Fd Chem. Toxic. Vol. 25, No. 9, pp. 703-707, 1987 0278-6915/87 $3.00 + 0.

00
Printed in Great Britain. All rights reserved Copyright © 1987 Pergamon Journals Ltd

F A I L U R E OF S H O R T - T E R M IN VITRO M U T A G E N I C I T Y
TESTS TO P R E D I C T THE A N I M A L C A R C I N O G E N I C I T Y
OF H A I R D Y E S
C. M. BURNETT and J. F. CORBETT
Clairol Research & Development Laboratories, Stamford, CT 06902, USA

(Received 3 November 1986; revisions received 16 April 1987)

Abstract--A number of hair-dye chemicals that have given positive results in various short-term
mutagenicity tests have shown no clear evidence of carcinogenicity in animal bioassays. Commercial HC
Blue No. 1 and its analogues HC Red No. 3 and HC Blue No. 2 are all mutagenic in the Ames test but
only the HC Blue No. 1 is carcinogenic in animals. A carcinogenicity study in mice was carried out on
both a commercial sample of HC Blue No. 1 and a highly purified sample which was negative in a battery
of short-term tests for mutagenic activity. Both samples, administered at 0.3% in the diet for up to 24
months, were carcinogenic to mice, inducing hepatocellular carcinomas in ) 89% of the mice examined.
Therefore the presence of mutagenic impurities is not responsible for the carcinogenicity of commercial
HC Blue No. 1. Discrepancies between the results of mutagenicity and carcinogenicity tests on hair dyes
and other chemicals are discussed, and the value of short-term mutagenicity tests for assessing chemical
safety is questioned.

INTRODUCTION
Over the past decade there has been a shift in concern
over chemical safety towards effects of chronic low-
level exposures, particularly carcinogenic effects. The
National Cancer Institute bioassay program began
testing selected chemicals for carcinogenic activity
in lifetime animal studies in the early 1970s. These
studies are costly and time-consuming (approxi-
mately $1 million/test and 5-7 yr from preliminary
studies to peer review and publication of the technical
report) and therefore there has been considerable
research into quicker and less expensive means of
detecting carcinogenic potential in thousands of
untested substances. A number of in vitro short-
term tests were developed which were based on the
premise that carcinogens are mutagens (Ames et al.
1973) and that chemicals could be screened for
carcinogenic potential by testing for mutagenicity or
D N A alterations.
Publications claiming a high correlation between
the results of tests for mutagenic activity and those
for carcinogenicity, as determined in definitive life-
time animal studies, began to appear in the middle
1970s. Using much the same group of chemicals
(mostly known carcinogens) claims of 90% cor-
relations were made for the Ames test (McCann et al.
1975) and tests for malignant transformation (Pienta,
1979) and many people in industry began exploring
this new technology. However, it soon became appar-
ent that the selection of chemicals and the frequency
of carcinogens in the samples evaluated could
markedly influence the degree of correlation (Cooper
et al. 1979; International Commission for Protection
Against Environmental Mutagens and Carcinogens,
1983). All was not well. Clearly a more objective
Abbreviation: NTP = National Toxicology Program.
703
evaluation of the predictive value of short-term tests
was in order. The First International Collaborative
Study set out to do this. Perhaps the most important
results of this massive undertaking, published 1981,
were the questions it raised regarding the ability of
short-term tests to distinguish between carcinogens
and non-carcinogens, and the ability of the Ames test
to detect most carcinogens, irrespective of chemical
class (de Serres & Ashby, 1980).
In 1980 the U K Environmental Mutagen Society
set up a study to assess whether in vitro tests for
genotoxicity are predictive of effects in whole ani-
mals. The results of the study showed no activity in
various in vivo experiments for the model c o m p o u n d
.4-chloromethylbiphenyl which the selection commit-
tee predicted would be a potent mammalian carcino-
gen and mutagen in vivo (Ashby et al. 1982). In the
meantime the International Program for Chemical
Safety had begun another collaborative study to
evaluate a group of ten chemicals (eight carcinogens,
two non-carcinogens) in various in vitro eukaryotic
cell assays as well as in short-term in vivo tests. The
results of the in vitro tests have been published
(Ashby et al. 1985). The following points were made:
(1) The two non-carcinogens benzoin and capro-
lactam had significant genotoxic activity.
(2) There is still an urgent need to define what is
meant by a positive result, especially in the gene
mutation assays.
(3) Cell transformation assays are not recommen-
ded for routine work.
(4) It looks as though chromosome assays are
nearest to being refined into a useful routine assay
complementary to the Ames test.
(5) The way forward for using short-term tests is to
realize that there is a difference between in vitro and
in vivo; define the potential in vitro and then assess
704 C. M. BURNETT and J. F. CORBETT

