater waterbinding capacity (Table 3). Matthews et al. (1970) also
Table 3 Rheological properties of wheat and dierent wheatsoy dough

Resting time (min)
45
90
135
Flour and flour inblends (%)
Water absorption (%)
Valorimeter value (BU)
E (cm2)*
R50 Ex
E (cm2)*
R50 Ex
E (cm2)*
R50 Ex
Control (WF: 100%)

WF: 97% + SF: 3%
WF: 93% + SF: 7%
65
65.5
69
49
52
56
42
45
46
0.94
0.99
1.34
48
50
52
1.14
1.34
1.91
48
55
55
1.17
1.73
2.52
WF, wheat flour; SF, defatted soy flour.

*Energy (E) required to break the strength of dough after 45, 90 and 135 min in the rest cabinets of the extensograph.
Resistance (R) measured after 50 mm transposition of the recorded paper. Extensibility (Ex) of dough in mm.
concluded that oil seed our added to wheat our

increased the farinograph water absorption, which is
similar with our ndings. Valorimeter value increased as
the percentage levels of soy our increased in blends.
Dough stability decreased as the level of fortication
increased. Flour containing 12% defatted soy our was
the least stable. Farinograms of wheat our bread and
dierent wheatsoy our blends are shown in Fig. 1.
The results of extensograph testing were expressed as
the resistance to constant deformation after 50 mm
stretching (R50); the extensibility (Ex) was described as
the distance travelled by the recorder paper from the
moment that the hook, touches the test piece until
rupture of the test piece and the ratio between the two of
them (R50 Ex). The results of the farinogram and
extensogram studies are shown in Table 3. The amount
of water (absorption) required to centre the farinogram
curve on the 500 BU (Brabender Units) line increased
steadily with every increment of our blends from 3% to
7% soy our blend. The presence of soy our increased
the water required for the optimum bread-making
absorption. Water absorption level was positively correlated with loaf volume; these water levels could not be
considered optimal because the workability of the dough
was impaired. The valorimetric values (BU) showed an
increase as the soy substitution level increased. The
inclusion of soy our blends delayed farinograph arrival
time and decreased dough stability when substituted for
wheat our in a bread system. As the level of our
blends in composite dough increased, farinograph
absorption (Table 3) increased, but mixing time and
dough stability decreased as the substitute level increases

from 3% to 7% soy our. Extensographs showed that
Ex required breaking the strength of dough after 45, 90
and 135 min in the rest cabinets increased as the
substituted level and the resting time increased from
3% to 7% and from 45 to 135 min, respectively. The
ratio R50 Ex increased as the proportion of wheat
substitution changed and appeared to be more pronounced after 90 min of resting time.
Matthews et al. (1970) concluded that oil seed our
added to wheat our increased the farinograph water
absorption, reduced dough tolerance and produced
bread with lower loaf volume. The changes in the
physicochemical properties of bread containing soy
our were reported by Shiraldi et al. (1996) who found
a signicant reduction of bread staling caused by soy
our. In a study when levels of 5%, 10% and 15% of
soybean were used to supplement cookies, extensograph
results indicated that soy our increased resistance to
extension, proportional number and energy and diminished dough extensibility (Hegazy & Faheid, 1990).
Conclusion
The present study conrmed that the blending of wheat

our with defatted soy our at dierent levels altered the
organoleptic properties of dierent blended breads even
though it is not signicantly dierent. Therefore, we
conclude that adding 3% or 7% defatted soy our
actually gives as good a loaf of bread as the 100% wheat
bread. Fortication with soy our is advantageous due
Figure 1 Farinograph testing of the control

our and dierent wheatsoy our blends
showed that water absorption, dough
development, dough stability and valorimeter
value increased as the level of fortication
increased.
1697
1698
to the increased nutritional value (higher mineral and

protein content) and higher water absorption with
acceptable consumer attitude in rheological and sensory
characteristics.
Acknowledgment
This work is supported by a grant from the deputy

research of the Isfahan University of Medical Sciences.
The authors would like to greatly thank the sta of the
laboratories of School of Health and Institute of
Nutrition for their assistance.
References
AACC. (1983). Approved methods of the AACC, 8th edn. Method 54
21, approved April 1961, revised October 1982. St. Paul, MN, USA:
American Association of Cereal Chemists.
AACC. (2000). Approved Methods of the American Association of
Cereal Chemists. St. Paul, MN, USA: American Association of
Cereal Chemists.
Ames, B.N., Shigenaga, M.K. & Hagen, T.M. (1993). Oxidants,
antioxidants, and the degenerative diseases of aging. Proceedings of
the National Academy of Sciences of the United States of America,
90, 1522.
AOAC. (1995). Method 14.026. Ofcial Methods of Analysis. Washington, DC, USA: Association of Ocial Analytical Chemists.
Austin, A. & Ram, A. (1971). Studies on chapatti making quality of
wheat, Indian Council of Agricultural Research, New Delhi.
Technical Bulletin, 31, 96101.
Castle, E.P. & Thrasher, J.B. (2002). The role of soy phytoestrogens
in prostate cancer. The Urologic Clinics of North America, 29, 71
81.
Dhingra, S. & Jood, S. (2004). Eect of our blending on functional,
baking and organoleptic characteristics of bread. International
El-Adawy, T.A. (1997). Eect of sesame seed protein supplementation
on the nutritional, physical, sensory properties of wheat our bread.
Faridi, H.A., Finney, P.L., Rubenthaler, G.L. & Hubbard, J.D.
(1982). Functional (bread making) and compositional characteristics
of Iranian at breads. Journal of Food Science, 47, 926929.
Gayle, P.E., Knight, E.M., Adkins, J.S. & Harland, B.F. (1986).
Nutritional and organoleptic evaluation of wheat breads supplemented with pigeon pea our. Cereal Chemistry, 63, 136138.
Giovannucci, E., Ascherio, A., Rimm, E.B., Stampfer, M.J., Colditz,
G.A. & Willett, W.C. (1995). Intake of carotenoids and retinol in
relation to risk of prostate cancer. Journal of the National Cancer
Institute, 87, 6776.
Hasler, C.M. (2002). The cardiovascular eects of soy products.
Journal of Cardiovascular Nursing, 16, 5063.
Hegazy, N.A. & Faheid, S.M.N.. (1990). Rheological and sensory
characteristics of dough and cookies based on wheat, soybean, chick
pea and lupine our. Food Nahrung, 34, 835841.
Hoseney, R.C. (ed.) (1994). Principles of Cereal Science and Technology. Pp. 197213. St. Paul, MN, USA: American Association of
Cereal Chemists.
ICC (1992). International Cereal Chemistry. Standard 1972 No. 15 and

No. 114 1 (Revised 1992).
Institute of Standards and Industrial Research of Iran. (1998). Cereal
and Cereal Products Flat Bread Taftoon Code of Practice, 1st
Revision, ISIRI 5808.
Khan, M.N. & Lawhon, J.T. (1980). Baking properties of oilseed
protein and isolates produced with industrial membrane systems.
Latta, M. & Eskin, M. (1980). A colorimetric method for phytate
determination. Journal of Agricultural Food Chemistry, 28, 1313
1315.
Matthews, R.H., Sharpe, E.J. & Clark, W.M. (1970). The use of some
oilseed our in bread. Cereal Chemistry, 47, 181189.
Mohamed, A.A., Rayas-Duarte, P., Shogren, R.L. & Sessa, D.J.
(2006). Low carbohydrates bread: formulation, processing and
sensory quality. Food Chemistry, 99, 686692.
Panse, Y.G. & Sukhatme, P.V. (1961). Statistical Methods for
Agricultural Workers, 2nd edn. New Delhi: Indian Council of
Agricultural Research.
Rastogi, A. & Singh, G. (1989). Eect of addition of full fat soy our
of dierent varieties on quality characteristics and bread making
quality of white our. Bulletin of Grain Technology, 27, 2634.
Rimm, E.B., Stampfer, M.J., Ascherio, A., Giovannucci, E., Colditz,
G.A. & Willett, W.C. (1993). Vitamin E consumption and the risk of
coronary heart disease in men. New England Journal of Medicine,
328, 14501456.
Sahni, S.K. & Krishnamurthy, K. (1975). Development of high protein
bread. Part II. Soya our utilization. Journal of Food Science and
Scheiber, M.D., Liu, J.H., Subbiah, M.T.R., Rebar, R.W. & Setchell,
K.D.R. (2001). Dietary inclusion of whole soy foods results in
signicant reductions in clinical risk factors for osteoporosis and
cardiovascular disease in normal postmenopausal women. Menopause, 8, 384392.
Serna-Saldivar, S.O., Lopez-Ahumada, G., Ortega-Ramirez, R. &
Dominguez, R.A. (1988). Eect of sodium stearoyl-2-lactylate on
the rheological and baking properties of wheat bread fortied with
defatted soybean and sesame meal. Journal of Food Science, 53,
211214.
Shiraldi, A., Piazza, L., Brenna, O. & Vittadini, E. (1996). Structure
and properties of bread dough and crumb: calorimetric, rheological,
and mechanical investigation on the eects produced by hydrocolloids, pentosans, and soluble proteins. Journal of Thermal analysis,
47, 13391360.
Shogren, R.L., Mohamed, A.A. & Carriere, C.J. (2003). Sensory
analysis of whole wheat soy our breads. Journal of Food Science,
68, 21412145.
Stampfer, M.J., Hennekens, C.H., Manson, J.E., Colditz, G.A.,
Rosner, B. & Willett, W.C. (1993). Vitamin E consumption and
the risk of coronary disease in women. New England Journal of
Medicine, 328, 14441448.
Ter-Sarkissian, N., Azar, M., Ghavifekr, H., Ferguson, T. & Hedayat,
H. (1974). High phytic acid in Iranian breads. Journal of the
American Dietetic Association, 65, 651653.
Vittadini, E. & Vodovotz, Y. (2003). Changes in the physicochemical properties of wheat-and soy- containing breads during storage
as studied by thermal analyses. Journal of Food Science, 68,
20222029.
Original article
Effects of gelatine type and concentration on the shelf-life stability
and quality of marshmallows
Johanna M. Tan & Miang H. Lim*
Department of Food Science, University of Otago, Dunedin 9054, New Zealand
(Received 6 March 2008; Accepted in revised form 11 April 2008)
Summary
The eects of gelatine concentration, bloom strength, and origin on the quality and shelf-life stability of
marshmallows were studied. All six sample treatments were carried out under accelerated storage conditions
of 25 C and 75% relative humidity (RH) for 25 weeks. Gelatine A 150 bloom had the highest viscosity
because of its highest concentration (2.54%), lowest density and greatest amount of moisture loss producing
the hardest marshmallows. Hardness and water activity measurements correlate for all sample treatments
indicating that moisture loss is the main mechanism for hardening. With the exception of Gelatine B 2.2%,
sugar crystallisation may have occurred in all sample treatments at week 20 which would have an impact on
hardness as well. Gel network formation may be contributing towards hardness in Gelatine B 2.2% as there
was an increase in hardness but no changes were perceived in water activity.
Keywords
Crystallisation, gel network formation, gelatine, marshmallow, shelf life stability.
Introduction
Marshmallow, known as Pate de Guimauve, originated

in France. They were made from root extracts of
Althaea ocinalis, a plant growing in marshy areas. The
extract is a viscous juice and forms a light foam when
beaten with eggs and sugar (Lees & Jackson, 1973).
Today, the marshmallow is an aerated confectionary
consisting of mainly sugar syrup and an aerating and
stabilising agent, usually gelatine. Gelatine is the most
popular aerating agent used for marshmallow as it
produces stable foams of a light and airy texture
(Groves, 1995; Jackson, 1995). It is a very eective
foam stabiliser in preventing air bubbles in the system
from collapsing (Groves, 1995; Francis, 2000). Commercial marshmallows have a limited shelf-life of
approximately 40 weeks as changes in texture, mainly
hardening and loss of elasticity of the foam, occur over
time. For commercial marshmallows sold in warm areas
there is a noticeable increase in hardness after 20
22 weeks of storage. Moisture loss, sugar crystallisation,
foam collapse, gel network formation or a combination
of the above could contribute to the hardening process.
Air is not readily perceived in marshmallow, but its
presence contributes greatly to a soft and light texture
and adds volume to the mixture generating foam of high
e-mail: miang.lim@stonebow.otago.ac.nz
volume and low density (Groves, 1995; Jackson, 1995;

Campbell & Mougeot, 1999). All the ingredients are
combined and whipped to incorporate air. Initially large
air bubbles form and are broken into smaller ones
during continuous whipping. An optimum foam is
formed when the desired density is obtained. It is then
either extruded or deposited into starch moulds. Factors
which aect the degree of aeration include choice and
concentration of aerating agent, type of whipping
process, speed and temperature of whipping, amount
of sugar, and the percentage of dry solids (Alikonis,
1979; Lees, 1991; Jackson, 1995).
The formulation of marshmallow is believed to aect
the rate of change from an initial desirable soft and
elastic foam to a hard and gritty foam (Lees, 1991;
Groves, 1995). Research on the eect of the types of
sweeteners and air cell size distributions on shelf-life of
marshmallow have shown that glucose syrup with
high dextrose equivalent reduced sugar crystallisation
(Edmond, 2000; Lim et al., 2006) and that as total air
cell surface area increased the rate of moisture loss
increased as did the hardness of the marshmallows
(Eyres, 2001; Heenan, 2004).
Gelatine source inuences the texture and stability of
marshmallow with gelatine from older animals having a
higher tensile strength and forming rmer gels. This is
because the amount of soluble collagen decreases with
animal age and the degree of intermolecular crosslinking increases thus forming a stable network of
doi:10.1111/j.1365-2621.2008.01756.x
1699
1700
Effects of gelatine type and concentration on stability of marshmallows J. M. Tan and M. H. Lim
trivalent linkages consisting of three collagen molecules

(Ledward, 2000). Gelatine gel network formation contributes towards hardening in marshmallows which is a
textural defect (Tosh et al., 2003). Bloom value characterises the strength of a gelatine gel. A moderate
bloom value between 180 and 220 is recommended for
the production of marshmallows (Lees & Jackson,
1973). A high bloom value indicates increasing gel
strength and clarity, it also signals a decrease in gel
colour, avour, setting time and stickiness (Jackson,
1995). The purpose of this research is to study specifically the eects of dierent gelatine concentrations,
bloom values, and origin on the premix viscosity, shelflife stability, and quality of marshmallow under accelerated storage conditions of 25 C and 75% RH.
Methods and materials
Marshmallow composition
For the purpose of this study, 1.5 kg batches were made

with ingredients obtained from a commercial manufacturer. Because of condentiality, the exact composition
of the syrup solution (consisting mainly of sucrose,
glucose syrup and water) cannot be disclosed. A basic
marshmallow composition normally consists of a syrup
solution (90.6%) and a gelatine solution (9.4%). As the
main aim of this study is to determine the type,
concentration and origin of the gelatine, the composition of the syrup solution was kept constant throughout
this study. Sample treatments were varied with respect
to gelatine concentrations, bloom values, and origin of
the gelatine. Four dierent types of gelatine were used
and were obtained from Gelita NZ Ltd (Christchurch,
New Zealand) (Table 1).
The equation below was provided by Gelita to
determine the amount of gelatine needed to maintain a
similar gel strength for gelatine of dierent bloom
values.
r
Bloom 1
concentration 2 concentration 1
Bloom 2
where bloom 1 and concentration 1 represents Gelatine

