Journal of Cereal Science 32 (2000) 7387

doi:10.1006/jcrs.2000.0320, available online at http://www.idealibrary.com on

The Effects of Bread Making Process and Wheat

Quality on French Baguettes
P. Baardseth, K. Kvaal, P. Lea, M. R. Ellekjr and E. M. Frgestad
MATFORSK, Norwegian Food Research Institute, Osloveien 1, N-1430 As, Norway
Received 10 June 1999

ABSTRACT
The quality of baguettes can be evaluated by defined sensory attributes and image analyses. The
eect of flour quality, production process (traditional French and industrially modified), mixing and
proofing time were studied. Process accounted for 40% of the variation in baguette quality whereas
flour quality accounted for 16% of the variation when principal component analysis was applied on
the sensory attributes. Baguettes produced using a soft dough and gentle treatment (traditional French
process) had a higher sensory score for porosity, elasticity, crispness of crust, crackles on the crust,
and porosity and volume as measured by image analysis, than baguettes produced using a sti dough
and rough treatment (modified industrial process). Mixing and proofing time also aected the porosity
and area of the cut surface. Porosity, crackles on the crust, glossiness and volume were related to
flour quality.
2000 Academic Press

Keywords: sensory analysis, image analysis, French baguette quality.

INTRODUCTION
French bread (baguettes or pain Parisian) is typically characterised by a crisp eggshell crust
34 mm thick, an open and random crumb cell
structure, a full-bodied flavour, a high specific
volume (mL/g) and a relatively high crust:crumb
ratio due to the diameter and the length of the
loaves1,2.
: ICC=International Association
for Cereal Science and Technology; ISO=International Association for Standardisation; AACC=
American Association for Cereal Chemists; SDSPAGE=sodium dodecyl sulphate polyacrylamide gel
electrophoresis; HMW=high molecular weight; WA=
water absorption; DU=dough development; DS=
dough stability; BU=Brabender unit; ANOVA=analysis of variance; PCA=principal component analysis;
PC=principal component; PCR=principal component regression; AMT=angle measure technique;
P1=traditional French process; P2=modified industrially process.
Corresponding author: P. Baardseth.
07335210/00/070073+15 $35.00/0

An important aspect of traditional French bread

production is the adjustment of the flour:water
ratio to give soft dough1,2. Mechanical dough handling in industrial processing, however, requires a
stier dough to avoid adhesion of dough to the
machinery.
To obtain the characteristic pore structure of
the crumb of a typical French bread, the dough
should be fully expanded with numerous large,
gas-filled bubbles at the end of the mixing and
fermentation process1,2. In general, the mixing
time necessary to obtain optimum dough development is dependent on flour quality and as
well as the mixing equipment, mixing intensity,
dough recipe, etc. Magnus3 found that the optimum fermentation time with respect to loaf volume, form ratio, loaf weight, pore size, firmness
and colour of small hearth loaves also varied
widely amongst wheats of dierent qualities, and
according to the mixing process used. In addition
to producing a light dough capable of retaining
gas produced by the yeast, the fermentation process also leads to the production of aromatic
2000 Academic Press

74

P. Baardseth et al.

substances that contribute to the taste of the baguettes4,5. However, there are no published studies
on the eect of mixing and fermentation time
on volume, pore structure and texture of French
bread.
Critical phases of French bread production follow fermentation. In this phase it is important to
avoid degassing the dough if an open and random
pore structure in the crumb is to be achieved1,2.
This is unlike most other bread baking processes,
which aim to obtain an even, fine pore structure
in the crumb. A further increase in gas and consequent enhancement of bubble size occurs during
the proofing stages which is the time from dough
dividing to baking in the oven. Aromatic substances also build up during the proofing stage46.
The cutting of the dough surface after the final
proof is an integral part of processing of French
bread and leads to a characteristic product. Cutting releases stresses in the dough during baking,
i.e. increases crust area during dough expansion
and in the bake out, gives an attractive appearance
to the loaf and improves the flavour7.
In determining the baking potential of flour,
both protein content and protein quality are important. Although protein quality is complex, it is
known that a major factor in determining the
protein quality is the composite of the high molecular weight (HMW) glutenin subunits8,9.
The present study focuses on how the baking
process and flour quality aect the characteristics
of French baguettes produced by traditional and
industrially modified methods. The following factors were investigated using a full factorial design:
flour quality, a modified industrial process versus
a traditional French process, mixing time and
proofing time. Baguette quality was assessed by
sensory and image analysis.
MATERIAL AND METHODS
Flour
Four wheats of dierent qualities were used: Tjalve
(a Norwegian spring wheat), Folke (a Norwegian
winter wheat) and two commercial baguette flours
(A: 100% French soft wheat; B: 20% American
wheat, 25% French wheat, 20% Norwegian spring
wheat, 35% Norwegian winter wheat). The wheats
were milled on a commercial mill (Simon, England
with Miag rollers). Extraction rates were 757%
for Tjalve and 775% for Folke. Ascorbic acid
(30 ppm) was added to the flour immediately after