the realization of this potential in vivo. F o r this
purpose A s h b y (1985) has r e c o m m e n d e d the use o f
Ames tests a n d in vitro cytogenetics followed by
in vivo tests on b o n e - m a r r o w cytogenetics a n d
perhaps also for unscheduled D N A synthesis
( U D S ) a n d sister-chromatid exchange in rat liver.
However, n o - o n e seems willing to address the very
t h o r n y a n d real p r o b l e m on h o w to deal with discor-
d a n t results in even a limited battery of in vitro a n d
in vivo tests, a situation that appears to have regu-
latory bodies which have insisted on h a v i n g these
kinds o f data, in a q u a n d a r y (Ashby, 1986; A s h b y et
al. 1983). Nonetheless, a n d in the face o f m o u n t i n g
evidence bringing into question the predictive value
o f s h o r t - t e r m tests for carcinogens, interest c o n t i n u e d
in exploring the situation. The N a t i o n a l Toxicology
P r o g r a m ( N T P ) set up a n in vitro testing p r o g r a m to
evaluate the use o f these tests in the selection of
chemicals for the carcinogenicity bioassays in ani-
mals. Preliminary results of this on-going effort have
recently been published (Zeiger, 1987). Shelby &
Stasiewicz (1984) pointed out the high level of mu-
tagenic activity of 70 n o n - c a r c i n o g e n s w h e n tested in
in vitro procedures for identifying genotoxic carcino-
gens. In a c o m p r e h e n s i v e analysis o f the d a t a
presented by Shelby & Stasiewicz (1984), A s h b y &
Purchase (1985) concluded that " t h e p h r a s e 'in vitro
genotoxin' clearly c a n n o t be used as a surrogate for
the term ' a n i m a l c a r c i n o g e n ' " . Zeiger & T e n n a n t
(1985) have presented data on the m u t a g e n i c activity
of 210 chemicals in the N T P , using the results of
Ames tests on all 210 a n d o f tests for sister-chromatid
exchanges a n d c h r o m o s o m a l a b e r r a t i o n s on 95 of
them. They concluded " t h e expectations o f the ability
of in vitro mutagenesis a n d clastogenesis assays to
u n a m b i g u o u s l y detect chemical carcinogens have not
been fulfilled. The results of this study show a
c o n s t a n t error rate in the identification of non-
carcinogens with mutagenic potential".
The p a p e r (Ames et al. 1975) t h a t implicated hair
dyes as h u m a n carcinogens on the basis o f positive
results in the Salmonella test (and in which it was
stated t h a t "despite the possibility t h a t particular
classes o f bacterial m u t a g e n s m a y be f o u n d that are
not carcinogenic, it appears t h a t each of the hair dye
c o m p o u n d s we have f o u n d to be mutagenic has a
high probability of providing to be a carcinogen")
spawned a high level of interest in these materials and
m a n y in vitro tests were p e r f o r m e d prior to the
publication of the results of the a n i m a l bioassays that
were in progress in the N T P . A c o m p a r i s o n of the
results of the n o w completed animal bioassays and
various s h o r t - t e r m tests is s h o w n in Table 1. The high
level of false-positive results in the short-term tests is
apparent. O f interest also is the inability of the Ames
test to distinguish the animal carcinogen H C Blue
No. 1 from its non-carcinogenic analogues H C Red
No. 3 a n d H C Blue No. 2, the structures of which are
s h o w n in Fig. 1.
The trend towards positive results in tests with
non-carcinogenic hair dyes is obvious but the
reason(s) for this are not. Possible explanations for
genotoxic effects of non-carcinogens discussed by
A s h b y (1985) include: (1) innate genotoxicity not
expressed in vivo as a result of n o n - a b s o r p t i o n ,
hydrolysis, rapid excretion, preferential enzymatic
detoxification a n d (2) genotoxicity due to the pres-
ence of mutagenic impurities. These explanations
suggest that there may be no-effect levels in vivo for
c o m p o u n d s t h a t are mutagenic in vitro a n d t h a t it is
possible that for some carcinogens the activity may
reside in potent m i n o r c o n t a m i n a n t s . If the latter was
so, then purification of the chemical to the point of
its being negative in a battery of short-term tests
would assure its safety, i.e. it would not be likely to
be carcinogenic if tested in a lifetime carcinogenicity
bioassay in a sensitive species.
The o p p o r t u n i t y to test this concept in our labora-
tory occurred recently. The N T P evaluated a sample