A 2.2%, and bloom 2 and concentration 2 represents the
other gelatine.
Marshmallow sample preparation
A syrup solution was heated to 50 C in a microwave

and then mixed thoroughly with the gelatine solution,
which has been hydrated for about 5 min at 65 C. The
nal solution mixture, called the premix, was allowed to
stay for about 2 min over the water bath before being
transferred to a mixing bowl. The temperature of the
premix was maintained at 50 C. The premix was
whipped for 4.5 min in a Kenwood mixer with the rst
3.5 min at level 5 and then the last min at level 6. The
marshmallow whip was piped into cylindrical shaped
starch moulds of 32 mm in diameter and 15 mm in
height. A thin layer of starch was dusted onto the tops
of the marshmallows. They were left to set for about
24 h at room temperature.
Marshmallows were removed from the moulds and
excess starch brushed o. The marshmallows were packed
into moulded plastic trays and sealed in plastic barrier
bags (25 lm nylon inside a 60 lm ve layer coextruded
bag, impervious to trichloroanisol, Shorko W40, Cadbury
Confectionary Ltd, Dunedin, New Zealand) and stored
under accelerated conditions of 25 C and 75% RH.
Gelatine premix viscosity
Viscosity of the gelatine solutions at the time of whipping

were analysed by a RheoStress I rheometer (Thermohakke, Germany) along with RheoWin Pro software
(version 2001) and DC 30 Haake thermo controller
(Waltham, MA, USA) was used. A cup (Z34 Din 53018)
and spindle (Z34 Din) was used. A rotational shear rate
sweep was carried out to measure viscosity and shear
stress. As gelatine solutions display a near Newtonian
behaviour, readings were taken at twenty dierent shear
rates between 1000 and 10 s for 20 s at each shear rate.
Temperature was controlled at 50 C, to mimic the
whipping temperature. The experiment was done in
duplicate and apparent viscosities were calculated from
the ratio of the stress and shear rate.
Table 1 Variations in gelatine origin, bloom value and concentration

of the different sample treatments
Foam density
Sample
Gelatine
Gelatine
Gelatine
Gelatine
Gelatine
Gelatine
A 2.0%
A 2.2%
A 2.4%
A 150 bloom
A 250 bloom
B 2.2%
Origin
Bloom
value
Gelatine
concentration
(%)
Young bovine
Young bovine
Young bovine
Young bovine
Young bovine
Old bovine
200
200
200
150
250
200
2.0
2.2
2.4
2.54
1.97
2.2
Density measurements were carried out in triplicate for

all of the six treatments. Immediately after whipping, the
marshmallow foam was piped evenly into a pre-weighed
cylindrical container with a volume of 60.20 mL, and
reweighed. Density was calculated using the density
equation
density
mass of foamg
volume of containermL
Initial moisture content of the marshmallow samples

were measured in triplicates at week 0 using the vacuum
drying method. The samples were dried in an individual
aluminium dishes in a vacuum oven at 75 C and a
pressure of 4 Pa for 24 h. After drying, the dishes were
cooled in a desiccator at room temperature for about 1 h
before weighing. Initial moisture content was calculated
from the change in weight of the marshmallows.
Moisture loss of marshmallows was determined during storage by the change in weight of nine marshmallows on a tray sealed in an opaque barrier bag. The tray
was stored under the same accelerated conditions. This
was done in triplicate, thus giving twenty-seven marshmallows per treatment. At each testing period, these
trays were weighed and moisture loss was determined by
the change in weight.
Water activity, aw
The water activity of the marshmallows was measured in

triplicate using a water activity metre (CX3, AquaLab,
Decagon Devices Inc., Pullman, WA, USA). Small
strips were cut from the marshmallow centre. The water
activity metre was calibrated with a salt standard (NaCl
6.0 Molal in water, aw = 0.760). Water activity tests
were carried out on the stored marshmallows at weeks 0,
1, 2, 6, 14, 20, 22 and 25.
Hardness
The marshmallows were tested in replicates of ves for

each treatment. A two-bite compression test was carried
out using the Instron Universal Testing Machine (Model
5564; Instron, High Wycombe, UK). The marshmallows
were compressed 7 mm (about 28% compression), with a
30-mm at-head probe, at a constant crosshead speed of
0.4 mm s. Hardness tests were carried out at weeks 0, 1,
2, 6, 10, 14, 18, 20, 22 and 25.
Statistical analysis of variance (anova) tests were

conducted to determine if there were signicant dierences in water activity and hardness. spss version 10 was
used to carry out two-way anova univariate analysis,
where the two factors of time and sample treatments
were used, and Tukey HSD post hoc test.
Gelatine premix viscosity
Marshmallow premix, which is the syrup solution with

gelatine, should be of a low to medium viscosity to trap
air into the system and to improve the whipping

properties. Viscosity of the premix is also important in
the setting quality of the nished marshmallow product
(Tiemstra, 1964). It may have an eect on whipping
properties, thus aecting the air cell size and subsequently leading to dierent rates of moisture loss
(Heenan, 2004). The premix demonstrated a slight shear
thinning behaviour consistent of a polymer solution,
which showed a slight decrease in apparent viscosity at
high shear rates (Fig. 1). The rheological behaviour of
gelatine was consistent with results reported in previous
studies (Tiemstra, 1964; Wulansari et al., 1998; Marcotte et al., 2001; Heenan, 2004).
Gelatine A 150 bloom had the highest viscosity
(Fig. 1, Table 2) possibly because of the high level of
gelatine used, 2.54%. The viscosity of Gelatine A 250
bloom was inuenced by both the low concentration of
gelatine and the high bloom value of 250. Although the
gelatine concentration was the lowest at 1.97%, it was
not the least viscous as the high bloom value had greater
impact. A high bloom value indicates increasing gel
strength from increased cross-linking between molecules, therefore viscosity could, as a result, be greater
(Jackson, 1995; Francis, 2000). The viscosity of the
premix made from Gelatine B 2.2% gelatine was higher
300
250
Shear stress (Pa)
Moisture content
200
150
100
50
0
0
200
400
600
800
1000
1200
Shear rate (1/s)

Figure 1 Viscosity of premixes of marshmallow sample treatments,
Gelatine A 2.0% ( ), Gelatine A 2.2% (), Gelatine A 2.4% ( ),
Gelatine A 150 bloom (), Gelatine A 250 bloom (s) and Gelatine B
2.2% (d), at 50 C (n = 2).
Table 2 Marshmallow premix viscosity at 200 shear s and foam

density measurements taken during marshmallow preparation
Sample
Gelatine
Gelatine
Gelatine
Gelatine
Gelatine
Gelatine
A 2.0%
A 2.2%
A 2.4%
A 150 bloom
A 250 bloom
B 2.2%
Viscosity (Pas)
Density (g mL)1)
0.234
0.237
0.240
0.300
0.277
0.253
0.64
0.63
0.70
0.58
0.68
0.65
0.061
0.068
0.008
0.018
0.063
0.045
Density values are mean standard deviation (n = 3).
1701
than that of Gelatine A 2.2% despite the same concentration being used. This was expected as Gelatine B
2.2% was sourced from old bovine. These results are
supported by Djagny et al. (2001) and Tiemstra (1964)
who documented that the viscosity of gelatine solutions
increases with increasing solids and is aected by
gelatine type and concentration.
Foam density
No signicant dierences were detected between the

densities of each sample treatment (P > 0.05). The
average density of marshmallow foam was 0.65 g mL.
The densities of each sample treatment are shown in
Table 2. Heenan (2004) and Campbell & Mougeot
(1999) have previously observed a correlation between
low foam densities with a high premix viscosity. A lower
density indicates that there is a higher degree of air
incorporation (Groves, 1995; Campbell & Mougeot,
1999). At low densities, there are more air cells thus
increasing the total interfacial area. This enhances the
rate of moisture loss, which will have an eect on the
rate of hardening of the marshmallow (Jackson, 1995;
Heenan, 2004).
marshmallow studies (Edmond, 2000; Eyres, 2001;

Heenan, 2004; Jia, 2004).
Marshmallows made from Gelatine B 2.2% had the
greatest moisture loss with 3.74% loss by the end of the
25 weeks of storage and was signicantly (P < 0.05)
dierent from all other formulations except the Gelatine
A 150 bloom (Table 3). Marshmallows made from
Gelatine A 150 bloom gelatine lost 3.61% moisture
and were signicantly (P < 0.05) dierent from the
2.94% loss by Gelatine A 2.2%. Faster moisture loss has
been correlated with foam which observed lower density
despite the densities were not signicantly dierent.
These results support other studies which reported that
constant moisture content in marshmallow is hard to
maintain and moisture loss is inevitable in marshmallow
(Tiemstra, 1964; Alikonis, 1979).
Water activity
Water activity is a measure of the free water in the

system. Water activity values for all samples increased
from week 20 to 22 with the exception of Gelatine B
2.2% (Fig. 3). This increase could be as a result of sugar
crystallisation and the crystallisation has been documented previously by Edmond (2000) as a hardening
Moisture loss
Table 3 Final measurements taken at week 25 for moisture loss, water
activity and hardness of the different sample treatments
Sample
Gelatine
Gelatine
Gelatine
Gelatine
Gelatine
Gelatine
2.00
2.00
10
15
20
25
5.00
6.00
0.600
30
Water
activity
Hardness
(N)
5.43
5.50
5.99
6.62
6.05
6.94
0.497
0.488
0.476
0.416
0.481
0.455
9.63
21.25
9.97
35.23
12.08
25.78
a a
a,b
b,c
0.550
b
3.00
4.00
Moisture
loss (%)
0.650
0.00
1.00
A 2.0%
A 2.2%
A 2.4%
A 150 bloom
A 250 bloom
B 2.2%
0.700
1.00
aw
% Moisture gain
Moisture loss over time is an indicator of marshmallow

quality deterioration as it has an eect on glass
formation and sugar crystallisation (Lees & Jackson,
1973; Alikonis, 1979; Jackson, 1995). A continual
decrease in moisture and increasing hardness may be
an indication of sugar crystallisation as previously
observed (Edmond, 2000). All the sample treatments
had signicant (P < 0.05) moisture loss during
25 weeks under accelerated storage conditions (Fig. 2).
This result was similar to that observed in previous
% Moisture loss
1702
0.500
0.450
c
d
0.400
e
f
7.00
g
h
8.00
0.350
Weeks
Figure 2 Moisture loss of marshmallow samples, Gelatine A 2.0% ( ),

Gelatine A 2.2% (), Gelatine A 2.4% ( ), Gelatine A 150 bloom (),
Gelatine A 250 bloom (s) and Gelatine B 2.2% (d), during 25 weeks
of accelerated storage at 25 C and 75% RH. Values represent the
mean standard deviation (n = 3).A
0.300
10
15
Weeks
20
25
30
Figure 3 Water activity of marshmallow samples, Gelatine A 2.0%

( ), Gelatine A 2.2% (), Gelatine A 2.4% ( ), Gelatine A 150 bloom
(), Gelatine A 250 bloom (s), and Gelatine B 2.2% (d), during
25 weeks of accelerated storage at 25 C and 75% RH, values are
mean standard deviation (n = 3).A
mechanism of marshmallow. An increase in water

activity while moisture loss continues is an indication
of crystallisation as sugar molecules expel water when
forming crystals (Hartel, 1993). The industry has indicated that crystallisation has been observed in their
products after 20 weeks of storage.
Initial water activities for all six samples ranged from
0.590 to 0.646. All sample treatments had a signicant
(P < 0.05) drop in water activity at week 25 (Fig. 3).
Final water activities ranged from 0.416 to 0.497
(Table 3). Both time and sample treatments were
signicant (P < 0.05) factors contributing to dierences
in water activity (Fig. 3, Table 3).
Texture
Hardening is undesired and is the main cause of

deterioration in marshmallow (Alikonis, 1979; Groves,
1995). Hardening can be due to a combination of sugar
crystallisation, the problem of entering the glassy state,
or gelatine gel network formation (Hartel, 1993; Normand et al., 2000).
Initial marshmallow hardness at week 0 ranged from
2.05 N to 4.33 N (Table 3) and increased signicantly
(P < 0.05) over 25 weeks of storage (Fig. 4). Both time
and sample treatments are signicant (P < 0.05) factors
in causing dierences in hardness (Table 3). The Gelatine A 150 bloom formulation was signicantly
(P < 0.05) harder than all the other samples with a
nal reading of 35.23 N (Fig. 4) which correlates to its
fast moisture loss (Fig. 2). This has been previously
observed by Heenan (2004), Jackson (1995) and Massey
et al. (2001). The hardness could also partly be attributed to the high gelatine concentration despite its low
bloom value. A higher gelatine concentration will lead
to the formation of stronger gel networks within the
structure (Normand et al., 2000).
The Gelatine A 250 bloom formulation was signicantly (P < 0.05) softer than Gelatine A 2.2% and
40.00
Hardness (N)
35.00
30.00
25.00
20.00
15.00
10.00
5.00
aa
a,b
b, c
a,b
0.00
10
15
20
25
Weeks
Figure 4 Hardness measurements of marshmallow samples, Gelatine A
2.0% ( ), Gelatine A 2.2% (), Gelatine A 2.4% ( ), Gelatine A 150
bloom (), Gelatine A 250 bloom (s), and Gelatine B 2.2% (d),
during 25 weeks of accelerated storage at 25 C and 75%RH, values
are mean standard deviation (n = 5).A
Gelatine A 150 bloom, with hardness values of 12.08,

21.25 and 35.23 N respectively (Table 3). As it is of a
high bloom, it would be expected to produce a harder
marshmallow with stronger gel networking (Jackson,
1995). However the low gelatine concentration has a
greater eect on the texture resulting in a much softer
marshmallow. This softness was also supported by its
slower moisture loss from lower interfacial surface area
thus playing a part in inhibiting the hardening process.
Gelatine A 2.0% marshmallows had a nal hardness of
9.63 N while that of Gelatine A 2.4% was 9.97 N, both
of which were softer than Gelatine A 2.2%. Once again
this signies that lower moisture loss is one of the
reasons for maintaining softness as previously reported
(Jackson, 1995; Massey et al., 2001; Heenan, 2004).
Hardness of the Gelatine B 2.2% formulation
increased gradually over the rst 20 weeks but rapidly
from week 20 to 25, with a nal reading of 25.78 N as
compared with 21.25 N of Gelatine A 2.2%. This
increase in hardness did not appear to correlate with a
decrease in water activity, except in week 25 (Figs 3 and
4). Therefore, gel cross-linking may be contributing to
the hardness of Gelatine B 2.2% marshmallows. As
Gelatine B 2.2% is sourced from an old bovine animal,
intermolecular cross-linking increases resulting with a
rmer gel of high tensile strength (Ledward, 2000).
With the exception of Gelatine B 2.2%, a decrease in
hardness from week 20 to 22 was recorded in all the
sample treatments, which could indicate an occurrence
of crystallisation of sugar (Fig. 4). As indicated in the
previous section all the other ve samples showed an
increase in water activity from week 20 to 22 (Fig. 3)
which might have been as a result of release of water of
crystallisation. When crystals formed, the gel structure
breaks down and softens, losing its integrity (Hartel,
1993). However, this was not observed with Gelatine B
2.2% marshmallows. Hence some gelatine gel network
formation could reduce the eect of structural breakdown.
Moisture loss seems to be the predominant factor in
marshmallow hardening for all the sample treatments,
except for Gelatine B 2.2%, as hardness correlates with
the water activity measurements (Figs 3 and 4). There is
a good correlation between hardness and water activity
(Fig. 5). Increases in hardness were demonstrated by
decreases in water activity in an exponential manner.
This exponential increase in hardness at low water
activities corresponds well with the glass transition (Tg)
concept, as hardness increases rapidly nearing the Tg of
sugars gelatin matrix. In a previous study carried out by
Lim et al. (2006), over a storage period of 17 weeks at
25 C and 21% RH, sucrose crystallisation occurred
only at the rst 2 weeks of storage. The degree of
crystallisation of sucrose was determined using X-ray
diraction method. Therefore, it would have been useful
to determine if sugar crystallisation has occurred over
1703
could be the choice of gelatine type, origin and

concentration for marshmallows.
40
35
30
Hardness (N)
1704
25
Acknowledgment
20
The authors thank Cadbury Confectionery Ltd, Dunedin, New Zealand, for the support and nancing of this
study.
15
10
5
0
0.3
References
0.35
0.4
0.45
0.5
0.55
0.6
0.65
0.7
Water activity
Figure 5 Correlation between water activity and hardness for marshmallow samples, Gelatine A 2.0% ( ), Gelatine A 2.2% (), Gelatine
A 2.4% ( ), Gelatine A 150 bloom (), Gelatine A 250 bloom (s), and
Gelatine B 2.2% (d), during 25 weeks of accelerated storage at 25 C
and 75% RH, with exponential regression lines.
the storage period in this study under a dierent storage

condition with the extended time-period. It is known in
the industry that marshmallow products seem to harden
rapidly after a storage period of 20 weeks. Despite the
fact that sucrose crystals were not measured directly, the
evidence of a sudden increase in water activity could
have indirectly indicated sucrose crystallisation has
occurred with a corresponding decrease in hardness.
At any given water activity, Gelatine A 2.2% was harder
than any of the other samples (Fig. 5). This is possibly
because it has low premix viscosity and low foam
density, which mean it has a high interfacial area that
allowed faster loss of moisture, thus leading to a rapid
increase in Tg and hasten hardening (Lim et al., 2006).
Conclusions
Moisture loss from the marshmallow occurred over the

25 weeks of storage. The formulation containing Gelatine A 2.0% had the slowest loss while Gelatine B 2.2%
had the fastest loss. Increasing hardness correlated well
with decreases in water activity. With the exception of
Gelatine B 2.2% formulation, crystallisation may have
occurred at week 20 for all samples which was indicated
through an increase in water activity and a corresponding decrease in hardness. Gelatine B 2.2% showed a
gradual increase in hardness although there was not
much change to its water activity. Therefore, increasing
gel network may have a role in its hardening mechanism. Overall, the main mechanism of hardening in
marshmallows is because of moisture loss with sucrose
crystallisation possibly playing a role while for Gelatine
B 2.2% formulation, gel networking seems to be a
factor. Under these experimental conditions, treatment
with gelatine A 2.0% seemed to be the most resistant to
hardening over the 25 weeks of storage and therefore it
Alikonis, J.J. (1979). Candy Technology. Connecticut: AVI Publishing