milling. The two Norwegian flours were from the

same wheats as used by Frgestad et al.10 The
commercial baguette flours also contained 30 ppm
ascorbic acid.
Protein content was analysed by Kjeldahl N
(ICC Standard no 105, N57, presented on a
dry weight basis). Water absorption and reological
properties of the flours were determined using a
Farinograph11 and Mixograph12 using water addition according to the ISO standard11. The Zeleny
sedimentation test was performed according to the
AACC procedure13. HMW glutenin subunits were
determined by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE)14.
Tjalve flour and baguette flour B had similar
protein contents, 127% and 121%, respectively,
and both flours contained the HMW glutenin
subunits 5+10 that are correlated with strong
protein quality15 (Table I). Baguette flour B also
contained the HMW glutenin subunits 2+12 and
in addition subunits 5+10. Baguette flour A and
Folke flour had low protein contents (102% and
90%, respectively), and both contained HMW
glutenin subunits 2+12, associated with poor protein quality. The sedimentation volume and mixograph peak times were higher for the baguette
flour B and Tjalve flour compared with the baguette flour A and Folke flour (Table I).
Baguette flour A from soft wheat had a lower
water binding capacity than the other flours, due
to the lower damaged starch1,2.
Experimental design
A full factorial experimental design16 without replicates was set up and 32 baguette productions were
performed in random order. The main experiment
consisted of two factors, namely the four flour
qualities and the industrial versus the traditional
process. The two other factors, mixing and
proofing time were nested within each process,
both at two levels.
Experimental baking
Figure 1 shows scheme for the traditional French
baking process (P1) and the modified industrial
process (P2).
The flour temperature (8 C), water temperature
(12 C for 2+4 min mixing times, 04 C for
2+8 min mixing times) and completion of mixing
dough temperature (24 C) were recorded together

Baguette making and wheat quality

Table I

75

Characteristics of the wheat flours

Farinogram
Mixograph
Parameters

Id.
No
1
2
3
4

Wheat variety
%
flour source protein
Tjalve
Folke
Baguette flour
A
Baguette flour
B

HMW glutenin
subunits

Zeleny
WA DU
sedimentation (%) (BU)

DS Peak
(BU) time
(min)

127
90
102

2, 7+9, 5+10
2, 6+8, 2+12
2, 7+8, 2+12

48
28
12

624
609
560

25
17
17

60 536
110 262
70 280

121

2, 7+9, 6+8, 5+10,

2+12

39

606

20

80 564

WA=water absorption, DU=dough development, DS=dough stability, BU=Brabender unit.

Process 1
Traditional French
process

Process 2
Industrial modified
process

Recipe
0.9% yeast
1% salt
soft dough conditions 500 BU6%
0.8 ascorbic acid

Recipe
0.3% yeast
1% salt
stiff dough conditions 500 BU +5%
0.8 ascorbic acid

Mixing time:
2 + 4 min (125 wh)
2 + 8 min (300 wh)

Mixing time:
2 + 4 min (180 wh)
2 + 8 min (400 wh)

20 min fermentation

Gentle dividing and moulding

by hand

Rough dividing
treatment

15 min pre-proofing

Figure 1
2).

Gentle sheeting of the dough

to final form

Moulding and sheeting

to final form

Proofing time:
65 min
80 min

Proofing time:
150 min
165 min

Baking
270C
220C
10 min

Baking
250C
220C
10 min

Flow sheet of the two processestraditional French process (process 1) and modified industrial process (process

76

P. Baardseth et al.

Table II

Average energy input (Wh) used for doughs from four flours and two processes
Mixing
time

Tjalve

Folke

Baguette flour A

Baguette flour B

Process 1
(soft dough)

2+4 min
2+8 min

110
290

115
255

145
355

130
280

Process 2
(sti dough)

2+4 min
2+8 min

160
385

160
360

215
440

180
440

Table III

Definition of the sensory attributes

Sensory attributes
Glossiness
Crackles on the crust

Porosity
Area of cut surface
Elasticity
Odour intensity
Fresh odour
Flavour intensity
Fresh flavour
Salt flavour
Firmness
Moistness
Crispness of the crust