Table I. The results of animal bioassays and short-term tests for carcinogenicity on hair-dye
chemicals
Results in:
Mouse
Ames lymphoma Transformation
Chemical Bioassay* test test* test*
2,4-Diaminoanisole + + + +
4-Amino-2-nitrophenol + + + NT
2-Nitro-p-phenylenediamine + + +
4-Nitro-o-phenylenediamine - + + +
m-Phenylenediamine - + + +
p-Phenylenediamine - + + +
p-Toluenediamine + NT NT
HC Blue No. I + + NT NT
HC Blue No. 2 - + NT NT
HC Red No. 3 - + NT NT
Disperse Blue No. 1 + + NT NT
2-Amino-4-nitrophenol - + NT NT
2-Amino-5-nitrophenol + NT NT
NT = Not tested
*For bioassays + = clear evidence of carcinogenicity, = negative result or carcinogenicity not
proven. The carcinogenic effect of Disperse Blue No. 1 was strongly associated with bladder
calculi and was not likely to be due to the chemical.
*For in vitro tests + = positive result. In vitro mammalian cell tests (the mouse lymphoma test
and transformation tests using Syrian hamster embryo cells) were discontinued on the more
recently bioassayed dyes when it became clear that these procedures were producing very
high levels of false positives on these and many other chemicals (National Toxicology
Program, unpublished data).
The bioassays were carried out as part of the US National Toxicology Program.
 Mutagenicity v. carcinogenicity tests on HC Blue No. 1 705

NHCHa Table 3. Body and liver weights of mice killed in wk 91 of treatment

(a)
~ NO2

N (CH2CH2OH)2
Group
A
B
C
F
Body weight Liverweight Relativeliver weight
(g)
33.5
32.7
30.7
30.3
(g)
1.70
8.34
7.99
3.34
(% of body weight)
5.05
25.30
26.00
I 1.30
*For details of treatment, see Table 2.
Values are means for ten mice/group.

NHCH2CH20H Table 4. Body and liver weights of mice killed at the end of the study
Body Liver
(bt ~ N O 2 No. of weight weight Relativeliver weight
Group mice (g) (g) (% of body weight)
A 30 30.8 1.7 5.6
B 17 29.9 5.9 19.7
NH2 C 24 30.5 6.2 20.2
D 20 31.4 3.7 12.2
E 14 29.8 4.5 15.3
F 32 29.8 2.8 9.4
*For details of treatment, see Table 2.
NHCH2CH2OH Values are means for the number of mice shown.
(c) ~ N O 2
gested by the E u r o p e a n E c o n o m i c C o m m u n i t y . A
sample of this purified n o n - m u t a g e n i c dye was ob-
N(CH2CH~OH)2 tained a n d we tested it as part o f a larger study
evaluating the effects of different dosing regimens on
Fig. 1. Structures of (a) the animal carcinogen HC Blue t u m o u r incidence. The following is a s u m m a r y o f the
No. 1 (N4,N4-bis(2-hydroxyethyl)-N'-methyl-2-nitro-p - i m p o r t a n t details of this study.
phenylenediamine) and the non-carcinogens (b) HC Red
No. 3 (N'-(2-hydroxyethyl)-2-nitro-p-phenylenediamine)
and (c) HC Blue No. 2 (N',N4,N4-tris(2-hydroxyethyl)- MATERIALS AND METHODS
2-nitro-p-phenylenediamine).
TWO h u n d r e d a n d forty female B 6 C 3 F l mice (6 wk
old) were r a n d o m l y assigned to four groups o f 48
of the dye H C Blue No. 1 ( N 4, N4-bis(2-hydroxy - a n d two groups o f 24 animals each. They were
ethyl)-N~-methyl-2-nitro-p-phenylenediamine) in the housed six/cage in a bio-clean r o o m ( H E P A filters,
s t a n d a r d lifetime a n i m a l bioassay a n d f o u n d that it 10 air changes/house; 70_+ 4°F; 50_+ 10% relative
resulted in a very high level o f liver t u m o u r s in mice humidity) a n d were fed R a l s t o n P u r i n a Certified
w h e n fed at levels o f 0.6 or 0.3% in the diet of female R o d e n t C h o w (Ralston P u r i n a Co., Inc, St Louis,
mice a n d at levels of 0.3 or 0.15% in the diet o f male M O ) with or w i t h o u t added H C Blue No. 1 (Table
mice. These levels caused no a p p a r e n t adverse effects 2). Fresh samples o f diet were supplied to the mice
o n the general health o f the animals. The sample once a week for the first 11 m o n t h s a n d twice weekly
tested was the item o f c o m m e r c e a n d was mutagenic thereafter. All diets were stored frozen until used. The
in the A m e s S a l m o n e l l a / m a m m a l i a n m i c r o s o m e test stability of the dye-containing diets for 1 wk u n d e r
with or w i t h o u t metabolic activation, in the m o u s e l a b o r a t o r y conditions h a d been established prior to
l y m p h o m a system, a n d in the in vitro rat hepatocyte the initiation of the study. The mice were weighed
U D S test for D N A repair (NTP, 1985). once a week for 13 wk a n d m o n t h l y thereafter. A
A t a b o u t the time that the results o f the N T P study gross autopsy was performed on all animals that died
o n H C Blue No. 1 became available, a series o f or were killed when m o r i b u n d a n d a complete set o f
papers was published by three groups of investigators tissues was saved unless they were badly autolysed.
( D a r r o u d i et al. 1983; L o p r i e n o et al. 1983; S h a h i n & After 91 wk ten mice from each o f groups A, B, C a n d
Bugaut, 1983) which showed n o genotoxicity when a F were weighed, killed a n d autopsied a n d all m a j o r
highly purified sample o f H C Blue No. 1 was tested organs a n d tissues a n d tissue masses a n d gross lesions
in a series of studies c o n f o r m i n g to a battery sug- were removed a n d preserved in 10% neutral buffered