Company, Inc. Pp. 985.
Campbell, G.M. & Mougeot, E. (1999). Creation and characterization
of aerated food products. Trends in Food Science and Technology,
10, 283296.
Djagny, K.B., Wang, Z. & Xu, S. (2001). Gelatine: a valuable protein
for food and pharmaceutical industries: review. Critical Reviews in
Edmond, J. (2000). The Effects of Ingredients on the Glass Transition
Temperature, Tg, and the Shelf Life of Marshmallow. Honours
Thesis, Dunedin: University of Otago.
Eyres, G. (2001). The Effects of Foam Air-Cell Size Distribution on the
Stability of Marshmallow. FOSC 304 Reports, Dunedin: University
of Otago.
Francis, F.J. (2000). Gelatine. In: Wiley Encyclopedia of Food Science
and Technology (edited by F.J. Francis), 2nd edn. New York: John
Wiley and Sons, Inc.
Groves, R. (1995). Technology and techniques in marshmallow
production. Candy Industry, 160, 4653.
Hartel, R.W. (1993). Controlling sugar crystallisation in food products. Food Technology, 47, 99107.
Heenan, S. (2004). Effects of Processing Parameters on Marshmallow
Air Cell Size Distribution, Shelf-Life Stability, and Quality. Honours
Thesis, Dunedin: University of Otago.
Jackson, E.B. (1995). Sugar Confectionery Manufacture. London:
Blackie Academic & Professional.
Ledward, D. (2000). Gelatine. In: Handbook of Hydrocolloids (edited
by G. Phillips & P. Williams). Pp. 6585. Cambridge: Woodhead
Publishing Limited.
Lees, R. (1991). Fundamental principles in the production and
characteristics of foam confectionery products. Confectionery Production, 57, 210211.
Lees, R. & Jackson, E.B. (1973). Sugar Confectionery and Chocolate
Manufacture. Plymouth: Clarke, Doble, & Brendon Limited.
Lim, M.H., Yin, J. & Heenan, S. (2006). The mystery of marshmallow
hardening. In: Food Preservation Technology Series Water
Properties of Food, Pharmaceutical and Biological Materials (edited
by P. Buera, J. Welti-Chanes, P.J. Lillford & H.R. Corti). Pp. 325
342. New York: CRC (Taylor &Francis).
Marcotte, M., Hoshahili, A.T. & Ramaswamy, H.S. (2001). Rheological
properties of selected hydrocolloids as a function of concentration and
temperature. Food Research International, 34, 695703.
Massey, A.H., Khare, A.S. & Niranjan, K. (2001). Air inclusion into a
model cake batter using a pressure whisk: development of gas hold-up
and bubble size distribution. Journal of Food Science, 66, 11521157.
Normand, V., Muller, S., Ravey, J. & Parker, A. (2000). Gelation
kinetics of gelatin: a master curve and network modelling. Macromolecules, 33, 10631071.
Tiemstra, P.J. (1964) Marshmallows: I. overrun. Food Technology, 18,
915920.
Tosh, S., Marangoni, A., Hallet, F. & Britt, I. (2003). Aging dynamics
in gelatine gel microstructure. Food Hydrocolloids, 17, 503513.
Wulansari, R., Mitchell, J., Blanshard, J. & Paterson, J. (1998). Why are
gelatine solutions newtonian? Food Hydrocolloids, 12, 245249.
Original article
Differences in chemical, microbial and sensory quality parameters
of the marinated ascidia Microcosmus sabatieri Roule, 1885 during
storage at 6 C under vacuum conditions
Nikolaos Stamatis,1* John Arkoudelos2 & Dimitris Vafidis3
1 National Agricultural Research Foundation, Fisheries Research Institute, 64007 N. Peramos, Kavala, Greece
2 National Agricultural Research Foundation, Institute of Technology of Agricultural Products, S. Venizelou 1, 14123 Lycovrisi, Attica, Greece
3 Department of Ichthyology and Aquatic Environment, School of Agricultural Sciences, University of Thessaly, 38446 Nea Ionia, Magnesia,
Greece
(Received 19 December 2007; Accepted in revised form 11 April 2008)
Summary
Chemical, microbial and sensory quality aspects of the marinated ascidia Microcosmus sabatieri Roule, 1885,
were examined over a 150-day period at 6 C, under vacuum and air (control) packaging conditions using
three dierent formulations (with 12% sodium chloride and 3%, 5% or 7% acetic acid). Signicant
dierences were found between chemical and sensory analysis of three dierent marinated groups (P > 0.05)
during the storage period. There were also signicant dierences in pseudomonads, lactic acid bacteria and
yeast and moulds of the marinated groups by which lower bacterial counts were determined. N-3
polyunsaturated fatty acids concentrations decreased signicantly (P < 0.05), while total viable counts,
ammonia and total saturated fatty acid concentrations increased signicantly (P > 0.05) in all three groups
during storage. The dierences in fatty acid and ammonia concentrations were found to be useful as indexes
of freshness and decomposition of marinated M. sabatieri in storage. Shelf life of M. sabatieri marinades was
found to be 5 and 4 months under vacuum and air (control) packaging conditions respectively, at 6 C.
Keywords
Fatty acids, Microcosmus sabatieri, marination, product quality, shelf life.
Introduction
Marinades are preserved sh and shellsh in a mixture

of organic acid and salt (Fuselli et al., 1994). It
involves increasing the ionic strength and decreasing
the pH. The aim is not only to prevent microorganism
growth but also to allow a way of valorisation other
than salting for dierent shery products. The resulting product has an extended shelf life and characteristic avour (Y. Feng & Y. W. Huang, unpublished
results). Shrimps have extra shelf life after marinating
(Dalgaard & Jorgensen, 1999). Pelagic sh such as
herrings and sardines, which have a high esh fat
content normally, provide the raw material for the
preparation of marinated products (McLay, 1972).
Although marination technology in some seafood
products, e.g. herring, sardine, anchovy, shrimp, sprat
and galoids, is well developed (Poligne & Collignan,
2000; Gokoglu et al., 2004; Cadun et al., 2005; Kilinc
*Correspondent: Fax: +30-2594022222;

e-mail: nikstam@inale.gr
doi:10.1111/j.1365-2621.2008.01761.x
& Cakli, 2005; Nielsen et al., 2005), there is no

information about marinated ascidia.
Ascidia are commonly known as sea squirts. The body
of an adult sea squirt is quite simple, being essentially a
sack with two siphons through which water enters and
exits. Water is ltered inside the sack-shaped body where
food and oxygen are extracted. These very interesting
small marine animals are easily found on almost every
marine solid surface: rocks, jetties and algae from the
intertidal zone to a depth of 6000 m. The investigations
on their chemical composition concentrated mainly on
secondary metabolites, because they have proved themselves as a rich source of biologically active compounds
(Faulkner, 1995), on antitumour agents (Rinehart, 2000)
and on their lipid and sterol composition (Slantchev
et al., 2002). Microcosmus sabatieri (Fig. 1), called fouska in Greece, is a solitary ascidia (Stolidobranchia,
Pyuridae), widely distributed throughout the Mediterranean Sea (Monniot, 1962, 1965; Tursi, 1980) with the
exception of the Levantine basin (Koukouras et al.,
1995). It is an unselective suspension feeder species with
large ltration abilities (Fiala-Medioni, 1974). Microcosmus sabatieri lives rmly attached through its basal part
1705
1706
Chemical, microbial and sensory quality parameters of the marinated ascidia N. Stamatis et al.
Figure 1 Microcosmus sabatieri growing on rock at 27-m depth

(Dodecanese, Southern Aegean Sea); the tunic (left) and the mantle
(right).
of the tunic on various hard substrates, as well as on

biogenic detritic substratum, from the upper sublittoral
to 200-m depths (Tursi, 1980; Monniot & Monniot,
1987). The shery of M. sabatieri in the Dodecanese area
Greece goes back to the middle of 1800, where spongeshing was developed at the islands of Kalymnos and
Symi. Today there are about thirty-ve vessels belonging
to the small-scale artisanal shery, harvesting and processing M. sabatieri, all harboured at Kalymnos Island.
There is no ocial data on M. sabatieri landings but a
mean annual production of 576768 tons year)1 is
estimated based on data derived from the SpongeFishermen Union of Kalymnos. All commercial catches
are usually marketed in the Dodecanese as fresh chilled in
ice or as processed and are distributed on local market,
especially to the islands of Kalymnos, and Symi, and no
exportations of fresh or processed product are carried
out, besides the fact that M. sabatieri is consumed at
several Mediterranean countries, mainly Spain, France
and Italy. The processed, traditional product called in the
Dodecanese spinialo has a light salted character and a
limited shelf life (c. 30 days), when it is stored at cooler
temperatures (46 C). Spinialo demonstrates an exceptional nutritional value in the human diet being rich in
minerals (especially iodine, 0.01% dry weight; seawater
contains 0.05 ppm iodine or c. 34 million tons; seaweeds
of the Laminaria family are able to extract and accumulate up to 0.45% dry weight), vitamins and long-chain n-3
polyunsaturated fatty acids (PUFA) (Stamatis et al.,
2004, 2007; Lyday, 2007). N-3 PUFA possess hypolipidaemic, anti-hypertensive, anti-inammatory, antithrombotic and anti-arrhythmic properties. All of these
have been implicated in their health-promoting eects
(Kamal-Eldin & Yanishlieva, 2002). Given that the shelf

life of spinialo is relatively short and that there is a
growing tendency of consumers for consumption of
processed rather than fresh products of fouska, research
on the application of new processing and preservation
methods focussing on development of new, high added
value and high nutritional value products with shelf life
extension is required. Furthermore, it is well known that
refrigeration and vacuum packaging (VP) alone and or
in combination have considerable potential in extending
the shelf life and the keeping quality of raw as well as of
processed shery products, oering multiple advantages
to the sh industry and consumer (Arkoudelos et al.,
2007; Stamatis & Arkoudelos, 2007a, 2007b).
The aim of this work was to study the M. sabatieri
marinade production by producing dierent marinades
and determining their shelf life under VP conditions at
6 C, evaluating chemical, microbiological and sensory
quality parameters.
Samples
Fresh specimens of M. sabatieri obtained by diving at

35-m depth, in the Kavala-bay, Greece, and transported
to the laboratory within 1 h after harvesting, in polystyrene foam boxes containing ice. In the laboratory, the
specimens were shucked by cutting the two siphons with
a sterile knife and individually gut washed. Samples of
fresh esh amounting to 1.0 kg (about 150 individuals)
were used for each experiment.
Marinating process and storage
Three dierent marinating solutions (MS1, MS2 and

MS3) were prepared with 3%, 5% and 7% acetic acid
(Merck, 818755 glacial 99100 for synthesis) (Merck,
Darmstadt, Germany) and 12% sodium chloride (Merck,
106406 99.99 Suprapur) respectively. Samples were
dipped in these solutions at a ratio 1:1.5 (w w) for 48 h at
6 0.5 C. In addition, some spices (20-g black peppercorns, 5 garlic seed) and 0.05& food colouring sunset
yellow E110 Cl-15985 were added. Then the samples were
drained at ambient temperature (20 0.5 C) for 10 min
and placed in Suprovac polyamide laminate bags (thickness 90 lm, gas permeability cm3 m)2 day)1 bar)1 at
24 C, 50% RH, c. 25, 90 and 6 for CO2, O2 and N2
respectively). The aerobic (control) package was obtained
by ashing vacuum packages with atmospheric air, and
the vacuum package was obtained by packaging in an
evacuated pack. Bags supplied from the plastic lms
factory Flexopack (Koropi by Athens, Greece) were heatsealed using a VP machine (Henkovac 1502, BAs
Hertogenbosch, the Netherlands). Samples were stored
at 6 0.5 C and analysed every 30 days, up to
150 days, in order to determine the shelf life. The average

weight of the esh was 60 g, and the air volume in the bag
was c. 300 cm3.
Analytical methods
Two replicated experiments were conducted, and three

pouches were analysed on each sampling occasion.
Results are recorded as the arithmetic means of the six
values.
Chemical and proximate analysis
The pH value was determined with a pH meter (WTW

526, Germany), the glass electrode being applied directly
to the esh (Drosinos & Nychas, 1997). Crude protein
(N 6.25), total lipid, moisture and ash content were
determined by the methods described in the AOAC
manual (AOAC, 1975). Total salt concentrations were
determined according to the methods described by Karl
(1992). For ammonia and lactic acid analysis, 10 g
ascidia esh was treated as described by Drosinos et al.
(1997). Lactic acid was assayed enzymatically by the
method described by Gawehn (1984), and ammonia was
determined colorimetrically by the method of Chaney &
Marbach (1962). For fatty acid (FA) analysis, lipids were
extracted from the homogenised fresh edible portion by
the Bligh & Dyer (1959) method and were further
prepared for FA analysis according to the procedure of
Kates (1972). Methyl esters were prepared by saponication with 0.5 N NaOH and methylation with 14%
boron triuoride-methanol (Metcalfe et al., 1966). Fatty
acid methyl esters were analysed by a Hewlett Packard
(5890-Series II) (Hewlett Packard, Wilmington, DE,
USA) gas chromatograph, separating at 177 C, 18 min
hold time and 2.3 C min)1 to 210 C and equipped with
a ame ionisation detector. The capillary column SGEBPX 70 (50 m 0.22 mm 0.25 lm) was used with
helium as carrier gas and nitrogen (auxiliary gas) ow
rate at 36 mL min)1, hydrogen at 30 mL min)1 and
compressed air at 330 mL min)1. For peak identication, solutions of reference substances were analysed
under the same conditions, and their retention times
(RTs) and chromatograms were compared to those of
samples. The contribution of each identied compound
was expressed as the percentage (%) of its peak area to
the total area of all peaks eluted in each chromatogram.
The precision of the results was always better than 5%.
For statistics on chromatograms, the HP GC-Chem
Station, Rev. A. 06.03 [509], 19901998 software
(Hewlett Packard, Wilmington, DE, USA) was used.
of sterile 1 4 Ringers solution with NaCl (0.85%, w v)

and homogenised for 1 min in a Seward 400 Stomacher
(Seward Medical UAC House, London, UK). Samples
(0.1 mL) of decimal serial dilutions homogenate were
spread on the surface of the appropriate dried media in
Petri dishes for enumeration of: total viable count
(TVC) on plate count agar (Oxoid code CM 325)
incubated at 20 C for 4 days; pseudomonads (Ps) on
cetrimide fusidin cephaloridine agar (Oxoid code CM
559, supplemented with SR 103) incubated at 20 C for
2 days and yeasts-moulds (YM) on Rose-Bengal Chloramphenicol agar (RBC, Oxoid code CM 549, supplemented with SR 78) incubated at 25 C for 72 h; For
lactic acid bacteria (LAB) samples, 1.0 mL of decimal
serial dilutions were inoculated into 10 mL of molten
(45 C) MRS medium (Lactobacillus de Man Rogosa
Sharpe agar, Oxoid code CM 361) respectively. After
setting, a 10-mL overlay of molten media was added and
plates were incubated at: 25 C for 5 days for MRS
plates. Three replicates of at least three appropriate
dilutions were enumerated. All plates were examined
visually for typical colony types and morphological
characteristics associated with each growth medium.
Microbiological data were transformed into logarithms
of the number of colony-forming units (CFU g)1).
Sensory analysis
Five experienced panelists assessed the sensory properties of fouska samples during cold storage by the
method for sensory evaluation described by Gokoglu
(2002). Odour, colour, texture and general appearance
were used as criteria for acceptability. Intensities of
responses were evaluated on the verbal scale using a
seven-point category scale labelled slight (one to three
demerit points; intrinsic fresh odour, bright colour, rm
texture, glossy general appearance), moderate (four to
six demerit points; putrid odour, opaque colour, tight
texture, smooth surface general appearance) and extreme (seven demerit points; oensive odour, grey
discoloured colour, soft texture, dull general appearance) (Cardello et al., 1982). The sensory evaluation was
performed, at a month interval.
Chemical, microbiological and sensory data were

treated by analysis of variance. Signicance was established at P < 0.05 (Sokal & Rohlf, 1987).
Chemical changes
Fresh specimens of M. sabatieri (25 g) were aseptically

placed into a sterile stomacher bag containing 225 mL
The proximate composition, pH, ammonia and lactic

acid values of the fresh and marinated ascidia samples in
1707
1708
dierent marinating solutions and packs for day 2 and

day 150 (data from each sampling occasion are not
shown) at 6 C are summarised in Table 1. Moisture
value in the fresh product was determined as
77.5 1.8 g per 100 g and decreased to a mean value
of 70.2 1.3 g per 100 g after marination. Ash content
of the fresh ascidia was 2.3 0.3 g per 100 g. After
marination ash content increased to a mean value of
5.9 0.5 g per 100 g following the same pattern as for
NaCl (from a mean value of 1.5 0.3 g per 100 g in the
fresh to a mean value of 5.7 0.5 g per 100 g in the
marinated product). On the other hand, no signicant
dierences were found in moisture, ash and NaCl
contents of the marinated products (P > 0.05) during
storage under air (control) or vacuum conditions. Fat
content of the fresh product (3.5 0.7 g per 100 g) was
slightly increased during marination and reached at the
end of the storage period a mean value of 4.5 0.7 g
per 100 g as a result of the water loss caused by the
penetration of salt into the esh. The concentration of
protein was also increased after marination (c.
3.0 0.5 g per 100 g, day 2), because of the moisture
loss and shows no signicant dierences during the
storage period (P > 0.05), after marination, following
the same pattern as for the fat. Similar results for
proximate composition dierences in marinated seafood
products have been reported also by other researchers
(Gokoglu et al., 2004; Cadun et al., 2005; Kilinc &
Cakli, 2005; Ozden, 2005). The pH value of the fresh
ascidia product was 5.10 0.02. Mostly, the pH value
of live seafood products is close to 7.00 and can vary
post mortem between 6.00 and 7.00 (Simeonidou et al.,
1998). In marinated products, pH value should not be
more than 4.80 (Rehbein & Oehlenschlager, 1996).
Organic acids have an eective antimicrobial function
(Y. Feng & Y. W. Huang, unpublished results) due to
the ability to depress pH below the growth range of