Definition
Glossiness on the crust surface
No intensity=non-glossy (matt) on the crust
High intensity=high glossiness on the crust
Crackle formation on the crust (after baking a fine network of cracks appears on the crust
like in pottery)
No intensity=no crackles, smooth, even crust
High intensity=crackles and uneven crust
Pore structure in the crumb measured using Dallmanns pore table19
No intensity=dense pore structure
High intensity=open and random pore structure
Cuts on the crust so the dough can expand to increase the crust area of the bread
No intensity=little expansion of the crumb
High intensity=high expansion of the crumb
Slices able to retain the shape after squeezing
No intensity=no elasticity, slices do not retain shape after squeezing
High intensity=high elasticity, slices retain shape after squeezing
Total odour intensity of the sample
No intensity=no odour
High intensity=strong odour
Fresh odour
No intensity=no fresh odour
High intensity=distinct fresh odour
The strength of total flavour in the sample
No intensity=no flavour
High intensity=strong flavour
Fresh flavour
No intensity=no fresh flavour
High intensity=distinct flavour
Related to the flavour of sodium chloride
No intensity=no salty flavour
High intensity=distinct salty flavour
Relates to the force needed to bite through the crust and the crumb
No intensity=little force needed to bite through
High intensity=high force needed to bite through
Fluid feeling in the mouth after 3 to 4 bites
No intensity=no moistness, no fluidity after 3 to 4 bites
High intensity=distinct moistness, much fluidity after 3 to 4 bites
Mechanical texture properties related to the ability of samples to retain the shape
No intensity=no crispness, tough
High intensity=distinctly crisp, fragile

with the energy input (Table II). The doughs were

mixed in a Bearmixer (Varimixer Programmable
Control System MK III, Wodschow & Co, Copen-

hagen, Denmark). The doughs were proofed at

25 C and 70% RH in a proofing cabinet (Lillnord
A/S, Odder Denmark). The baguettes were pre-

Baguette making and wheat quality

baked for 10 min in a rotating hearth oven

equipped with a fan (Bago-line Type BEX 1.0,
Fjellbroen A/S, Faaborg, Denmark). Live steam
was injected during the first 35 s of baking, and
the temperature was reduced from 270 C (P1)
or 250 C (P2) to 220 C, immediately after the
baguettes were put in the oven. The baguettes
were packed in a modified atmosphere (80% N2
and 20% CO2) and stored at room temperature
(20 C) for a maximum of 10 days prior to sensory
and image analysis.
Preparation of samples for sensory and image
analyses
Pre-baked baguettes were finally baked (220 C,
11 min and live steam during the first 15 s of
baking) and left for 30 min at room temperature
prior to testing. Each panellist was given half a
baguette on a white dish. The other half was used
for textural analysis.
Sensory evaluation
Baguettes were analysed by conventional sensory
profiling using 10 trained panellists. Members were
selected and trained according to guidelines in
ISO/DIS 85861:198917 and the method for sensory profiling according to ISO 6564:1985-E18.
Panellists developed a lexicon by describing
dierences between extreme samples and developed a consensus list of 13 sensory attributes
for profiling. These were glossiness, crackles on
the crust, porosity (grain) scored according to
Dallmanns pore table19, area of cut surface, elasticity, odour intensity, fresh odour, flavour intensity, fresh flavour, salt flavour, firmness,
moistness and crispness of crust. Panellists were
trained in use of definitions of sensory attributes
(Table III) and rating anchors by pre-testing extreme samples. A continuous non-structured scale
was used for evaluation. The left side of the scale
corresponded to the lowest intensity (value 1.0)
and the right side corresponded to the highest
intensity (value 9.0). Evaluations were performed
in laboratory equipment as described in ISO 8589:
1988-E20. Each panellist evaluated the samples at
their own pace using a computerised system for
direct recording of data (CSA Compusense, version 4.2, Canada). Samples were served in a randomised order. Five to six samples were served
each session (two sessions per day). No replicates