Table 2. Outline of treatment of dietary groups

Test No. of
group Test material* Duration of treatment mice
A None (control) Continuous, 24 months 48
B 0.3% HC Blue No. 1 (2842) Continuous, 24 months 48
C 0.3% HC Blue No. 1 (2-117) Continuous, 24 months 48
D 0.3% HC Blue No. 1 (2-117) 9 months (then control diet for 15 months) 24
E 0.3% HC Blue No. 1 (2-117) 15 months (then control diet for 9 months) 24
F 0.3% HC Blue No. l (2-117) Every other week for 24 months 48
*The highly purified sample of HC Blue No. I was coded 2842. The Clairol sample 2-117 was a commercial
sample that was recrystallized and charcoal filtered and was negative in the Ames test.
706 C . M . BURNETTand J. F. CORBETT
Table 5. Incidence of liver tumours in B6C3F~ female mice fed HC Blue No. 1 at a level of
0.3% in the diet
No. of mice
Group* Initially Examined Adenomas Carcinomas
A 48 48
B 48 46
C 48 46
D 24 22
E 24 23
F 48 46
*For details of treatment, see Table 2.
formalin. Following removal, the liver, kidneys, heart
and brain were weighed before fixation. After 24
months all survivors were weighed and killed and a
gross autopsy was performed and tissues were taken
as at 91 wk.
RESULTS
The mice consuming the dye-containing diets
showed body-weight gains approximately equal to
that of the controls for the first 26 wk, after which
body weights were slightly less in all of the dye-fed
groups. However, the mice all appeared healthy and,
except in group F which was given the dye-containing
diet every other week, the average body weights of all
dye-fed groups were at least 90% of that of the
controls. In group-F mice the average body weight
was 85-90% of the control value. Survival was
excellent in all groups during the first year; no more
than one mouse died in any group. Between months
12 and 21 the numbers of mice that died or were
killed when moribund were as follows: G r o u p A, 7;
G r o u p B, 12; G r o u p C, 9; G r o u p D, 3; G r o u p E, 7;
G r o u p F, 3.
Beginning in month 18 of the study hepato-
cytomegaly was noted in animals that died or were
killed in groups B and C. The number of moribund
animals increased in these groups between months 18
and 20. The mean body weights and absolute and
relative liver weights of ten selected mice that were
killed in wk 91 are shown in Table 3. The liver
enlargement was striking and virtually every animal
in groups B and C was affected. In most animals
every lobe was affected, giving the liver the appear-
ance of a large mass with rounded edges. In addition
to liver enlargement the most noteworthy gross
findings were dark thyroid glands, blue urine, and
blue bile. Much the same picture was seen grossly in
mice surviving to the end of the study (Table 4).
The livers of all animals that had not been auto-
lysed were prepared and examined microscopically by
Experimental Pathology Laboratories, Herndon, VA.
The results are presented in Table 5. The carcinogenic
response was overwhelming. Nearly every mouse in
each test group had liver tumours and most of these
were hepatocellular carcinomas. Feeding for 9 or 15
months or every other week did not markedly reduce
the percentage of mice with liver tumours but did
result in a lower percentage of mice with malignant
neoplasms in groups fed the dye-containing diets for
only 9 months or every other week.
On the basis of the results of the N T P bioassay and
the study described above it is concluded that:
(1) A commercial sample of H C Blue No. 1 tested
No. with hepatoceUular Total no. with
liver tumours
I 2 3
10 41 45
10 44 46
10 13 20
4 19 20
21 25 40
by N T P was genotoxic and was carcinogenic in a
standard 2-yr bioassay at non-toxic dose levels.