microorganisms, resulting in metabolitic inhibition by
the undissociated acid molecules. The pH values of the
marinated ascidia in MS1, MS2 and MS3 were 3.61,
3.45 and 3.29 respectively after marination and slightly
increased during the storage period. The 1.52 units
decrease (P < 0.05) in the initial pH of ascidia by about
after marination in variant of acetic acid and salt
solutions is consistent with those reported for anchovies
(Engraulis encrasicolus), sardine (Sardina pilchardus)
and Pacic saury (Cololabis saira) (Gokoglu et al.,
2004; Kilinc & Cakli, 2005; Sallam et al., 2007). When
comparing pH values of the product in VP and air
(control) conditions at the beginning and at the end of
the storage period, no signicant dierences were
determined (P < 0.05).
Ammonia, the simplest possible amine, is ubiquitous
and one of the most abundant amines in fresh seafood
products (Lin et al., 1983). As quality parameter,
ammonia is used for determination of spoilage level of
shery products during dierent storage conditions and
packages (Stamatis and Arkoudelos, 2007a, 2007b). The
initial concentration of ammonia in fresh ascidia was
measured as 8.5 1.2 mg per 100 g. Ammonia content
in fresh shery products ranges from 4 to 11 mg per
100 g (Wang & Brawn, 1983). After marinating, the
concentrations of ammonia increased greatly to 12.3
and 12.2 mg per 100 g in MS1 samples stored in air
(control) and in VP conditions respectively. Masniyom
et al. (2002) reported comparable increases (P > 0.05)
for fresh seafood storage under similar packaging
conditions. Ammonia concentrations were increased
slightly during the storage time of marinated ascidia
up to 22.3 0.9 mg per 100 g and 19.4 0.9 (day
150) in MS1 samples stored in air (control) and in VP
conditions respectively (Fig. 2). Ammonia production
Table 1 Chemical and pH changes of fresh Microcosmus sabatieri samples during marination and cold storage in vacuum and air conditions at
6 C
Fresh
MS1 (2 days, VP)
MS1 (150 days, VP)
MS1 (2 days, AIR)
MS1 (150 days, AIR)
MS2 (2 days, VP)
MS2 (150 days, VP)
MS2 (2 days, AIR)
MS2 (150 days, AIR)
MS3 (2 days, VP)
MS3 (150 days, VP)
MS3 (2 days, AIR)
MS3 (150 days, AIR)
Moisture
Ash
77.5
70.9
68.3
70.9
67.9
69.5
67.2
69.5
68.2
70.1
67.1
70.2
67.5
2.3
5.3
4.9
5.1
5.7
6.2
5.4
6.2
5.2
6.4
5.7
6.4
6.2
1.8
1.4
1.4
1.5
1.4
1.4
1.6
1.7
1.5
1.5
1.6
1.7
1.5
NaCl
0.3
0.4
0.3
0.3
0.4
0.5
0.4
0.5
0.4
0.5
0.3
0.5
0.4
1.5
5.1
4.5
4.7
5.3
5.9
4.9
5.9
4.7
6.4
5.0
6.4
5.7
Fat
0.3
0.4
0.4
0.4
0.3
0.4
0.4
0.5
0.4
0.5
0.4
0.5
0.4
3.5
3.0
4.7
3.0
4.4
3.9
5.3
3.8
4.1
3.3
4.6
3.3
4.1
0.7
0.5
0.6
0.4
0.5
0.4
0.5
0.4
0.4
0.3
0.5
0.4
0.4
Protein
pH
15.8
19.5
22.1
19.5
20.2
19.2
21.5
19.2
21.4
17.6
21.7
17.6
19.9
5.10
3.61
3.83
3.61
3.89
3.45
3.61
3.45
3.59
3.29
3.46
3.29
3.43
1.4
1.5
1.7
1.4
1.6
1.5
1.4
1.5
1.7
1.6
1.6
1.4
1.7
NH3
0.02
0.01
0.02
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.02
8.5
12.2
19.4
12.3
22.3
11.8
16.9
11.8
19.9
13.5
15.5
13.5
15.8
Lactic
1.2
0.9
0.8
0.9
0.9
0.9
1.0
0.8
1.2
1.1
1.3
0.9
1.2
15.2
14.6
15.9
14.6
20.9
17.7
15.7
17.7
24.7
9.3
14.2
9.3
19.0
2.2
1.9
1.8
2.0
2.3
1.9
1.8
1.8
2.1
1.5
1.4
1.4
1.9
a
% SD for moisture, ash, NaCl, fat, protein and mg per 100 g for NH3, lactic; mean values from two different experiments. Each number is the mean of
three samples taken from the different experiments (coefficient of variation <5%). VP, vacuum packaging.
(a) 25
Vacuum storage
MS1
(b) 25
MS2
MS2
MS3
NH3 (mg 100 g1)
NH3 (mg 100 g1)
Air storage
MS1
20
15
MS3
20
15
10
10
0
2
3
Storage time (months)
2
3
Figure 2 Ammonia changes of marinated Microcosmus sabatieri samples during cold storage in vacuum (a) and air (b) conditions at 6 C.
by bacterial deamination of aminoacids has been

assumed to be the possible reason for the increase of
pH during the storage time of marinates (ICMSF, 1986).
The signicant increase of ammonia concentrations
correlates gut with the increase of TVCs during storage
of ascidia marinates. For that reason, ammonia may
provide an adequate quality index of freshness and
decomposition of marinated M. sabatieri in air (control)
and VP storage. Furthermore, ammonia concentrations
increased only slightly after marinating and storage in
MS2 and MS3 samples in air and in VP conditions
(P > 0.05; Fig. 2). This could be explained in as much
as the lower pH is not in favour of ammonia formation.
Finally, lactic acid values showed no signicant dierences (P < 0.05) between fresh and marinated ascidia
product, and therefore lactate seems not to be a sensible
chemical parameter for marinated ascidia quality
determinations.
Fatty acid changes
The FA contents in fresh and marinated ascidia product

stored in vacuum and air (control) conditions are shown
in Table 2. The FA composition for the fresh product
was typical for higher marine plants and invertebrates.
The main FAs were palmitic (16.42%) and palmitoleic
(9.94%), while the concentration of the long-chain n-3
PUFAs eicosapentaenoic acid (EPA, 17.61%) and
docosahexaenoic acid (DHA, 9.99%) typical for marine
organisms was relatively high. It is very well known that
these long-chain PUFAs are biologically important and
have been associated with a decreased risk of cardiovascular disease (Kromhout et al., 1985). Lower levels
(below 12%) of n-3 PUFA than those in the fresh
M. sabatieri have been reported for other two fresh
tunicates, namely Styela sp. and Phallusia sp. from the
eastern Mediterranean region (Slantchev et al., 2002).
Total contents of polyenes (TPUFA) in the fresh
M. sabatieri were 39.96%, somewhat higher than in the
marinated products (37.50%). Ozden (2005) reported

higher contents in anchovy and rainbow trout marinates
as in the corresponding fresh products. There was
signicant dierence in the content of total saturates
(TSFA) between fresh (31.75%) and marinated
(32.61%) ascidia. Total monoenes (TMUFA) in the
marinated ascidia were present at 24.63% while those in
the fresh product were present at 22.26%. Both fresh
and marinated ascidia contained high amounts of
palmitic, palmitoleic, EPA and DHA. The long-chain
PUFAs, EPA and DHA have been reported to be the
main PUFAs of Aegean Seas chub mackerel Scomber
colias japonicus (Stamatis & Arkoudelos, 2007b), sardine
Sardina pilchardus (Stamatis & Arkoudelos, 2007a) and
much other seafood (Zlatanos & Sagredos, 1983). The
total lipids and lipid composition in seafood may vary
with season, species, sex and or food intake (Jhaveri &
Constantinides, 1981). Rosa & Nunes (2004) stated that
16:0 (palmitic), 18:1n-9 (oleic), 20:5n-3 (EPA) and
22:6n-3 (DHA) are the main FAs in aquatic organisms
that are responsible for the peculiar taste and odour of
marinated products. Furthermore, in the frame of this
work, dierences in FA concentrations in relation to
freshness were examined in the marinated ascidia
samples during cold storage, at 6 C. Total PUFA
concentrations decreased signicantly (P < 0.05) in
air-packaging conditions but not (P > 0.05) in VP
conditions, while total SFA concentrations increased
signicantly (P < 0.05) in both packages, c. 4% and
2% in the air and vacuum package respectively, during
storage. TMUFA concentrations decreased signicantly
(P < 0.05) in VP conditions. Signicant dierences of
the MUFA (P > 0.05) have not been examined in the
air package, during storage. The results of this work for
FA values of raw material, marination stage and during
cold storage stay in agreement with those found for
other seafood species treated with similar processes
(Ozden, 2005; Voldrich et al., 1991; Odalapa et al.,
1984). In the present study, FA increases and decreases
1709
1710
Table 2 Fatty acid proles of fresh ascidia Microcosmus sabatieri samples during marination and cold storage in vacuum and air conditions at
6 C
Vacuum packaging (months)
Peak number (acid: no
double bonds)
1
2
3
4
5
6
7
8
9
(12:0)
(13:0)
(14:0)
(15:0)
(16:0)
(17:0)
(18:0)
(20:0)
(24:0)
TSFA
10 (14:1)
11 (15:1)
12 (16:1n-7)
13 (17:1)
14 (18:1n-9)
15 (18:1n-7)
16 (20:1n-9)
TMUFA
17 (18:2n-6)
18 (18:3n-6)
19 (18:3n-3)
20 (18:4n-3)
21 (20:3n-6)
22 (20:4n-6)
23 (20:4n-3)
24 (20:5n-3)
25 (22:5n-6)
26 (22:5n-3)
27 (22:6n-3)
TPUFA
Total unknown
Air packaging (months)
Fresh
product
Marinated
product
0.04
0.06
8.72
1.41
16.42
0.72
3.37
0.35
0.65
31.75
0.22
0.52
9.94
0.29
4.52
6.70
0.07
22.26
3.27
0.29
2.48
2.56
0.18
2.64
0.39
17.61
0.10
0.45
9.99
39.96
6.03
0.05
0.07
9.32
1.38
16.12
1.03
3.92
0.40
0.31
32.61
0.3
0.57
11.71
0.08
5.41
6.4
0.16
24.63
3.15
0.26
2.13
3.02
0.13
2.06
0.27
17.71
0.04
0.27
8.46
37.50
5.26
0.04
0.07
10.25
1.12
16.47
0.93
3.81
0.31
0.38
33.38
0.26
0.48
10.76
0.02
5.24
6.12
0.19
23.07
3.35
0.24
2.52
3.69
0.25
1.53
0.00
16.8
0.06
0.38
7.9
36.72
6.83
0.04
0.07
9.98
1.11
14.91
0.97
3.19
0.30
0.50
31.07
0.25
0.48
10.97
0.07
4.83
5.96
0.07
22.63
3.31
0.3
2.53
3.64
0.25
1.96
0.41
17.26
0.07
0.38
8.36
38.47
7.83
0.03
0.06
10.12
1.02
16.03
0.88
4.65
0.51
0.55
33.85
0.21
0.41
10.22
0.06
5.11
5.78
0.05
21.84
3.14
0.29
2.22
3.11
0.26
2.13
0.42
17.04
0.07
0.39
7.99
37.06
7.25
0.03
0.06
8.44
0.97
15.77
0.60
5.44
0.62
0.61
32.54
0.22
0.52
8.76
0.03
5.43
5.87
0
20.83
3.33
0.26
1.74
2.67
0.44
2.51
0.38
17.5
0.06
0.33
7.48
36.70
9.93
0.04
0.07
9.54
1.14
16.81
0.62
5.46
0.58
0.55
34.81
0.25
0.56
9.14
0.02
5.61
6.24
0
21.82
3.38
0.19
1.8
2.72
0.49
2.73
0.39
17.23
0.08
0.32
6.87
36.20
7.17
0.04
0.07
11.16
1.14
15.22
0.97
2.66
0.00
0.11
31.37
0.29
0.44
11.83
0.02
5.94
6.11
0.15
24.78
3.48
0.32
2.87
3.27
0.05
1.71
0.42
17.31
0.00
0.11
7.98
37.52
6.33
0.04
0.06
10.24
1.05
16.02
0.90
3.42
0.30
0.40
32.43
0.24
0.45
10.86
0.07
5.01
5.86
0.19
22.68
3.36
0.3
2.53
3.97
0.23
1.52
0.41
17.46
0.02
0.38
7.56
37.74
7.15
0.03
0.05
10.65
1.12
17.56
0.92
4.11
0.31
0.45
35.20
0.22
0.39
10.48
0.06
5.12
5.99
0.18
22.44
3.31
0.25
2.21
3.45
0.31
1.85
0.48
17.12
0.03
0.35
7.21
36.57
5.79
0.04
0.07
10.40
1.19
18.22
0.70
5.44
0.52
0.47
37.05
0.25
0.56
10.13
0.02
5.68
6.4
0.22
23.26
3.37
0.29
1.8
2.61
0.43
2.23
0.38
15.89
0.04
0.27
6.41
33.72
5.97
0.03
0.07
10.65
1.06
18.95
0.62
4.60
0.43
0.41
36.82
0.27
0.52
10.22
0
6.17
7.07
0.26
24.51
3.65
0.27
1.93
2.55
0.35
2.45
0.39
16.78
0.02
0.12
6.55
35.06
3.61
Percentage of the total fatty acid present. Mean values from two different experiments; each number is the mean of three samples taken from the
different experiments (coefficient of variation <5%). TSFA, total saturates; TMUFA, total monoenes; TPUFA, total polyenes.
were found to be useful as an index of freshness

and decomposition of marinated ascidia in cold storage under vacuum and air (control) packaging
conditions.
Microbiological changes
The changes in microora of the fresh ascidia product

and marinated product stored in vacuum and air
(control) packaging conditions for days 2 and 150 (data
from each sampling occasion are not shown) at 6 C are
given in Table 3. The initial microbiological quality of
the fresh product was as follows: 2.0, 1.2, <1 and 1.5
log CFU g)1 for TVCs, Ps counts, LAB counts and YM
counts respectively. These low initial counts indicate
very good ascidia quality. The TV counts did not exceed
the maximum limit of log CFU g)1, considered as the
upper acceptability limit for marine species (ICMSF,
1986). Pseudomonads and YM load decreased after

marination (day 2) and were at very low levels (not
determined) during the storage period. LAB counts were
detected during storage almost always <1, except three
times: 1.8 (after 150 days in VP, MS1), 2.0 (after
150 days in AIR, MS1) and 2.0 (after 150 days in VP,
MS2). Log TV counts of MS1, MS2 and MS3 at the day
150 were determined as: 4.5 and 6.4, 4.2 and 6.2, 3.9 and
5.1 in VP and air-packaging conditions respectively. At
the end of the storage period, the samples showed the
following microbiological conditions: MS1, log
CFU g)1 4.5 and 1.8 of TV and LAB counts in VP
conditions and log CFU g)1 6.4 and 2.0 of TV and LAB
counts in air-packaging conditions; MS2, log CFU g)1
4.2 and 2.0 of TV and LAB counts in VP conditions and
log CFU g)1 6.2 and not determined of TV and LAB
counts in air packaging conditions; MS3, log CFU g)1
3.9 and not determined of TV and LAB counts in VP
and bright, moist general appearance. On the other hand,

stale ascidia product had a detectable odour, opaque, grey
colour, very soft texture and wrinkled, viscous general
appearance. Sensory scores signicantly (P < 0.05)
charged with storage time (Table 4). Moderate intensity
was observed between 2 and 3 months after storage in air
conditions and between 3 and 4 months after storage in
vacuum conditions for all sensory characteristics.
Although the intensities of intrinsic fresh odour, bright
colour, rm texture, glossy general appearance (acceptable characteristics) signicantly (P < 0.05) decreased,
the remaining characteristics (unacceptable nature)
signicantly (P < 0.05) increased during cold storage.
According to the results of the analysis of variance test;
the dierences between VP and air storage samples were
found to be statistically signicant (P < 0.05) only for
odour in all three formulations. However, there was no
signicant dierence (P > 0.05) in colour, texture or
overall appearance between the dierent storage conditions apart from marinating solution. According to the
results of sensory analysis, the VP samples held at 6 C
were acceptable up to 5 months and the air-packaged
samples were acceptable up to 4 months.
Table 3 Microbial changes of fresh ascidia Microcosmus sabatieri
samples during marination and cold storage in vacuum and air

conditions at 6 C
Fresh
MS1 (2 days, VP)
MS1 (150 days, VP)
MS1 (2 days, air)
MS1 (150 days, air)
MS2 (2 days, VP)
MS2 (150 days, VP)
MS2 (2 days, air)
MS2 (150 days, air)
MS3 (2 days, VP)
MS3 (150 days, VP)
MS3 (2 days, air)
MS3 (150 days, air)
TV count
Ps count
LAB count
YM count
2.0
2.3
4.5
2.3
6.4
2.3
4.2
2.3
6.2
2.3
3.9
2.3
5.1
1.2
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
1.8
<1
2.0
<1
2.0
<1
<1
<1
<1
<1
<1
1.5
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
a
Values expressed in log CFU g)1 (each number is the mean of three
samples taken from the different experiments, coefficient of variation
<5%). TV, total viable count; Ps, pseudomonads; LAB, lactic acid bacteria;
YM, yeasts and moulds; VP, vacuum packaging.
conditions and log CFU g)1 5.1 and not determined of