77

were performed. Between tasting, panellists

cleaned their mouth with tap water (20 C) and
salt free crackers.
Image analysis
Images of the baguettes were produced using a
Canon EX2 video camera. The images were recorded with the Screen Machine II frame grabber.
The signal from the camera was composed of
luminance and chrominance signals (Y/C or Hi8).
This signal is common in semi-professional cameras and gives a good separation of crominance
and intensity signals. The baguette slices were
illuminated at a 45 angle from both sides using
four tungsten lamps, two on each side. The recorded images were converted from true colour
images (512 by 512 pixels) to 256 level grey scale.
Images were centred using a thresholding technique together with calculations of the centre of
gravity of the image. A central rectangular cutout and a resizing to obtain a 256256 greyscale
image was performed to ensure that no baguette
was smaller than the quadrate. The images were
enhanced with the 33 mask, unsharp, convolution, and the porosity calculated. The area of
the baguette slices was calculated from images by
counting the number of pixels covering the bread
sample.
Statistical analysis
Several ANOVA models were employed to investigate the eects of the dierent factors. First,
an overall model including: Process, Assessor and
Flour was used. Of these, Assessor and all interaction
terms involving the Assessor were regarded as random eects, whereas the remaining eects were
regarded as fixed. Mixing time and Proofing time were
nested within the Process eect. Since these factors
cannot be compared for the two processes. Thus
a given mixing time will act dierently on a soft
dough vs a stier dough as seen by the dierent
mixing energies (Fig. 1). In addition, for proofing
time there were dierent time settings for the two
processes. Separate models for each Process were
employed using the following factors: Assessor,
Flour, Mixing time, Proofing time. For the attributes
with a significant Flour eect, the dierences were
studied in more detail by Tukeys Multiple Comparison Test. As the fixed eects were on two
levels only, it was not necessary to apply for these

78

P. Baardseth et al.

eects. The multivariate Principal component analysis (PCA) that treats all variables simultaneously21, was performed to obtain an overview
of the sensory data using the Unscrambler software
(CAMO A/S, Trondheim, Norway). In PCA, the
information in the data is projected down to a
small number of new variables called principal
components (PCs), which are linear combinations
of the original data. The dierent PCs are orthogonal to each other, and are estimated to give,
in decreasing order, the best description of the
variability in the data. The first few PCs will
contain most of the relevant information in the
data. The PCs are described by loadings for the
variables and scores for the samples. The data are
modelled in terms of significant factors, plus errors
or residuals.
Image analysis using the angle measure technique (AMT)22 has been shown to be feasible for
modelling sensory porosity23. The images of the
baguette slices from 32 doughs presented to three
random panellists giving a total of 96 data sets were
vectorized by AMT and modelled with sensory
attributes and process variables (Y-variables) using
principal component regression (PCR). The feature extraction method of AMT provides a data
vector (X-variables) for each image. Principal component regression (PCR)24 was performed to predict the baking process and sensory property (Yvariables) from image analyses (X-variables). PCR
combines PCA and traditional multiple regression
analyses. First PCA is performed among the Xvariables (the image variables). Thereby, the few
first significant PCs from the PCA analyses are used
as X-variables in a traditional multiple regression
analyses. In this way the multicollineary problem
among the original X-variables is solved. The
reference values were represented by the mean
sensory attributes over all panellists. Each of the
three data sets consisted of 32 images, and the
mean sensory attributes (Y) were the same for each
of the three data sets. Two of the data sets were
used as calibration sets and one was the test set.
Cross validation among the three data sets shows
similar results. We also used a test set validation,
and plotted the designed variable loading and

scores together with the sensory reference values

to graph the relationship between process and
sensory variables22.

RESULTS AND DISCUSSION

PCA of sensory attributes
Principal component analysis (PCA) of the sensory attributes of baguettes showed that 56% of
the variations could be explained by the two
first principal components (PCs), 40% by the
first factor and 16% by the second factor. The
loadings plot of the two firsts PCs [Fig. 2(a)]
showed that texture attributes explained the
main variation among the baguettes tested in
this study. PC1 in the loadings plot separated
firmness and elasticity i.e. internal characteristics.
The corresponding scores plot [Fig. 2(b)] showed
that PC1 separated the two processestraditional
French at the right and industrially modified at
the left. Thus, baguettes made by the industrial
adapted process with rough mixing and handling
and sti dough (P2) were firmer, but less elastic
than baguettes made by the traditional French
process made with soft dough (P1) and gentle
mixing and handling. Baguettes with high intensity of elasticity also had a high porosity,
moistness and low firmness. Fresh odour and
flavour were also associated with the traditional
French process [Fig. 2(a)].
PC2 in the loadings plot separated the external
characteristics, glossiness and area of cut surface.
The cutting of the dough surface after final
proof is an integral part of processing French
bread. The two other external quality characteristics, crispness and crackles on the crust,
were highly correlated with the area of cut
surface. The eects of crust crackles were more
complex, but the scores plot shows flour quality
to be important [Fig. 2(b)]. High glossiness was
obtained on baguettes baked with Tjalve flour
(1), whereas baguettes with a high area of cut
surface were obtained with baguette flour A (3),
from French soft wheat.