TV and LAB counts in air packaging conditions. These
ndings are in accordance with those obtained for this
kind of product made with other seafood species (Fuselli
et al., 1994; Cadun et al., 2005; Kilinc & Cakli, 2005).
Both bacterial load and the ammonia concentration of
MS3 samples were lower (P < 0.05) than that of MS2
and MS1 samples. These ndings have been conrmed
by the panelists during the sensory evaluation.
Conclusion
In this study, the ascidia M. sabatieri was marinated

according to three dierent formulations. Fresh specimens from the Aegean Sea were used as raw material.
Shelf life of three groups of marinated product was
estimated as 150 days at 6 0.5 C under vacuum
storage conditions. Ammonia and FA values together
with sensory evaluation help to indicate the shelf life of
the marinated products. The ammonia, a by-product
of decomposition, can be utilised as an indicator of
decreasing quality during storage. Increasing SFA and
decreasing PUFA concentrations indicate that oxidation
Sensory analysis
Fresh product of ascidia had a typical seaweed odour,

glossy orange colour, slightly rm, elastic, rm texture
a
Table 4 Sensory evaluation of Microcosmus sabatieri marinades stored in air and vacuum packaging conditions, at 6 C
Air packaging
0
1
2
3
4
Vacuum packaging
0
1
2
3
4
5
MS1
MS2
MS3
MS1
MS2
MS3
MS1
MS2
MS3
MS1
MS2
MS3
1.0
1.8
3.4
6.0
7.0
1.0
2.0
3.5
6.2
7.0
1.0
1.6
3.3
6.1
7.0
1.0
2.1
3.4
6.2
7.0
1.0
2.0
3.5
6.2
7.0
1.0
2.0
3.7
6.4
7.0
1.0
2.5
3.9
6.2
7.0
1.0
2.6
4.2
6.6
7.0
1.0
2.6
4.3
6.5
7.0
1.0
2.2
4.3
6.3
7.0
1.0
2.1
4.2
6.9
7.0
1.0
2.2
4.0
6.8
7.0
1.0
1.2
1.8
3.1
3.9
7.0
1.0
1.2
2.0
3.3
5.2
7.0
1.0
1.0
1.6
3.3
3.8
7.0
1.0
1.3
2.1
3.4
5.2
7.0
1.0
1.4
2.0
3.5
5.2
7.0
1.0
1.4
2.0
3.6
5.3
7.0
1.0
1.1
2.5
3.8
5.2
7.0
1.0
1.1
2.6
4.0
5.5
7.0
1.0
1.1
2.6
4.0
5.5
7.0
1.0
1.0
2.2
4.0
6.0
7.0
1.0
1.0
2.1
4.2
5.9
7.0
1.0
1.0
2.2
4.0
5.8
7.0
a
Each number is the mean of three samples taken from the different experiments, coefficient of variation <5%; the sensory evaluation scale was 17
demerit points.
1711
1712
is in progress, and this has an important eect on

quality. The dierences in FA concentrations were
found to be useful as an index of freshness and
decomposition of marinated M. sabatieri in storage.
This study aimed to increase the value of the ascidia
M. sabatieri, a very popular seafood product in Dodecanese Greece, by marination. The marinated product
of fouska gained new delicious taste, acceptable texture
and an extra-long shelf life.
Acknowledgment
This study was supported by the Greek National project

Development of sheries activities in Kalymnos Island
aiming to the promotion of the local shery products
funded by the Region of Southern Aegean (P.E.P.
Notiou Aigaiou).
References
AOAC (1975). Ofcial Methods of Analysis. 12th edn (edited by S.
Williams). Pp. 129146. Washington, DC: Association of Ocial
Analytical Chemists.
Arkoudelos, J., Stamatis, N. & Samaras, F. (2007). Quality attributes
of farmed eel (Anguilla anguilla) stored under air, vacuum and
modied atmosphere packaging at 0 C. Food Microbiology, 24,
728735.
Bligh, E.G. & Dyer, W.J. (1959). A rapid method of total lipid
extraction and purication. Canadian Journal of Biochemistry and
Physiology, 37, 911917.
Cadun, A., Cakli, S. & Kisla, D. (2005). A study of marination of
deepwater pink shrimp (Parapenaeus longirostris, Lucas, 1846) and
its shelf life. Food Chemistry, 90, 5359.
Cardello, A.V., Sawyer, F.M., Maller, O. & Dygman, L. (1982).
Sensory evaluation of the texture and appearance of 17 species of
North Atlantic sh. Journal of Food Science, 47, 18181823.
Chaney, A.L. & Marbach, E.P. (1962). Modied reagents for
determination of urea and ammonia. Clinical Chemistry, 8, 130132.
Dalgaard, P. & Jorgensen, L.V. (1999). Cooked and brined shrimps
packed in a modied atmosphere have a shelf life of >7 months at
0 C, but spoiled in 46 days at 25 C. International Journal of Food
Drosinos, E.H. & Nychas, G.J.E. (1997). Production of acetate and
lactate in relation to glucose content during modied atmosphere
storage of gilt-head seabream (Sparus aurata) at 0 1 C. Food
Research International, 30(9), 711717.
Drosinos, E.H., Lambropoulou, K., Mitre, E. & Nychas, G.J.E.
(1997). Attributes of fresh gilt-head sea bream (Sparus aurata) llets
treated with potassium sorbate, sodium gluconate and stored under
a modied atmosphere at 0 1 C. Journal of Applied Bacteriology,
62, 403412.
Faulkner, D. (1995). Marine natural products. Journal of Natural
Product Reports, 12, 256258.
Fiala-Medioni, A. (1974). Ethologie alimentaire dinvertebres benthiques ltreurs (ascidies). II. Variations des taux de ltration et de
digestion en fonction de lespe`ce. Marine Biology, 28, 199206.
Fuselli, S.R., Casales, M.R., Fritz, R. & Yeannes, M.I. (1994).
Microbiology of the marination process in anchovy (Engraulis
anchoita) production. Lebensmittel-Wissenschaft und Technologie,
27, 214218.
Gawehn, K. (1984). D(-) Lactate. In: Methods of Enzymatic Analysis
(edited by H.U. Bergmeyer & M. Grassl). Pp. 588592. Weinheim,
VI: Verlag Chemie.
Gokoglu, N. (2002). A descriptive method for sensory evaluation of

mussels. Lebensmittel Wissenschaft und Technologie, 35, 563567.
Gokoglu, N., Cengiz, E. & Yerlikaya, P. (2004). Determination of the
shelf life of marinated sardine (Sardina pilchardus) stored at 4 C.
Food Control, 15, 14.
ICMSF (1986). International commission on microbiological specications for foods. Sampling plans for sh and shellsh. In: ICMSF,
Microorganisms in Foods. Sampling for Microbiological Analysis:
Principles and Scientic Applications. 2nd edn, Vol. 2. Pp. 181196.
Toronto, Canada: ICMSF, University of Toronto Press.
Jhaveri, S.N. & Constantinides, S.M. (1981). Chemical composition
and shelf life study of graysh (Squalus acanthias). Journal of Food
Science, 47, 188192.
Kamal-Eldin, A. & Yanishlieva, V.N. (2002). N-3 fatty acids for
human nutrition: stability considerations. European Journal of Lipid
Karl, H. (1992). Vergleich versciedener Methoden zur Essigsaeure
und Kochsalz-Bestimmung in Marinaden. Informationen fur die
Fischwirtschaft, 39, 137143.
Kates, M. (1972). Techniques of Lipidology: Isolation, Analysis and
Identication of Lipids. Pp. 362364. New York: North-Holland
Pub. Co.
Kilinc, B. & Cakli, S. (2005). Determination of the shelf life of sardine
(Sardina pilchardus) marinates in tomato sauce stored at 4 C. Food
Control, 16, 639644.
Koukouras, A., Voultsiadou-Koukoura, E., Kevrekidis, T. & Vadis,
D. (1995). Ascidian fauna of the Aegean Sea with a checklist of the
eastern Mediterranean and Black Sea species. Annales Institutes
Oceanographique Paris, 71, 1934.
Kromhout, D., Bosschiter, E.B. & Lezenne, C.C. (1985). The inverse
relationship between sh consumption and 20-year mortality from
coronary heart disease. The New England Journal of Medicine, 312,
12051209.
Lin, J.-K., Lee, Y.-J. & Chang, H.W. (1983). High concentrations of
dimethylamine and methylamine in squid and octopus and their implications in tumour aetiology. Food and Chemical Toxicology, 21, 2.
Lyday, P.A.(2007). Minerals Information. Pp. 3637. Washington DC:
Geological Survey, U.S. Department of the Interior. 20440.
Masniyom, P., Benjakul, S. & Visessanguan, W. (2002). Shelf life
extension of refrigerated sea bass slices under modied atmosphere
packaging. Journal of the Science of Food and Agriculture, 82, 873880.
McLay, R.(1972). Marinades. Torry Advisory Note 56. Pp. 110. Torry
Research Station: Ministry of Agriculture, Fisheries and Food.
Metcalfe, L.D., Schmitz, A.A. & Pelka, J.R. (1966). Preparation of
fatty acid esters from lipids for gas chromatographic analysis.
Analytical Chemistry, 38, 514516.
Monniot, C. (1962). Les Microcosmus des cotes de France. Vie et
Milieu, 12, 397432.
Monniot, C. (1965). Etude Faunistique et evolutive de la famille des
Pyuridae (Ascidiacea). Memoires du Museum National dHistoire
Naturelle, Serie A, Zoologi, XXXVI, 204.
Monniot, C. & Monniot, F.(1987). Ascidies (Ciones at Violet). In:
Fiches FAO didentication des espe`ces pour les besoins de la peche
(revision I) Mediterranee et Mer Noire (edited by W. Fischer, M.L.
Bouchon & M. Scneider). Pp. 743760. Rome: Zone de peche 37.
Nielsen, D., Hyldig, G., Nielsen, J. & Nielsen, H.H. (2005). Sensory
properties of marinated herring (Clupea harengus) processed from
raw material from commercial landings. Journal of the Science of
Food and Agriculture, 85, 127134.
Odalapa, A., Akin, M.A.S. & Olusegun, L.O. (1984). Quality changes
of Nigerian traditionally processed freshwater sh species, II
Chemical composition. Journal of Food Technology, 19, 341348.
Ozden, O. (2005). Changes in amino acid and fatty acid composition
during shelf-life of marinated sh. Journal of the Science of Food and
Poligne, I. & Collignan, A. (2000). Quick marination of anchovies
(Engraulis enchrasicolus) using acetic and gluconic acids. Quality
and stability of the end product. Lebensmittel-Wissenschaft und

Technologie, 33, 202209.
Rehbein, H. & Oehlenschlager, J. (1996). Fische und Fischerzeugnisse.
Pp. 395411. Paul Parey Verlag, Hamburg-Berlin: Krebs und
Weichtiere.
Rinehart, K.L. (2000). Antitumor Compounds from Tunicates.
Medical Care Research and Review, 20, 127.
Rosa, R. & Nunes, M.L. (2004). Nutritional quality of red shrimp,
Aristeus antennatus (Risso), pink shrimp, Parapenaeus longirostris
(Lucas), and Norway lobster, Nephrops norvegicus (Linnaeus).
Sallam, Kh.I., Ahmed, A.M., Elgazzar, M.M. & Eldaly, E.A. (2007).
Chemical quality and sensory attributes of marinated Pacic saury
(Cololabis saira) during vacuum-packaged storage at 4 C. Food
Chemistry, 102, 10611070.
Simeonidou, S., Govaris, A. & Vareltzis, K. (1998). uality assessment
of seven Mediterranean sh species during storage on ice. Food
Slantchev, K., Yalcin, F., Ersoz, T. et al. (2002). Composition of
Lipophylic Extracts from Two Tunicates, Styela sp and Phallusia sp.
from the Eastern Mediterranean. Zeitschrift der Naturforschung,
57C, 534540.
Sokal, R.R. & Rohlf, F.J.(1987). Introduction to Biostatistics. P. 363.
New York: W. H. Freeman and Company.
Stamatis, N. & Arkoudelos, J. (2007a). Eect of modied atmosphere
and vacuum packaging on microbial, chemical and sensory quality
indicators of fresh, lleted Sardina pilchardus at 3 C. Journal of the

Science of Food and Agriculture, 87, 6.
Stamatis, N. & Arkoudelos, J. (2007b). Quality assessment of Scomber
colias japonicus under modied atmosphere and vacuum packaging.
Food Control, 18, 4.
Stamatis, N., Vadis, D. & Chintiroglou, C. (2004). Quality improvement of spinialo and processed products from sea urchin. NAGREF Agricultural Research & Technology, 15, 1921.
Stamatis, N., Monios, G., Stergiou, D. & Vadis, D.(2007).
Evaluation of some quality parameters in spinialo of sea squirt
(Microcosmus sabatieri) from Kalymnos Island Greece. In:
Proceedings of the 5th International Congress in Food Technology
(edited by E.S. Lazos). Pp. 624631, Vol. III. Thessaloniki,
Greece: Hellenic Association of Food Technologists.
Tursi, A. (1980). Guide per il riconoscimento delle specie animali delle
acque lagunari e costiere italiane. AQ 1 93, 4. Ascidiacei. Genova:
Consiglio Nazionale delle Ricerche.
Voldrich, M., Doblos, J., Kalac, P. & Curda, D. (1991). Changes of
fatty acid composition during the processing of sh. Nahrung, 9,
663664.
Wang, M.Y. & Brawn, W.D. (1983). Eects of elevated CO2
atmosphere on storage of freshwater craysh (Pacifastacus leniusculus). Journal of Food Science, 43, 158162.
Zlatanos, S. & Sagredos, A.N. (1983). The fatty acids composition of
some important Mediterranean sh species. Fat Science and Technology, 95, 6669.
1713
1714
Original article
Effect of freezethaw cycles and additives on rheological and
sensory properties of ready to bake frozen chapaties
Deep N. Yadav,* Prakash E. Patki, Mohammad A. Khan, Gopal K. Sharma & Amrindar S. Bawa
Defence Food Research Laboratory Siddarthanagar, Mysore 570 011, India
(Received 01 October 2007; Accepted in revised form 15 April 2008)
Summary
The wheat grains were conditioned at 18% moisture level for 1 h followed by heating for 80 s in a microwave
oven to reduce the activity of polyphenol oxidase as well as other lipolytic and oxidative enzymes. The dough
samples were prepared from whole wheat our (milled from microwave treated wheat grains) using 68%
water and 2% salt on our weight basis as common ingredients. The additives like glycerol monostearate and
glycerol (5% and 1% respectively on our weight basis) were also used in order to study their eects on the
quality of ready to bake frozen chapaties (R-BFC) during freezethaw cycles. Alveographic properties of
R-BFC as well as textural prole and sensory quality of chapaties prepared from R-BFC samples were
evaluated during repeated freezethaw (FT) cycles. Results showed that the alveographic properties (P, L,
and W) decreased signicantly (P 0.05) in all the samples after each FT cycle, while glycerol added samples
showed the least changes. Chapati hardness increased in all the samples up to fourth FT cycle, thereafter
sharply decreased. However, chapaties with 1% glycerol were rated better in terms of mouth feel and texture
during FT cycles.
Keywords
Additives, alveographic properties, freezethaw cycles, ready to bake frozen chapaties, texture prole.
Introduction
People in India have been using wheat since times

immemorial for making chapaties, a at unleavened hot
plate baked product prepared from whole wheat our.
About 8590% wheat consumption in India is in the
form of chapati and its culinary preparations such as
tandoori roti, nan, paratha and poori (Gujral & Gaur,
2002). During the last few years, consumers and buyers
are becoming increasingly aware of the importance of
safe and high-quality food products. Interest becomes
greater as new products are introduced into the market
and modern technologies are being used even in the
production of traditional or conventional food products (Use of freezing in bakery products), (Giannou
et al., 2003). Frozen dough creates the opportunity for
users to steer away from the use of skilled labour, as a
result, many more restaurants are now able to oer
freshly baked product to their customers at reasonable
prices. However, during frozen storage, various physical, rheological and sensory changes occur in the
food products, which determine the consumers acceptability. The frozen food sometimes thawed due to the
e-mail: dnyadav1977@yahoo.co.in
electricity failure as well as during the transportation.

Several studies have shown freezethaw cycles may
damage protein structure (Varriano et al., 1980; Berglund et al., 1991; Autio & Sinda, 1992; Inoue & Bushuk,
1992; Rasanen et al., 1998). The use of additives to
minimise the damage in the gluten network caused due
to frozen storage and freezethaw cycles has become a
common practice in the baking industry. Kobbs (1997)
has suggested that there could be a possibility of
reducing the freezethaw damage by incorporating
gums in frozen dough. Ward & Andon (1993) mentioned that gums such as carboxymethyl cellulose,
carragenan, gum arabic and locust bean gum may be
used to alleviate the problems associated with frozen
dough. Scwharzla et al. (1996) observed positive
eects on the shelf life of bread by partially replacing
our with locust bean gum and guar gum. The eect of
sodium stearoyl-2-lactylate, diacetyl tartaric acid esters
of monoglyceride, glycerol monostearate (GMS) and
distilled glycerol monostearate surfactants gels on
dough rheological characteristics and quality of bread
was investigated by several authors (Wolt & DApplonia, 1984; Inoue et al., 1995; Sharadanant & Khan,
2003; Azizi & Rao, 2004; Asghar et al., 2005). There
are quite a number of reports on frozen dough of yeast
raised products and most of them have identied
doi:10.1111/j.1365-2621.2008.01763.x
Effect of freezethaw cycles and additives D. N. Yadav et al.
factors inuencing the quality of bread produced from

frozen dough. Only a few studies have been reported on
non-yeasted frozen dough (Varriano et al., 1980; Indrani et al., 2002; Yadav et al. (2007a)). Earlier Yadav
et al. (2007a) developed ready to bake frozen chapati
(R-BFC) and studied the rheological and sensory
changes for prolonged storage (6 months) and reported
that R-BFC samples retained their texture and sensory
quality up to 6 months of frozen storage at )18 C;
however, they observed that extensibility of dough
increased and maximum resistance to extension decreased during storage. The hardness and chewiness of
chapati prepared from R-BFC samples also increased
during frozen storage. However, the eect of freeze
thaw cycles and additives on the quality of R-BFC still
needs investigation, because the freezethaw cycle
damage the gluten structure, hence alters the texture
of chapaties and additives may minimise these changes.
Therefore, the experiments were conducted to study the
eects of additives such as GMS and glycerol and
repeated freezethaw cycles on the rheological characteristics of R-BFC and textural as well as sensory
quality of chapaties, baked from R-BFC.
investigation were of analytical grade and procured

from M s E-Merck, Mumbai, India.
Preparation of ready to bake frozen chapati (R-BFC)
Dough from the wheat our samples was prepared by

mixing the our in Hobart dough kneader (HL 120 Hz
50 60, Hobart, Troy, OH, USA) with 68% water and
2% salt as common ingredients (on our weight basis)
for 10 min. Three batches were prepared, i.e. (1) control
(without additives), (2) 5% GMS (2:1:1, water:oil:GMS)
and (3) 1% Glycerol. The concentration of GMS and
glycerol were selected on pre-trial basis. The dough was
allowed to rest for 15 min and divided in to 40 g pieces.
Each dough piece was formed into spherical shape and
pressed using pneumatic chapati pressing machine
(Sunrise Industry, Mysore, India) into a attened
circular dough sheet (Diameter 150.0 mm and thickness
2.00 mm). Four attened circular dough sheets were
packed in a polyethylene pouches (75 lm), with polypropylene lining (75 lm) in between each chapati and
heat sealed.
Freezing and storage
Microwave treatment and milling of wheat grains
In the present study, commercial, medium hard variety

(Triticum aestivum cv. bansi), grown during 2007, in the
state of Karnataka was procured from the local market,
manually cleaned and used in the investigation.
Wheat grains were microwave treated to reduce the
activity of polyphenol oxidase (PPO) as well as other
lipolytic and oxidative enzymes (Yadav et al., 2007a).
Initially wheat grains were conditioned at 18% moisture
level for 1 h followed by heating for 80 s in a microwave
oven (LG Electronics India Pvt. Ltd, Delhi, IndiaModel
MG 607 APR) at full power (900 W, 2450 MHz). After
treatment, the samples were immediately allowed to cool
under ambient temperature conditions (23 2 C) and
then milled to 400500 lm particle size in an emery disc
mill (Model No. EGM-467K, Dia: 18 inch, Ganesha &
Company, Chennai, India) to obtain whole wheat our
(100% extraction rate), which is commonly used in
households for chapati making.
The samples were frozen in a blast freezer (Model no.

15L-U2-300; Hull Corporation, Hatboro, PA, USA) at
)30 C for 30 min to achieve the core temperature of
)18 C. After freezing, the samples were immediately
stored in a chest freezer (Model No 425; Blue Star Ltd,
Mumbai, India) at )18 C for further study.
Freezethaw cycles
To see the eect of freezethaw cycles, the R-BFC

samples were partially thawed at 45 C in a refrigerator
for 16 h and after that three packets (each packets have
four chapaties) of each samples were immediately baked
on a hot plate (230 5 C), each side baked for 2.0
3.0 min with the application of hydrogenated vegetable
oil to a golden brown colour. The rest of the R-BFC
samples were again kept in freezer at )18 C for 48 h,
this completes one freezethaw (FT) cycle. The frozen
stored samples of R-BFC were withdrawn again from
the freezer after each 48 h and thawing and baking were
done as above for the second FT cycle and the process
was continued till the completion of seven FT cycles.
Chemical composition of flour
The chemical parameters like protein, fat, ash and

gluten contents were determined according to approved
American Association of Cereal Chemist (1983) methods 4611, 3025, 0811 and 3811 respectively. The
moisture content of samples was recorded using Brabender Rapid-Moisture Meter (Type 890100; Brabender,
Duisburg, Germany). All the chemicals used in the
Dough quality testing
Two packets (300 g) of partially thawed samples (at 4

5 C for 16 h) were again thawed at room temperature
(23 2 C) for 6 h and then mixed in consistograph for
45 min after that dough quality characteristics were
measured using the Alveograph (Alveolink NG Consistograph; Villeneuve La Garenne, Chopin, France)
1715
1716
following the approved (American Association of

Cereal Chemist, 2000) method (55-21). The tenacity or
maximum pressure required for the deformation (P),
extensibility (L), conguration ratio of the curve (P L)
as well as baking strength (W) were recorded from the
alveograph curve.
Texture profile analysis
Textural prole analysis of chapaties was performed

using a Texture Analyzer Plus (Model No.
01 TALS LXE UK; LLOYD Instruments, Hampshire,
UK). The bite test on chapati was carried out using the
volodke bite set (part no. 01 2663). The volodke bite
set is designed to imitate incisor teeth shearing through
a food sample. The set comprises upper and lower
teeth which during the test penetrate the sample twice
to obtain the peak force and the texture prole at
the preset speed and position. The chapati strips
(2 15 mm) were axially compressed to 90% of their
original height, avoiding fracture. Forcetime deformation curve were derived with a 50.0 Kg load cell applied
at a crosshead speed of 10.0 mm min. Attributes were
calculated as follows: hardness value is the peak force
(N) of the rst compression of the sample; cohesiveness
is the area of work during the second compression
divided by the area of work during the rst compression (dimensionless); springiness is the distance (mm)
that the sample recovers after the rst compression;
chewiness (Nmm) is the product of the attributes
(gumminess springiness) and in sensory terms corresponds to the energy required to chew a solid food
product. Ten measurements per replication were taken

for all textural analysis.
Colour value
The colour values in terms of L, a & b for R-BFC and

baked chapatis were measured using a Hunter Colour
Meter (Data lab; Silvasa, Gujarat, India) with illuminant D65 and 100 observer. The partially thawed samples
were re-thawed at room temperature (23 2 C) for
2 h before measuring the colour. A higher L value
indicated a brighter or whiter sample. Values of a and b
indicated the redgreen and yellowblue chromaticity
respectively.
Sensory analysis
The sensory attributes of chapati in each experimental

block were evaluated in terms of colour, aroma, mouth
feel, texture and overall acceptability by a trained panel
consisting of ten scientic sta members of the laboratory with knowledge of consumers preference for
chapaties using a nine-point hedonic scale having a
score of 9 for extreme liking and 1 for extreme disliking
(Larmond, 1977).
All the experiments were performed in triplicate. Analysis of variance was carried out as per Snedecor &
Cochran (1968) using statistica software version-7(State
soft Corporation; Tulsa, OK, USA).
Table 1 Alveographic parameters of R-BFC as affected by additives and freezethaw (FT) cycles (n = 5)
Alveographic parameters
P
PL
No. of FT cycles
GMS
Gly
GMS
Gly
GMS
Gly
GMS
Gly
0
1
2
3
4
5
6
7
LSD (P 0.05) a
b
ab
SEM a
b
ab
158.3
157.5
156.3
153.0
150.0
145.0
139.0
135.6
159.5
158.7
157.5
155.0
153.0
149.0
145.0
142.5
0.54
0.88
1.53
0.19
0.31
0.54
156.3
156.0
155.0
154.0
154.0
152.0
150.0
148.3
34.6
34.7
34.0
32.3
30.0
28.3
26.8
25.3
29.5
29.7
29.3
28.4
27.5
25.8
23.7
22.7
0.10
0.17
0.30
0.04
0.06
0.10
31.0
31.2
29.8
29.5
29.0
28.3
27.8
27.5
4.59
4.47
4.72
4.90
5.09
5.22
5.57
6.06
5.41
5.34
5.38
5.46
5.56
5.78
6.12
6.28
0.03
0.05
0.08
0.01
0.02
0.03
5.04
5.00
5.20
5.22
5.31
5.37
5.40
5.39
158
157
154
152
149
145
138
134
155
155
151
148
143
138
132
127
1.30
2.13
3.68
0.46
0.75
1.29
156
155
153
152
150
147
145
142
C, control; GMS, 5% glycerol monostearate; Gly, 1% glycerol; a, type of sample; b, no. of FT cycles; LSD, least significance difference; SEM, standard
error of mean.
The our samples were analysed for its moisture, gluten,

total protein, fat, total carbohydrate and ash contents
and the corresponding values were 10.20 0.15%,
9.7 0.05%, 13.3 0.1%, 1.3 0.08%, 73.6
0.08% and 1.6 0.05% respectively.
Dough quality
Dough extension properties are very important when

evaluating frozen dough as they inuence the quality of
baked product. Many researchers used the extensiograph for measurement of extension properties of frozen
dough, in the present study we used Chopin alveograph
for the same purpose and results obtained have been
presented in Table 1. Higher resistance to extension (Pvalue in alveograph) value indicates strong wheat our
and correlates with good baking characteristics (Inoue &
Bushuk, 1992). The P-value (158.3) decreased signicantly (P 0.05) in control samples after rst FT cycle
but the decrease was not signicant in GMS added
samples, thereafter it showed signicant (P 0.05)
decrease; however, it remained insignicant in glycerol
added samples up to second FT cycle. The variation in
P-value of all the three samples was found signicant
(P 0.05) after seven FT cycles and the least decrease
was observed (156.3148.3) in glycerol added samples
followed by GMS (159.5142.5) and control (158.3
135.6) samples. Sharadanant & Khan (2003) also
reported continuous decrease in maximum resistance
to extension in yeasted dough during 16 weeks of frozen
storage. They also reported that addition of CMC, kcarragenan and locust bean gum increased the maxi-
mum resistance of extension in yeasted dough as

compared with control (without gum) samples. Earlier,
we (Yadav et al., 2007a,b) observed an increase in
dough (non-yeasted) extensibility during continuous
frozen storage up to 6 months, but in this study it was
observed that extensibility increased slightly at rst
freezethaw cycle, thereafter it continuously showed a
decreasing trend. It may be due to the rst weakening of
the bonds in the gluten network, thereafter breakage of
the covalent bonds in the network, resulting less
extensible gluten (Inoue & Bushuk, 1991). At the end
of seventh FT cycle, the extensibility (27.5) of glycerol
added samples were signicantly (P 0.05) higher than
GMS added samples (22.7) and control (25.3). Higher
retention of extensibility value in glycerol added samples
indicated that the quality of gluten in the dough has
been retained to a greater degree than the control and
GMS added samples. Wolt & DApplonia (1984) also
found a decrease in extensibility with an increase in
frozen time, which was attributed to overall gluten
network deterioration. The baking strength (W) also
showed decreasing trend due to FT cycles. The decrease
was not signicant up to three FT cycles in glycerol
samples whereas it showed signicant (P 0.05) reduction in GMS (151) and control (154) samples after
second FT cycle. The maximum decrease was observed
in GMS samples (155.0127.0) followed by control
(158.0134.0) and glycerol samples (156.0142.0) after
seven FT cycles. The glycerol added samples showed
signicantly (P 0.05) higher retention of the alveographic properties (P, L, and W) over the control and
GMS samples. It showed that glycerol works as a
cryoprotectants and can minimise the dough quality
losses caused due to frozen storage and FT cycles,
Table 2 Colour values of R-BFC as affected by additives and freezethaw (FT) cycles (n = 3)
L
No. of FT cycles
GMS
Gly
GMS
Gly
GMS
Gly
0
1
2
3
4
5
6
7
LSD (P 0.05) a
b
ab
SEM a
b
ab
64.86
64.72
64.63
64.55
64.12
63.73
62.56
61.86
65.13
64.89
64.76
64.65
64.16
63.78
62.88
62.58
0.02
0.04
0.06
0.008
0.01
0.02
64.96
64.83
64.70
64.48
64.32
63.93
63.15
62.45
3.63
3.63
3.65
3.68
3.71
3.73
3.76
3.78
3.60
3.62
3.63
3.65
3.69
3.71
3.75
3.78
0.01
0.02
0.03
0.004
0.006
0.01
3.62
3.63
3.65
3.67
3.68
3.70
3.72
3.75
15.85
15.67
15.45
15.27
15.12
14.96
14.85
14.78
15.92
15.73
15.52
15.32
15.18
15.05
14.92
14.85
0.02
0.03
0.05
0.006
0.009
0.002
15.88
15.72
15.57
15.40
15.22
15.15
15.06
14.95
error of mean.
1717
probably due to the fact that glycerol being a humectants, absorb the free water of the dough and thus
making the water unavailable for the recrystalisation
and thus preventing the damage of gluten protein caused
by the ice crystals.
Colour value of R-BFC
The colour values of R-BFC samples have been shown

in Table 2 and the corresponding values for baked
chapaties were presented in Fig. 1. Brightness (L) values
L value
(a) 68.5
68
67.5
67
66.5
66
65.5
65
64.5
64
63.5
63
Control
GMS
Gly
No.of FT cycles
(b) 3.35
Control
a value
3.3
GMS
Gly
3.25
3.2
3.15
3.1
3.05
No. of FT cycles
(c) 13.7
13.6
13.5
b value
1718
13.4
13.3
13.2
Control
13.1
GMS
13
Gly
12.9
12.8
0
No. of FT cycles
Figure 1 (a) Changes in L value of baked chapaties prepared from
R-BFC samples during freezethaw (FT) cycles. (b) Changes in a value
of baked chapaties prepared from R-BFC samples during FT cycles.
(c) Changes in b value of baked chapaties prepared from R-BFC
samples during FT cycles.
decreased continuously after each FT cycle; however,

the decrease was not signicant up to fourth FT cycle in
all the samples, thereafter it showed signicant
(P 0.05) decrease. Though all the samples were
acceptable upto seven FT cycles as indicated by overall
acceptability scores of the baked chapaties (Table 4).
The redness values showed increasing trend; however,
the increase was not signicant in any sample while the
yellowness (b) value showed a decreasing trend, but all
the samples were acceptable at the end. Eects of GMS
and glycerol on colour values of R-BFC and chapaties
baked from it were not signicant. Generally lightness
(L) value decreases continuously in dough during
storage, but in our study we have used the our samples,
which was obtained from microwave treated wheat
grains. The microwave treatment of wheat grains
reduces the PPO activity in resultant our, retarding
the dough disclouration during storage (Yadav et al.,
2007a); hence less change in colour values of R-BFC
samples were observed in this study.
Texture profile of chapaties
Chapaties baked from R-BFC samples were subjected to

textural prole analysis and the values have been
reported in Table 3. It was interesting to note that the
hardness value increased signicantly (P 0.05) up to
fourth FT cycle in all the samples; thereafter it showed a
decreasing trend. This may be due to the severe loss of
polymer cross-linking in gluten network after fourth FT
cycle as well as decrease in starchprotein interactions
(Autio & Sinda, 1992). Glycerol added samples showed
signicantly (P 0.05) less increase in hardness value
(4.434.68 N) followed by GMS added (4.324.91 N)
and the control samples (4.555.32 N) up to fourth FT
cycle; however, the increase was signicant (P 0.05) in
all the samples. At the end of the seventh FT cycle, the
control samples had least hardness (3.12 N) followed by
GMS (3.75 N) and glycerol (3.95 N) added samples, but
the dierences were signicant (P 0.05). Sensorily, it
was observed that after fourth FT cycle the control
chapaties were more brittle as compared with GMS and
glycerol added ones. Brittleness of chapati is not a
desired characteristic (Haridas Rao et al., 1986). All the
other textural parameters such as cohesiveness, springiness and chewiness decreased signicantly (P 0.05)
during each FT cycles and the reduction was signicantly (P 0.05) higher in control samples as compared
with glycerol and GMS added ones. The springiness
value after seventh FT cycles in all the samples diered
signicantly (P 0.05), showing maximum (0.94 mm)
for glycerol added followed by GMS (0.85 mm) and
control (0.50 mm). Chewiness, the force required to
chew a sample was also signicantly (P 0.05) higher in
glycerol added samples (0.85 N) as compared with GMS
added (0.64 N) and control (0.23 N). A chewiness value
Table 3 Textural characteristics of chapati prepared from R-BFC as affected by additives and freezethaw (FT) cycles (n = 10)
Hardness (N)
Cohesiveness
Springiness (mm)
Chewiness (Nmm)
No. of FT cycles
GMS
Gly
GMS
Gly
GMS
Gly
GMS
Gly
0
1
2
3
4
5
6
7
LSD (P 0.05) a
b
ab
SEM a
b
ab
4.55
4.84
5.01
5.25
5.32
3.85
3.76
3.12
4.32
4.58
4.75
4.84
4.91
4.00
3.88
3.75
0.03
0.05
0.08
0.01
0.02
0.03
4.43
4.48
4.58
4.63
4.68
4.25
4.12
3.95
0.31
0.29
0.22
0.18
0.17
0.17
0.15
0.15
0.34
0.32
0.31
0.29
0.26
0.24
0.22
0.20
0.006
0.009
0.02
0.002
0.003
0.006
0.36
0.35
0.33
0.32
0.31
0.28
0.25
0.23
1.12
1.03
0.98
0.90
0.82
0.70
0.57
0.50
1.32
1.24
1.15
1.08
1.05
0.92
0.88
0.85
0.01
0.02
0.03
0.004
0.006
0.01
1.38
1.32
1.24
1.15
1.10
0.98
0.96
0.94
1.56
1.43
1.09
0.85
0.74
0.46
0.32
0.23
1.96
1.82
1.67
1.50
1.36
0.89
0.75
0.64
0.02
0.04
0.06
0.008
0.01
0.02
2.22
2.09
1.89
1.71
1.58
1.15
0.99
0.85
error of mean.
Table 4 Sensory characteristics of chapati prepared from R-BFC as affected by additives and freezethaw (FT) cycles (n = 10 panelists)
Colour
Aroma
Mouth feel
Texture
Overall acceptability
No. of FT cycles
GMS
Gly
GMS
Gly
GMS
Gly
GMS
Gly
GMS
Gly
0
1
2
3
4
5
6
7
LSD (P 0.05) a
b
ab
SEM a
b
ab
8.8
8.7
8.5
8.3
8.2
8.0
8.0
7.8
8.7
8.7
8.5
8.4
8.2
8.1
8.0
7.7
0.12
0.08
0.20
0.04
0.03
0.05
8.8
8.7
8.6
8.4
8.3
8.2
8.1
7.8
8.4
8.3
8.2
8.2
8.0
8.0
7.9
7.8
8.3
8.2
8.2
8.0
8.0
7.9
7.9
7.7
0.09
0.06
0.13
0.03
0.02
0.05
8.4
8.3
8.3
8.2
8.1
8.0
8.0
7.9
8.5
8.4
8.3
8.1
8.0
8.0
7.8
7.7
8.4
8.3
8.3
8.2
8.1
8.1
8.0
8.0
0.08
0.15
0.21
0.03
0.04
0.06
8.4
8.4
8.4
8.3
8.2
8.0
8.0
8.0
8.5
8.5
8.4
8.2
8.0
7.8
7.7
7.4
8.7
8.6
8.5
8.4
8.2
8.0
7.8
7.7
0.10
0.08
0.15
0.03
0.02
0.05
8.7
8.6
8.6
8.5
8.5
8.4
8.2
8.0
8.5
8.3
8.3
8.1
8.0
7.8
7.8
7.6
8.6
8.5
8.4
8.3
8.1
8.0
7.9
7.8
0.12
0.06
0.17
0.04
0.03
0.06
8.7
8.6
8.5
8.4
8.2
8.2
8.1
8.0
error of mean.
up to certain extent is also a desired characteristic of

chapaties (Haridas Rao et al., 1986). Springiness and
chewiness data of the chapaties were correlated to the
tenacity (P) value of the dough and observed a
signicant (P 0.05) correlation (r = +0.88) between
tenacity of the dough and springiness as well as tenacity
of the dough and chewiness (r = +0.87) of the chapaties. With these results it has been concluded that the
dough with weak gluten network results in a chapati
with lower springiness and chewiness values, suggesting
that such chapaties may be physically more brittle or
subjected to easy tear.
Sensory characteristics
Sensory data for baked chapaties are given in Table 4.