Figure 2 Loading (a) and score plots (b) for the two first factors obtained by PCA for sensory attributes of baguettes
formulated from wheat flours of four dierent qualities and made by two dierent processes. The first number is flour quality
1 (Tjalve), 2 (Folke), 3 (Baguette flour A), 4 (Baguette flour B), the second number indicates the mixing time 1 (2+4 min), 2
(2+8 min), the third number indicates the proofing time 1 (65/180 min), 2 (80/165 min) and the fourth number 1 (traditional
process) and 2 (modified industrial process).

Baguette making and wheat quality

79

(a)
0.6
glossiness
0.5
0.4
elasticity
0.3

odour intensity
porosity

PC2

0.2

flavour intensity

moistness
salt flavour

0.1

firmness

0
fresh flavour
fresh odour

0.1
0.2

crispness of the crust

0.3

crackles
on the crust

Area of
cut surface

0.4
0.4

0.3

0.2

0.1

0
PC1

0.1

0.2

0.4

0.3

(b)
4
1121
3
1122

4121

1222
1111 1221

1212 1112

4122

PC2

2122
2222

3122

1211

4211
2111
4221

3111
4112
3221
2221 4222
3211
4212
3121

0
3112
2112

4111

2121

2211

3222
2212

4
3212
5
5

0
PC1

80

P. Baardseth et al.

Process 2

Process 1

Figure 3 Images of slices of baguettes produced with four dierent flours (from top to bottomTjalve (1), Folke (2), baguette
flour A (3) and baguette flour B (4)) and two dierent processes (from left to rightprocess 2, mixing time 2+4 min and
2+8 min at proofing time 150 min, mixing time 2+4 min and 2+8 min at proofing time 165 min, process 1, mixing time
2+4 min and 2+8 min, fermentation time 20 min, pre-proofing time 15 min and proofing time 65 min, mixing time 2+4 min
and 2+8 min, pre-proofing 20 min+15 min and proofing time 80 min, see Fig. 1)

0.10
odour intensity
F2
F3
0.05

MT4

crackles
on the crust
P1

0
PC2

crispness of the crust

elasticity
porosity

0.05

area of
cut surface

salt flavour
F1

PT2

PT1

firmness

flavour intensity

P2

glossiness

moistness

MT8
fresh odour
fresh flavour

area
0.10
F4

0.15
50

150
PC1

250

Figure 4 The PCR-modelling of sensory attributes and process variables based on features extracted from images. The
symbols in the loading plot are the four factors: Process P1 and P2; flour quality (F1=Tjalve, F2=Folke, F3=baguette flour
A, F4=baguette flour B), mixing time (MT4=2+4 min, MT8=2+8 min), proofing time (PT1=65/150 min, PT2=80/
165 min).

Baguette making and wheat quality

p = 0.0028

p = 0.0001

11

9
Porosity

Crackles on the crust

11

81

7
5
3

7
5
3

1
P1

P2

P1

P2

p = 0.0099
11

Fresh flavour

Elasticity

p = 0.0224
11

7
5
3

5
3

1
P1

P2

P1

p = 0.0001

11

P2

p = 0.0001

5000

4000
Area mm2

Crispness of the crust

7
5
3

3000
2000
1000

1
P1

P2

P1

P2

Figure 5 Sensory attributes (crackles on the crust, porosity, elasticity, fresh flavour and crispness of the crust) and slice area
(mm2) aected significantly by process (P1 (traditional French) and P2 (modified industrial).

Correlation coefficient between porosity

measured by image analysis and by sensory
analysis
Figure 3 shows slices of 32 baguettes used in image
analysis based on a multivariate feature extraction.
The porosity was well predicted from image analysis using modelling with AMT extracted features.
The correlation coecient between porosity predicted from images and porosity evaluated by the

sensory panellist was 0.93 with a corresponding

root mean square error of prediction (RMSEP) of
0.62. This indicates that image analysis may be
used as a reference method for describing bread
porosity. The correlation coecient between
area of baguette slices and sensory porosity was
0.73.
The prediction of all sensory attributes and
process variables (Y-variables) based on AMT
feature vectors from image analysis (X-variables)

82

P. Baardseth et al.

Process 1 (traditional)
(a)
p = 0.0086
Area of cut surface

Porosity

9
7
5
3
1

2+4

7
5
3
1

2+8

p = 0.0005

2+4

2+8

Mixing time (min)

p = 0.0005
Area of cut surface

Porosity

9
7
5
3
1

65

7
5
3
1

80

p = 0.0001

65

80

Proofing time (min)

Figure 6 Sensory attributes aected significantly by mixing time and proofing time within each process. (a) Process 1 mixing
time 2+4 min (125 wh) and 2+8 min (300 wh), and 20 min fermentation time, 15 min pre-proofing, proofing time 65 min and
80 min. (b) Process 2 mixing time 2+4 min (180 wh) and 2+8 min (400 wh), and proofing time 150 min and 165 min (see Fig. 1).

by using principal component regression (PCR)modelling is shown in Figure 4.