All the sensory parameters were signicantly (P 0.05)
aected by FT cycles. A sensory score of more than 7.0
was considered acceptable and based on this assumption
all the samples, control as well as GMS and glycerol
added ones were acceptable in terms of all the sensory
parameters even after seven FT cycles. The maximum
scores for mouth feel (8.0), texture (8.0) and overall
acceptability (8.0) were observed in glycerol added
samples at the end of seventh FT cycle and were
1719
1720
signicantly (P 0.05) higher from control (7.7, 7.4 and

7.6 respectively) and GMS added samples (8.0, 7.7, and
7.8 respectively). The sensory scores for texture continuously decreased during the FT cycles (Table 4),
whereas instrumental data (Table 3) for hardness
showed an increase in all the samples up to fourth FT
cycles followed by a sharp decrease. This showed that
for a better acceptability, chapaties should have hardness between 3.9 and 4.4 N and chewiness between 0.85
and 2.2 Nmm.
Conclusion
The study revealed that additives like GMS and glycerol

do not completely overcome the negative eect of frozen
storage and repeated freezethaw cycles. Addition of
GMS (5%) did not show any signicant improvement in
the quality of R-BFC. However, glycerol (1%) minimised the rheological changes in ready to bake frozen
chapaties occurring during freezethaw cycles and
frozen storage as well as improved their textural and
sensory properties. This will pave the way for the
popularisation of frozen dough technology for the
popular Indian traditional at bread, i.e. chapaties.
References
American Association of Cereal Chemist. (1983). Approved Methods of
the AACC, 8th edn. St. Paul, MN: The Association.
American Association of Cereal Chemist. (2000). Approved Methods of
the AACC, 10th edn. St. Paul, MN: The Association.
Asghar, A., Anjum, F.M., Tariq, M.W. & Hussain, S. (2005). Eect
of carboxymethyl cellulose and gum arabic on the stability of
frozen dough for bakery products. Turkish Journal of Biology, 29,
237241.
Autio, K. & Sinda, E. (1992). Frozen doughs: rheological changes and
yeast viability. Cereal Chemistry, 69, 409413.
Azizi, M. & Rao, G. (2004). Eect of surfactants gels on dough
rheological characteristics and quality of bread. Critical Review in
Berglund, P.T., Shelton, D. & Freeman, T. (1991). Frozen bread
dough ultra structure as aected by duration of frozen storage and
freeze thaw cycles. Cereal Chemistry, 68, 105107.
Giannou, V., Kessoglou, V. & Tzia, C. (2003). Quality and safety
characteristics of bread made from frozen dough. Trends in Food
Gujral, H.S. & Gaur, S. (2002). Eect of barley our, wheat gluten and
liquid shortening on the texture and storage characteristics of
chapati. Journal of Texture Studies, 33, 461469.
Haridas Rao, P., Leelavathi, K. & Shurpalekar, S.R. (1986). Test
baking of chapati development of a method. Cereal Chemistry, 63,
297303.
Indrani, D., Prabhashankar, P., Rajiv, J. & Rao, G.V. (2002). Eect of
storage on the rheological and parotta-making characteristics of
frozen parotta dough. European Food Research and Technology, 215,
484488.
Inoue, Y. & Bushuk, W. (1991). Studies on frozen dough I. Eects of
frozen storage and freezethaw cycles on baking and rheological
properties. Cereal Chemistry, 68, 627631.
Inoue, Y. & Bushuk, W. (1992). Studies on frozen dough II. Flour
quality requirements for bread production from frozen dough.
Inoue, Y., Sapirstein, H.D. & Bushuk, W. (1995). Studies on frozen
dough. IV. Eect of shortening system on baking and rheological
properties. Cereal Chemistry, 72, 221226.
Kobbs, L. (1997). Designing frozen foods. Food Product Design, 1, 27
43.
Larmond, E. (1977). Laboratory Methods for Sensory Evaluation of
Foods. Pp. 16371662. Ottawa: Canada Department of Agric
Publication.
Rasanen, J., Blanshard, J.M.V., Mitchell, J.R., Derbyshire, W. &
Autio, K. (1998). Properties of frozen wheat doughs at subzero
temperatures. Journal of Cereal Science, 28, 114.
Scwharzla, S.S., Johnson, J.M., Barbeau, W.E. & Duncan, S. (1996).
Guar and Locust bean gums as a replacers of all purpose our in
bread: an objective and sensory evaluation. Journal of Food Quality,
19, 217229.
Sharadanant, R. & Khan, K. (2003). Eect of hydrophilic gums on
frozen dough. I- dough quality. Cereal Chemistry, 80, 764772.
Snedecor, G.N. & Cochran, W.G.. (1968). Statistical Methods, 6th
edn. Pp. 234257. New Delhi: Oxford and IBH Pub, Co.
Varriano, M.E., Hsu, K.H. & Mahdi, J. (1980). Rheological and
structural changes in frozen dough. Bakers Digest, 54, 3236.
Ward, F. & Andon, S. (1993). The use of gums in bakery foods from
American Institute of Baking. Technical Bulletin, 15, 18.
Wolt, M.J. & DApplonia, B.L. (1984). Factors involved in the
stability of frozen dough. II. The eects of yeast type, our type and
dough additives on frozen dough stability. Cereal Chemistry, 61,
213221.
Yadav, D.N., Patki, P.E., Sharma, G.K., Srihari, K.A., Harilal, P.T. &
Bawa, A.S.. (2007a). Rheological and sensory quality of ready to
bake chapati (R-BFC) during frozen storage. Journal of Food
Quality (In press).
Yadav, D.N., Patki, P.E., Sharma, G.K. & Bawa, A.S.. (2007b). Eect
of microwave heating of wheat grains on the browning of dough and
quality of chapaties. International Journal of Food Science and
Technology. [Online: 19 December 2007, doi: 10.1111/j.1365-2651.
2007.01595.x].
Short communication
Enzymatic coagulation of milk: animal rennets and microbial
coagulants differ in their gelation behaviour as affected
by pH and temperature
Doris Jaros,* Katrin Seitler & Harald Rohm
Institute of Food Technology and Bioprocess Engineering, Technische Universitat Dresden, 01069 Dresden, Germany
(Received 8 January 2008; Accepted in revised form 28 March 2008)
Keywords
Gelation, milk clotting, rheology.
Introduction
The enzymatic coagulation of milk is the rst important

step during cheese-making and undoubtedly aects the
cheese quality (Harboe & Budtz, 1999). After cleaving
j-casein, which, mainly by steric repulsion, is responsible
for the stability of the casein micelles in native milk
(Walstra, 1990; Holt & Horne, 1996), the destabilised
casein micelles aggregate, gel formation starts and a
three-dimensional network develops as a consequence of
casein micelle fusion. Technologically important parameters, such as coagulation time, gel rmness and elasticity,
the fraction of hydrolysed j-casein or syneresis depend
not only on the intrinsic factors of the milk [e.g. casein
concentration or polymorphism, pre-heat treatment, pH,
temperature, fat homogenisation, calcium content (Lomholt & Qvist, 1999; Karlsson et al., 2007a)], but also on
specic properties of the coagulant (OCallaghan et al.,
2000). Gel rmness at cutting below a critical level leads
to a loss of nes in the whey, thus decreasing cheese yield
(Riddell-Lawrence & Hicks, 1989; Johnson et al., 2001);
however, a gel rmness above an upper critical value
might also result in a higher loss of nes because more
energy is needed for cutting (Walstra et al., 2006).
Apart from traditional rennet extracted from the
abomasum of ruminants (mainly calves), there are a
number of substitutes used in cheese making, because
the demand for coagulating enzymes started exceeding
the supply almost 50 years ago (de Koning, 1978).
Commercial animal rennets contain mainly chymosin,
which cleaves j-casein specically at the Phe105Met106
peptide bond, yielding the hydrophobic para-j-casein
and the hydrophilic casein macropeptide (residues 106
169; Hyslop, 2003). The pepsin fraction in calves rennet
(approx. 1020%) varies with the age of the animals;
*Correspondent: E-mail: doris.jaros@tu-dresden.de
doi:10.1111/j.1365-2621.2008.01749.x
adult bovine rennet has a much higher pepsin content,

usually 8090% (Harboe & Budtz, 1999). Prerequisite
for rennet substitutes is a weak general proteolytic
activity and a narrow specicity. Aspartic proteinases
used as rennet substitutes are, apart from coagulants of
plant origin (i.e. from Cynara cardunculus L. and Cynara
humilis L.), applied in artisanal protected denomination
of origin (PDO) cheese production (Sousa & Malcata,
1996; Freitas et al., 2000; Esteves et al., 2003; Silva &
Malcata, 2005), enzymes excreted by Rhizomucor mihei,
Rhizomucor pusillus and Cryphonectria parasitica. It is
generally acknowledged that fungal proteinases show a
pH optimum similar to that of chymosin, whereas their
temperature optimum is signicantly higher [40 C vs.
approx. 60 C for chymosin and microbial coagulants,
respectively (Hyslop, 2003)]. Nowadays, most of the
commercially available microbial coagulants are destabilised by oxidation, thus showing reduced heat stability
and non-specic proteolytic activity (Harboe & Budtz,
1999). Finally, recombinant chymosin, which is identical
to bovine chymosin as concerns its amino acid sequence,
is synthesised by genetically modied organisms (e.g.
Aspergillus niger or Kluyveromyces lactis) and available
since approx. 15 years.
When producing cheeses with specic designations
(e.g. organic labels) in central Europe, cheese makers are
forced to use either animal rennet or coagulants of
fungal origin (fermentation-produced chymosin is not
allowed, and plant coagulants are not commercially
available). When comparing the action of dierent types
of coagulants, one aspect of economic importance is
obviously cheese yield. In model experiments with
Cheddar cheese, Emmons (1990) and Emmons et al.
(1990) observed, on the basis of nitrogen balances, a
lower cheese yield when fungal coagulants were used.
Ustunol & Hicks (1990) found a signicantly lower yield
when cheeses were manufactured with C. parasitica
1721
1722
Enzymatic coagulation of milk D. Jaros et al.
protease instead of using calf rennet. The decreased

cheese yield is presumably related to some unspecic
proteolysis, as has been demonstrated by Vanderporten
& Weckx (1972), Shammet et al. (1992) and Krause
et al. (1998, 2001).
There is, in our opinion, a signicant lack as concerns
the knowledge of the gelation behaviour of calf rennets
and destabilised microbial coagulants. To establish a
methodical basis for further work on dierent enzymes,
we analysed the gel formation after enzymatic clotting
of milk induced by dierent coagulants and systematically varied temperature and pH in a range relevant to
the production of hard and semi-hard cheeses.
Base milk and enzymes for gelation experiments
Reconstituted skim milk (12% dry matter) was prepared

from low-heat skim milk powder (moisture content:
3.7%; Alpavit Kaserei Champignon Hofmeister GmbH
& Co. KG, Lauben, Germany) and distilled water. After
dissolution, 0.3% sodium acide was added to the milk
which was then stored in a cold room overnight. Prior to
usage, the milk was warmed to room temperature and
adjusted to a particular pH by adding 10% (v v) lactic
acid dropwise.
Two calf rennets and two microbial enzyme preparations from R. mihei, with clotting activities determined
according to International Dairy Federation (IDF)Standard 157A (Anonymous, 1997), were used (Table 1).
After diluting each rennet solution to 24 International
Milk-clotting Units (IMCU) mL)1 with distilled water,
one volume unit was added to 400 volumes of milk to
induce gelation.
ature using a DC8 thermocontroller (Thermo Haake

GmbH, Karlsruhe, Germany) 50 lL enzyme solution
was added, and the milk was transferred into the prethermostatted cup of the rheometer. Measurements were
started within 1 min after enzyme addition. During
gelation, a strain of 0.003 low enough to ensure
undisturbed gelation (Karlsson et al., 2007b; Sandra &
Dalgleish, 2007) was applied at an angular frequency
of 1 rad s)1. Storage modulus G (Pa), loss modulus G
(Pa) and the loss tangent tan d ()) were monitored as a
function of time.
Experimental setup
A central composite design was established to determine

pH and temperature eects on gelation. The independent variables were coded at ve levels ()20.5, )1, 0, +1,
+20.5), with the central point referring to pH0 = 6.4
and T0 = 34 C (Table 2). Using a 24 IMCU mL)1
enzyme preparation and inducing gelation by adding
one volume unit of enzyme to 400 units of milk means
that the relative enzyme concentration was at the upper
level of a suppliers dosage recommendations. One
experiment was carried out at each star point, and the
number of repetitions at the central point were three
(Kleppmann, 2003).
Statistical evaluation
A commercial software package (SAS Institute, 2001)

was used to t second-order models and to generate
response surface plots. After rescaling the input variables to their normalised values pHN or TN, a model of
the following form was applied.
Yi b0 b1 pHN b2 TN b11 pHN2 b22 TN2
b12 pHN TN
Rheological measurements
Gelation proles were monitored using a strain-controlled ARES RFS3 rheometer (TA Instruments, Eschborn, Germany) with a cup-and-bob-geometry (inner
diameter: 32 mm; outer diameter: 34 mm; bob length:
33.5 mm), and temperature was maintained with an
accuracy of 0.1 C by a computer-controlled circulator.
After thermostatting 20 mL milk to the desired temperTable 1 Coagulants used in the study
Yi is the response variable, b0 is an equation constant, b1,

b2, b11 and b22 are linear and quadratic regression terms,
and b12 is the cross-product regression term. After
estimating the regression coecients of the model by the
general linear model (GLM) procedure, equation terms
with insignicant F-values were eliminated stepwise
until only signicant (P < 0.05) factors remained in
the model.
Table 2 Variables, nominal and coded levels in the gelation experi-
ments
Product code
Origin
Clotting activity
(IMCU)
Chymosin:
pepsin (ratio)
Sample
Sample
Sample
Sample
Microbial (type TL)

Microbial (type TL)
Animal (calf)
Animal (calf)
285
262
186
361
100:0
100:0
80:20
94:6
A
B
C
D
pH ())
T (C)
Normalised value
6.26
6.30
6.40
6.50
6.54
31.2
32.0
34.0
36.0
36.8
)1.414
)1
0
1
1.414
Chymosin:pepsin ratios were provided by the manufacturers.
gel formation; and that (ii) G40 refers to a late event,

the process of gel assembly.
Effects of enzyme concentration on the gelation profiles
Figure 1 depicts the results of the replicate gelation

experiments for sample B (microbial coagulant) and
sample C (animal rennet). Generally, the gelation curves
resemble those given in literature (e.g. Lomholt & Qvist,
1999). Based on several suggestions (Horne, 1998, 2003;
Ste et al., 1999), we decided to extract the following
parameters from the individual curves for further
analysis: (i) the gelation time, tg (s), refers to the time
needed for G to exceed 1 Pa, a value above the noise
level of the rheometer; (ii) the increase of G per unit
time in the linear part (i.e. usually between 30 and
60 Pa) of the gelation curve was denoted as gelation
rate, rg (Pa s)1), which can be regarded as an equivalent
of the rming rate obtained from traditional formagraph measurements (Daviau et al., 2000); (iii) the setto-cut time t20 (s) corresponds to the time needed to
reach a G of 20 Pa; and (iv) G40 (Pa) is gel rmness
40 min after enzyme addition. The averaged variation
coecients were 4.3% and 2.4% for the time-based
measures tg and t20, respectively, and 2.9% and 2.5%
for the parameters related to rmness (rg and G40,
respectively).
The interrelationship between the parameters
extracted from the gelation curves is shown in
Fig. 2. Although all linear correlation coecients were
signicant (P < 0.001), the data point distribution
indicates that one of the time-based measures (tg or
t20) is clearly redundant, as is either rg or G40 for the
rmness-based parameters. We therefore decided to
interpret our further results in terms of gelation time
(tg) and gel rmness (G40) to achieve as much
information as possible despite reducing the number
of parameters. The rationale behind this decision was
that: (i) tg refers to an event at the very beginning of
Effects of coagulant type, pH and temperature
Figure 3 shows the gelation proles of milk as aected

by coagulant, pH and temperature. There is a general
trend that, for a particular temperature, gelation time
decreases and rmness increases with decreasing pH.
The boundary values of pH = 6.55 and pH = 6.25
apply when either a direct starter or a few percentage of
a stock culture is added to the base milk, respectively. It
is also evident that, by decreasing coagulation temperature at a particular pH, there is a pronounced decrease
in gel rmness and a tendency towards reduced gelation
time. Finally, there is a trend that the systems coagulated with animal rennet (samples C and D) provided
rmer gels within a shorter period of time, especially
when considering conditions typical for the manufacture
of traditional Swiss-type cheese (pH and temperature on
the lower end of the applied range).
By using animal rennet at 33 C, Daviau et al. (2000)
found a decrease in gelation time by the factor of 2 for
6.5 > pH > 6.2, but no further decrease below
pH = 6.2 and a twofold increase in rmness. For
suciently aged rennet gels, Zoon et al. (1989) showed
that rmness increased linearly on reducing pH from
approx. 6.7 to 6.2, but then decreased on further
lowering of the pH. Esteves et al. (2003) found a vefold decrease in gelation time when lowering pH from
6.7 to 6.0 at T = 32 C. However, the dierences in gel
rmness attributable to pH were much lower than in our
study. Najera et al. (2003) varied acidity, temperature
and calcium concentration, and found that, with
decreasing pH, coagulation time decreased and gel
rmness increased, and by decreasing temperature from
44 to 28 C, coagulation time slightly decreased but gel
rmness increased.
Figure 1 Development of storage modulus