The loading plot shows the relationship between
texture attributes and process variables and flour
qualities. Porosity, elasticity and juiciness correlated with the traditional French process and
firmness correlated with the industrial process.
Thus, the texture attributes could also be predicted
by AMT from the image analysis.
Effect of process on baguette quality
Baguettes produced by the traditional French process had significantly higher porosity measured
both by sensory analyses (Fig. 5) and image analyses (Fig. 4). This is also clearly seen in Figure
3. The area of the slices i.e. volume was also
significantly larger for baguettes produced by the

traditional French process compared with baguettes baked with the industrial adapted process
(Figs 35). Good French baguette quality is normally characterised by large volume and open
porosity25,26, which were obtained by baking with
the traditional French process. These results confirmed the importance of avoiding degassing of
the dough (gentle treatment) to achieve the open
and random pore structure of the crumb1,2. Baguettes produced by the traditional French process
also had significantly higher elasticity, crust crispness and crackles on the crust compared with
baguettes produced using the industrial adapted
process with sti dough and rough treatment (Figs
4 and 5). Furthermore, baguettes produced by the
traditional French process had significantly higher
fresh flavour (Fig. 5) which is due to higher concentration of yeast used in the recipe (09% vs

Baguette making and wheat quality

Process 2 (industrial)

Area of cut surface

p = 0.0001

Porosity

7
5
3

7
5
3

2+4

2+8

p = 0.0001

9
7
5
3
1

p = 0.0001

2+4

Crispiness of the crust

Crackles on the crust

(b)

83

2+8

p = 0.0001

9
7
5
3
1

2+8

2+4

2+4

2+8

Crispness of the crust

Mixing time (min)

p = 0.0012

9
7
5
3
1

150

165
Proofing time (min)

Figure 6(b)

03%) and longer fermentation time4,5. Mixing

time aected significantly the sensory attribute
porosity and area of cut surface for both processes
[Fig. 6(a,b)]. For the industrial process mixing time
also aected crackles on the crust and crispness of
the crust [Fig. 6(b)]. Fermentation/proofing time
aected significantly porosity and area of cut surface of the baguettes baked with the traditional
process [Fig. 6(a)], whereas crispness of the crust

was aected by proofing time when the industrial

process was used [Fig. 6(b)].
Effects of flour quality on baguette quality
Slice area i.e. volume was significantly larger for
baguettes produced using baguette flour B (4)
than baguettes produced using Tjalve flour (1) or
baguette flour A (3), which again was significantly

84

P. Baardseth et al.

Area mm 2

(a)

P1

2000

P2

1000

Porosity (image)

3000

15

10

(b)
P1

6
4

P2

2
0

15

10

Protein %

Protein %

(c)
3000
Area mm 2

2500
2000
1500
1000
500
0

P1

P2

Figure 7 (a) Eect of flour quality on area of the baguette slices, measured in mm2 by image analysis, Tjalve (), Folke
(), baguette flour A () and baguette flour B (), (b) Eect of flour quality on porosity, measured by image analysis, and
(c) interaction between flour quality and process on slice area.

6
ab

5
4

ab ab

a
b

b
a

ab

b b b

ab

a a a
ab

b b b

ab ab b

a
b

a
b
bc c

c bc

3
2
1

Glossiness

Crackles

Porosity

Surface

Elasticity

Fresh
odour

Fresh
flavour

Firmness

Juiciness

Crispness

Figure 8 Average values of each sensory attribute that were significantly dierent between baguettes baked from flours from
wheats of four dierent qualities. () Tjalve, (C) Folke, (C) baguette flour A, () baguette flour B. attributes, with no
significant level dierence according to Tukeys test are indicated by similar letter (ac).

larger than baguettes produced using Folke flour

(2) [Figs 3, 4, 7(a)]. Baguettes produced using
baguette flour B (4) also had a significantly more
open crumb structure (higher crumb porosity) than
Tjalve flour. Baguette flour A and Folke flour had
the densest crumb structure [Figs 3, 4, 7(b) and
8]. Baguette flour B (4) and Tjalve flour (1) have

both stronger protein quality and higher protein

content than the remaining flours (Table I). The
reason for the dierent quality of baguettes produced with flour B compared with Tjalve is not
known. Further experiments are needed to identify
more precisely optimum level of protein and protein quality for baguettes.

Baguette making and wheat quality

85

Table IV ANOVA for each sensory attribute aected by the interaction between flour
quality and process, flour quality and mixing time, flour quality and fermentation/proofing
time. The p-value gives the lowest significance value at which the two groups are dierent
(i.e. if the p-value is less than 005, the two groups are significantly dierent at the 005
level)

Glossiness
Crackles on the crust
Porosity
Area of cut surface
Elasticity
Odour intensity
Fresh odour
Flavour intensity
Fresh flavour
Salt flavour
Firmness
Moistness
Crispness of crust

Flour
qualityprocess

Flour
qualitymixing
time

Flour
qualityfermentation
/proofing time

00020
00003
00001
00002
00074
03921
00679
01662
00369
01183
01414
04389
00003

00782
00112
00003
00002
02113
01585
01416
01974
00538
03129
02544
00243
00001

04600
01142
00323
02754
00642
08692
05884
06362
02198
05907
00036
07071
00001

The area of cut surface was significantly higher

for baguettes produced using baguette flour A (3)
than with other flours (Figs 2 and 8). Dough made
from this flour was stable during mixing (70), but
had short development time (1.7) (Table I). On
the other hand, higher energy input during mixing
was recorded for these doughs compared to the
others (Table II). The moistness of baguettes produced with baguette flour A was also low (Fig. 8).
It is possible that this property relates to the quality
of the soft wheat, although water was added,
based on the water uptake determined using the
Farinograph.
Interaction between process and flour qualities
on baguette quality
The eect of flour depends on the process (Table
IV). Baguettes made from flours of all qualities
had a denser crumb structure when produced by
the modified industrial process compared with the
more traditional process, but there were small
dierences among the flours used in the modified
industrial process (Fig. 9). Despite these small
dierences, baguette flour B still gave the most
open crumb structure (Figs 3 and 8). Slice area of
the baguettes produced using flours of all qualities
also decreased significantly in the modified industrial process compared with the traditional
French process, but baguettes from baguette flour
B retained a relatively high slice area [Fig. 7(c)].

Crackles on the crust, crispness of the crust and

elasticity were significantly higher in baguettes
produced using traditional process compared with
industrial process for all flours. The dierences
between these attributes from process 1 to process
2 was, however, somewhat dierent for the various
flours (Fig. 7).
For the area of cut surface the interaction plots
(Fig. 9) were complex. This study also demonstrated (Table II) that the mixing time necessary
to obtain optimum dough development varies
among dierent flours and varies according to
mixing equipment, mixing intensity and recipe3.
When making the dough, the four flour qualities
behaved dierently using a fixed mixing time
(Table II). Interactions between flour quality and
mixing time was also found for the following
sensory attributes crackles on the crust, porosity,
area of cut surface, moistness and crispness of the
crust (Table IV). Porosity, firmness and crispness
of the crust were aected by interactions between
flour quality and fermentation/proofing time
(Table IV).

Acknowledgements
Technical assistance from the bakers Alf O. Nielsen
and Leif A. Fardal is greatly appreciated. We also wish
to thank Grethe Enersen and Bjrg Narum Nilsen for
skilful technical assistance during image recording.

86

P. Baardseth et al.

7
Crackles on the crust

Glossiness

6
5
4
3
2
1

P1

4
3
2

7
Area of cut surface

Porosity

P2

6
5
4
3
2
1

P2

P1

P2

P1

P2

6
5
4
3

P1

P2

5
4

5
4

P1

Elasticity

Crispness of the crust

P1

P2

P1

P2

Fresh flavour

11
9
7
5
3

Figure 9 Average values of the sensory attributes showing significant interaction (Table IV) between flour quality (()
Tjalve, () Folke, () baguette flour A, () baguette flour B) and process (P1 (traditional French) and P2 (modified industrial).