(G), loss modulus (G) and loss tangent (tan
d) in skim-milk gels induced by a microbial
coagulant (sample B) and an animal rennet
(sample C). The three replicate experiments
were performed at T = 34 C with milk
adjusted to pH = 6.4.
1723
1724
Figure 2 Inter-relations between gelation

time tg, set-to-cut time t20, gelation rate rg and
gel rmness G40 in gels coagulated with
dierent enzymes at dierent conditions.
Along with the coecients of determination, the

signicant (P < 0.05) regression coecients giving
information on the dependency of gelation time or gel
rmness on acidity and temperature of the milk are
summarised in Table 3. For both animal rennets, tg and
G40 showed a signicant linear dependency on pH and
temperature. As regards the microbial coagulants,
additional quadratic and interaction terms were observed. After performing the statistical analyses, the
coecients were rescaled to basic units of 1 C and 1
pH. Hence, b0 given in Table 3 refers to the nominal
value of tg or G40 of gels made from milk of 34 C and
pH 6.4; b1 and b2 reect the increase or decrease per unit
pH and temperature, respectively, and the same applies
for b11, b22 and b12.
The contour plots (Fig. 4) illustrate the dependency of
the gelation time and gel rmness on pH and temperature. Generally, the time for the onset of gelation
decreased with decreasing pH, but increased with
decreasing temperature (positive and negative coecients, respectively), and the quadratic terms are responsible for the curvilinear contours. The tg values
calculated for highest pH and T (6.54 36.8 C) are
1144 and 1225 s for coagulants A and B, respectively,
but 1165 and 1152 s for rennets C and D, respectively.
Decreasing T to 31.2 C at this particular pH resulted in

a tg of 1366 s (A), 1517 s (B), 1301 s (C) and 1374 s (D).
At pH = 6.26 and T = 31.2 C, gelation times were
695, 754, 608 and 614 s for coagulants A, B, C and D,
respectively, and the corresponding times at low pH
(6.26) but high temperature (36.8 C) were 472, 461, 472
and 392 s. Especially the conditions in the third quadrant (low pH and low temperature) resemble those
applied for a wide variety of semi-hard or hard cheeses
(Hyslop, 2003; Kammerlehner, 2003).
As regards gel rmness, there is a positive, and in case
of animal rennets, linear dependency on temperature
(Fig. 4) and rmness also increases linearly with
decreasing pH. Quadratic and interaction terms were
found to be signicant for coagulants A and B. By using
the coecients from Table 3, G40 calculated for
T = 32 C and pH = 6.3 (third quadrant) is 82.5 and
76.3 Pa for milk coagulated with microbial coagulants A
and B, respectively. When using animal rennets, the
estimated rmness values are 103.6 Pa (sample C) and
105.9 Pa (sample D).
Analysis of variance and multiple t-tests were applied
to the results of the replicate central experiments
(T = 34 C, pH = 6.4). For these particular conditions, gelation time for samples A and B (microbial
Figure 3 Gelation proles as a function of

clotting temperature and milk pH. Full lines,
microbial coagulants A and B; broken lines,
animal rennets C and D.
Table 3 Signicant coecients (P < 0.05) of the regression models for
the individual coagulants
origin) is signicantly lower than for samples C and D

(animal rennets). Gel rmness is signicantly higher
when milk was clotted with animal rennets (Table 4).
Coagulant code
Conclusion
Parameter
Gelation time tg (s)

775.4
b0
2400
b1
b2
)39.72
3520
b11
9.561
b22
b12
0.987
R2
Gel firmness G40 (Pa)
75.41
b0
b1
)336.0
6.682
b2
314.3
b11
)2.173
b22
b12
)37.87
0.989
R2
803.2
2728
)52.20
4213
13.21
0.995
887.1
2475
)24.19
0.981
883.5
2715
)39.70
70.22
)363.6
8.596
)1.676
)31.75
0.982
82.76
)346.6
6.898
0.992
81.03
)389.1
7.014
0.980
0.966
b0 values represent gelation time or gel firmness at pH = 6.4 and

T = 34 C. b1 is associated with pH and b2 is associated with T. R2 is the
coefficient of determination. Regression model, see eqn. 1.
It is generally accepted that clotting temperatures of

approx. 3035 C produce gels with the highest rmness
(Lucey, 2002). The results of our study clearly show
that, although working in a small temperature range, the
gel rmness is aected to a large extent. Depending on
the pH of the milk, gel rmness (G after 40 min) varied
between approx. 25 and 140 Pa. Especially with respect
to temperature changes between 34 C and 31 C,
animal rennets are more stable than microbial coagulants: it is evident from the contour plots in Fig. 4 that
the corresponding dierences in gel rmness are approx.
25% and 5060% when using animal rennets and
microbial coagulants, respectively. A similar impact on
gelation time was observed; although tg at central point
conditions was approx. 10% higher when using animal
rennet, this eect is compensated when gelation temperature is decreased. An increased gel rmness at a
xed time after enzyme addition obviously implies that
comparable rmness (e.g. for starting curd cutting) may
1725
1726
Table 4 Gelation time and rmness of the milk gels at pH = 6.4 and
T = 34 C
Coagulant code
Gelation time tg(s)
Gel firmness G40 (Pa)
A
B
C
D
775a
804a
875b
874b
73.9a
70.4a
84.8b
79.3b
24.2
18.4
27.7
16.0
1.19
2.46
1.69
1.88
Column values with different superscripts differ significantly at P < 0.05.

Mean values and standard deviations are based on three replicate
experiments.
References
Figure 4 Contour plots showing the dependency of gelation time (tg)

and gel rmness (G40) on clotting temperature and milk pH. Letters
refer to coagulant code, and numbers in the contour plots refer to tg in
s (left column) or G40 in Pa (right column).
be reached in a shorter period of time, or on the other

hand, lower amounts of animal rennets are required to
obtain similar gel rmness within the same time frame.
The analysis of gel formation by dynamic rheometry
gives a solid basis for further investigations of the
impact of coagulant type on cheese yield and quality.
Anonymous (1997). Bovine Rennets. Determination of Total Milk

Clotting Activity. FIL IDF-Standard 157A. Brussels, Belgium:
International Dairy Federation.
Daviau, C., Famelart, M.H., Pierre, A., Goudedranche, H. &
Maubois, J.L. (2000). Rennet coagulation of skim milk and curd
drainage: eect of pH, casein concentration, ionic strength and heat
treatment. Le Lait, 80, 397415.
Emmons, D.B. (1990). Milk clotting enzymes. 2. Estimating cheese
yield losses from proteolysis during cheese making. Journal of Dairy
Science, 73, 20162021.
Emmons, D.B., Beckett, D.C. & Binns, M. (1990). Milk clotting
enzymes. 1. Proteolysis during cheese making in relation to
estimated losses of yield. Journal of Dairy Science, 73, 20072015.
Esteves, C.L.C., Lucey, J.A., Wang, T. & Pires, E.M.V. (2003). Eect
of pH on the gelation properties of skim milk gels made from plant
coagulants and chymosin. Journal of Dairy Science, 86, 25582567.
Freitas, A.C., Macedo, A.C. & Malcata, F.X. (2000). Review:
technological and organoleptic issues pertaining to cheeses with
denomination of origin manufactured in the Iberian Peninsula from
ovine and caprine milks. Food Science and Technology International,
6, 351370.
Harboe, M. & Budtz, P. (1999). The production, action and application of rennet and coagulants. In: Technology of Cheesemaking
(edited by B.A. Law). Pp. 3365. Sheeld, UK: Sheeld Academic
Press.
Holt, C. & Horne, D.S. (1996). The hairy casein micelle: evolution of
the concept and its implications for dairy technology. Netherlands
Milk and Dairy Journal, 50, 85111.
Horne, D.S. (1998). Casein interactions: casting light on the black
boxes, the structure in dairy products. International Dairy Journal, 8,
171177.
Horne, D.S. (2003). Casein micelles as hard spheres: limitations of the
model in acidied gel formation. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 213, 255263.
Hyslop, D.B. (2003). Enzymatic coagulation of milk. In: Advanced
Dairy Chemistry. Volume 1. Proteins (edited by P.F. Fox & P.L.H.
McSweeney). Pp. 839878. New York, NY: Kluwer Academic Plenum Publishers.
Johnson, M.E., Chen, C.M. & Jaeggi, J.J. (2001). Eect of rennet
coagulation time on composition, yield, and quality of reduced-fat
cheddar cheese. Journal of Dairy Science, 84, 10271033.
Kammerlehner, J. (2003). Kasetechnologie. Freising, Germany: Freie
Kunstlerpresse.
Karlsson, A.O., Ipsen, R. & Ardo, Y. (2007a). Inuence of pH and
NaCl on rheological properties of rennet-induced casein gels made
from UF concentrated skim milk. International Dairy Journal, 17,
10531062.
Karlsson, A.O., Ipsen, R. & Ardo, Y. (2007b). Rheological properties
and microstructure during rennet induced coagulation of UF
concentrated skim milk. International Dairy Journal, 17, 674682.
Kleppmann, W. (2003). Taschenbuch Versuchsplanung: Produkte und

Prozesse optimieren. Munchen, Germany: Carl Hanser Verlag.
de Koning, P.J. (1978). Coagulating enzymes in cheesemaking. Dairy
Industries International, 43, 712.
Krause, W., Partzsch, M., Hassan, Z.M.R. & Haufe, T. (1998).
Substrate and binding specity of aspartic proteases with milk
clotting properties. Nahrung Food, 42, 162165.
Krause, W., Partzsch, M. & Hugel, S. (2001). Hydrolysis of isolated
beta-casein with chymosin and other rennet enzymes. 1. Characterization of terminal amino acids of the peptides in beta-casein (CN)and non casein (NCN)-fractions. Milchwissenschaft, 56, 146149.
Lomholt, S.B. & Qvist, K.B. (1999). The formation of cheese curd. In:
Technology of Cheesemaking (edited by B.A. Law). Pp. 6698.
Sheeld, U.K: Sheeld Academic Press.
Lucey, J.A. (2002). Formation and physical properties of milk protein
gels. Journal of Dairy Science, 85, 281294.
Najera, A.I., de Renobales, M. & Barron, L.J.R. (2003). Eects of pH,
temperature, CaCl2 and enzyme concentrations on the rennetclotting properties of milk: a multifactorial study. Food Chemistry,
80, 345352.
OCallaghan, D.J., ODonnell, C.P. & Payne, F.A. (2000). On-line
sensing techniques for coagulum setting in renneted milks. Journal of
Food Engineering, 43, 155165.
Riddell-Lawrence, S. & Hicks, C.L. (1989). Eect of curd rmness on
stirred curd cheese yield. Journal of Dairy Science, 72, 313321.
Sandra, S. & Dalgleish, D.G. (2007). The eect of ultra high-pressure
homogenization (UHPH) on rennet coagulation properties of
unheated and heated fresh skimmed milk. International Dairy
Journal, 17, 10431052.
SAS Institute (2001). SAS Stat guide for personal computers. 8.02 edn.
Cary, USA: SAS Institute.
Shammet, K.M., Brown, R.J. & McMahon, D.J. (1992). Proteolytic
activity of some milk-clotting enzymes on k-casein. Journal of Dairy
Science, 75, 13731379.
Silva, S.V. & Malcata, F.X. (2005). Studies pertaining to coagulant
and proteolytic activities of plant proteases from Cynara cardunculus. Food Chemistry, 89, 1926.
Sousa, M.J. & Malcata, F.X. (1996). Eects of processing conditions
on the caseinolytic activity of crude extracts of Cyanara cardunculus
L. Food Science and Technology International, 2, 255263.
Ste, A., Schreiber, R., Hafenmair, M. & Kessler, H.G. (1999). Eect
of denatured whey proteins on the rennet-induced aggregation of
casein micelles. International Dairy Journal, 9, 401402.
Ustunol, Z. & Hicks, C.L. (1990). Eect of calcium addition on cheese
yield manufactured with Endothia parasitica protease. Journal of
Dairy Science, 73, 1725.
Vanderporten, R. & Weckx, M. (1972). Breakdown of casein by rennet
and microbial milk-clotting enzymes. Netherlands Milk and Dairy
Journal, 26, 4759.
Walstra, P. (1990). On the stability of casein micelles. Journal of Dairy
Science, 73, 19651979.
Walstra, P., Wouters, J.T.M. & Geurts, T.J. (2006). Dairy Science and
Technology. Boca Raton, FL: CRC Press.
Zoon, N., van Vliet, T. & Walsta, P. (1989). Rheological properties of
rennet-induced skim milk gels. 4. The eect of pH and NaCl.
Netherlands Milk and Dairy Journal, 42, 249269.
1727

Get File PDF

Загружено:

Сведения о документе

Исходное описание:

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Get File PDF

Загружено:

Авторское право:

Доступные форматы

International Journal of Food Science & Technology

Volume 43 ,Issue 9 , Pages.1507 - 1727 (September 2008)

Probiotic foods: consumer perception and attitudes (p 1577-1580)

International Journal of Food Science and Technology 2008, 43, 15071511

Eating and drinking should be pleasurable. The sensory

and how to join, can be found at http://www.IFST.

Gamlath investigated a novel use for over-ripe

In the area of quality:

In the areas of shelf-life and storage:

International Journal of Food Science and Technology 2008

Freeze-thaw cycles were found to have a detrimental

In the area of nutrition and health:

It is not possible to review all the papers in detail here,

orange avours and provided more information, in that

performing consumer studies in order to establish

International Journal of Food Science and Technology 2008

Short- and long-term health effects on

International Journal of Food Science and Technology 2008

Replication should be used where appropriate, such

Sensory evaluation is a growing, dynamic eld. It

physico-chemistry, psychophysics, psychology, physiology, neuroscience and genomics.

Galmarini, M.V., Zamora, M.C., Baby, R., Chirife, J. & Mesina, V.

International Journal of Food Science and Technology 2008

International Journal of Food Science and Technology 2008, 43, 15121519

Amino acids, aroma, Maillard reaction, glucose.

The parameters that inuence the overall avours and

sugar will yield hundreds of volatile compounds

l-alanine, l-arginine hydrochloride, l-aspartic acid,

Glucose-amino acids Maillard reaction aromas K. H. Wong et al.

hydroxide pellets of analytical grade were obtained from

d-glucose monohydrate (10.0 mmol in 20 mL distilled

These combinations were based on the amino acid

With cysteine and methionine

Amount of amino acids

Amount of amino acids

2008 Institute of Food Science and Technology

International Journal of Food Science and Technology 2008, 43, 15121519

Glucose-amino acids Maillard reaction aromas K. H. Wong et al.

6 Caramel: burned sugar.

Colour references were prepared for the samples of

International Journal of Food Science and Technology 2008, 43, 15121519

Results and discussion

Flavour notes of individual amino acids

Temperature, substrates ratio, pH and time may aect

2008 Institute of Food Science and Technology

Glucose-amino acids Maillard reaction aromas K. H. Wong et al.

Table 2 Colour and odour of Maillard

Fruity** (Persimmon), pleasant/sweet**,

Ciku Manilkara achras mill.; pandan Pandanus odorus, Ridl.

Table 3 Colour and odour of Maillard

Fruity** (ciku, fresh dates), pleasant/sweet**, flowery**

Ciku Manilkara achras mill.; pandan Pandanus odorus, Ridl.

2008 Institute of Food Science and Technology

International Journal of Food Science and Technology 2008, 43, 15121519

Glucose-amino acids Maillard reaction aromas K. H. Wong et al.

Pleasant/sweet**, fruity* (ciku, fresh dates), caramel-like**,

Table 4 Colour and odour of Maillard

During the Strecker degradation, sulphur-containing

International Journal of Food Science and Technology 2008, 43, 15121519

reaction medium, and that the panellists were more

At 6 m HCl (which is the concentration of HCl used in

2008 Institute of Food Science and Technology

Fruity (Persimmon), pleasant/sweet,

Fruity (ciku, fresh dates), pleasant/sweet, flowery**

Pleasant/sweet, caramel-like, fruity, sour

Pleasant/sweet**, fruity, sour

Soy sauce**, marmite, salted fish-like