Baguette making and wheat quality

REFERENCES
1. Collins, T.H. Chorleywood process for French bread.
British Baker 16 (1978) 1118.
2. Tweed, A.R. A look at French Baguettes. Cereal Foods
World 28 (1983) 397399.
3. Magnus, E.M., Brathen, E., Sahlstom, S., Frgestad,
E.M. and Ellekjr, M.R. Eects of wheat variety and
processing conditions in experimental bread baking
studied by univariate and multivariate analyses. Journal
of Cereal Science 25 (1997) 289301.
4. Zehentbauer, G. and Grosch, W. Crust aroma of baguettes I. Key odorants of baguettes prepared in two
dierent ways. Journal of Cereal Science 28 (1998) 8192.
5. Zehentbauer, G. and Grosch, W. Crust aroma of baguettes II. Dependence of the concentrations of key
odorants on yeast level and dough processing. Journal of
Cereal Science 28 (1998) 9396.
6. Stear, C.A. Handbook of Bread Making Technology.
Elsevier Applied Science, London and New York (1990).
7. Kamel, B.S. and Stauer, C.E. Advances in baking.
Blackie Academic & Professional (Chapman & Hall)
Glasgow, England (1993).
8. Payne, P.I., Corfield, K.G. and Blackman, J.A. Identification of a high-molecular-weight subunit of gluteinin
whose presence correlates with bread making quality in
wheat of related pedigree. Theoretical Applied Genetic 55
(1979) 153159.
9. Payne, P.I., Nightingale, M.A., Krattiger, A.F. and Holt,
L.M. The relationship between HMW glutenin subunit
composition and the bread-making quality of British
grown wheat varieties. Journal of Science Food & Agriculture
40 (1987) 5165.
10. Frgestad, E.M., Magnus, E.M., Sahlstom, S. and Ns,
T. Influence of flour quality and baking process on hearth
bread characteristics made using gentle mixing. Journal
of Cereal Science 30 (1999) 6170.
11. International Organization for Standardization. International Standard. Wheat flourPhysical characteristics
of doughsPart 1: Determination of water absorption
and rheological properties using a Farinograph. ISO
5530-1: 1988 (E).
12. Finney, K.F. and Shogan, M.D. A 10 g mixograph for
determining and prediction functional properties of wheat
flour. Bakers Digest 46 (1972) 3235, 3842, 77.
13. American Association of the Cereal Chemists (AACC).
Approved Methods of the AACC, 8th ed. American

14.

15.
16.
17.

18.

19.
20.
21.
22.

23.

24.
25.

26.

87

Association of Cereal Chemists. Inc., St Paul, Minnesota.

Vol 2 (1995) 5661 A pp 14.
Uhlen, A.K. The composition of high molecular weight
glutenin subunits in Norwegian wheat and their relation
to bread-making quality. Norwegian Journal of Agricultural
Science 4 (1990) 117.
Finney, K.F. An optimized, straight-dough, bread-making method after 44 years. Cereal Chemistry 61 (1984)
2027.
Box, G.E.P., Hunter, W.G. and Hunter, J.S.Statistics for
experiments. John Wiley & Sons, New York (1978).
International Organization for Standardization. Draft
International Standard. Sensory analysisMethodologyGeneral guidance for the selection, training and
monitoring of assessors. Part 1: Selected assessors. ISO/
DIS 8586-1: 1989.
International Organization for Standardization. International
Standard.
Sensory
analysisMethodologyFlavor profile methods. 1st ed. 1985-10-15,
Geneve. ISO 6564-1985 (E).
Dallmann, H.Porentabelle. Schafer Inc., Detmold, Germany (1958).
International Organization for Standardization. International Standard. Sensory analysisGeneral guidance
for the design of test rooms. ISO 8589: 1988 (E).
Mardia, K.V., Kent, J.T. and Bibby, J.M.Multivariate
Analysis. Academic Press, London (1980).
Kvaal, K., Wold, J.P., Indahl, U.G., Baardseth, P. and
Ns, T. Multivariate feature extraction from textural
images of bread. Chemometrics and Intelligent Laboratory Systems 42 (1998) 141158.
Kvaal, K, Baardseth, P., Indahl, U.G. and Isaksson,
T. Relationships between sensory measurements and
features extracted from images. In Multivariate Analysis
of Data in Sensory Science. (T.Ns and E.Risvik eds),
Elsevier Science B. V. (1996) pp 135157.
Martens, H. and Ns, T.Multivariate Calibration. John
Wiley and Sons (1989).
Mani, K., Eliasson, A-C., Lindahl, L. and Tragardh, C.
Rheological properties and bread making quality of wheat
flour doughs made with dierent dough mixers. Cereal
Chemistry 69 (1992) 222225.
Payne, P.I., Corfield, K.G., Holt, L.M. and Blackman,
J.A. Correlations between the inheritance of certain highmolecular weight subunits of glutenin and bread-baking
quality in progenies of six crosses of bread wheat. Journal
of Science Food and Agriculture 32 (1981) 5160.