(2) Samples of H C Blue No. 1 which were non-
genotoxic--sample 2 117 in the Ames test, and
sample 2482 in a six-test battery (Darroudi et al.
1983; Loprieno et al. 1983; Shahin & Bugaut,
1983)--were highly carcinogenic for the liver of
female B6C3Fj mice when tested as in the N T P
study.
(3) The pure compound N4,N4-bis(2-hydroxy-
ethyl)-N ~-methyl-2-nitro-p-phenylenediamine is car-
cinogenic for mouse liver.
(4) The presence of mutagenic impurities in com-
mercial H C Blue No. 1 is not responsible for its
carcinogenicity. HC Blue No. I that had been
purified to an extent at which it was non-genotoxic
was carcinogenic in an animal bioassay.
DISCUSSION
The failure of a battery of tests to predict the
carcinogenicity of H C Blue No. 1 is no less surprising
than the similar failure of many positive in vitro test
results to predict the outcome of two bioassays
showing p-phenylenediamine to be non-carcinogenic
(Imaida et al. 1983; N T P , 1979). Clearly, the results
of short-term tests fail to predict the carcinogenicity,
or lack thereof, of hair-dye chemicals. As Brusick
(1986) has stated "the objective of genetic toxicology
to detect carcinogens has not been successfully
accomplished by application of the array of test
methods currently available. A principal task in
the reassessment of genetic toxicology should be the
development of a set of realistic objectives for this
technology."
In view of the accumulating evidence of the inabil-
ity of short-term tests to predict accurately the out-
come of animal carcinogenicity bioassays of a broad
spectrum of chemical types, and particularly hair
dyes, it is clear that presentation of the results of
short-term tests with the implication that these results
are relevant for assessing chemical safety is without
scientific merit and is misleading.
Acknowledgements--We wish to thank T. Re, R. Loehr,
S. Rodriguez and S. Posbergh for their contributions to the
conduct of the study on the purified dyes.
REFERENCES
Ames B., Durston W., Tamasaki E. & Lee F. (1973).
Carcinogens are mutagens: a simple test system com-
bining liver homogenates for activation and bacteria for
detection. Proc. nam. Acad. Sci. U.S.A. 70, 2281.
 Mutagenicity v. carcinogenicity tests on HC Blue No. 1
Ames B., Kammen H. & Yamasaki E. (1975). Hair dyes are
mutagenic: identification of a variety of mutagenic ingre-
dients. Proc. hath. Acad. Sci. U.S.A. 72, 2423.
Ashby J. (1985). The International Program on Chemical
Safety Collaborative Study on In-vivo Short-Term Tests.
Paper presented at the Sixteenth Annual Meeting of the
US Environmental Mutagens Society, Las Vagas, 25
February, 1985.
Ashby J. (1986). The prospects for a simplified and inter-
nationally harmonized approach to the detection of pos-
sible human carcinogens and mutagens. Mutagenesis l, 3.
Ashby J., de Serres F., Draper M., Ishidate M., Margolin
B., Matter B. & Shelby M. (Editors) (1985). Progress in
Mutation Research: Evaluation of Short-Term Tests for
Carcinogens--Report of the International Program on
Chemical Safety's Collaborative Study on In-Vitro Assays.
Elsevier, Amsterdam.
Ashby J., de Serres F., Draper M., Ishidate M., Matter B.
& Shelby M. (1983). The two IPCS collaborative studies
on short-term tests for genotoxicity and carcinogenicity.
Mutation Res. 109, 123.
Ashby J., Lefevre P., Elliott B. & Styles J. (1982). An
overview of the chemical and biological reactivity of
4CMB and structurally related compounds: possible rele-
vance to the overall findings of the UKEMS 1981 study.
Mutation Res. 100, 417.
Ashby J. & Purchase I. (1985). Significance of the genotoxic
activities observed in-vitro for 35 of 70 NTP non-
carcinogens. Envir. Mutagen. 7, 747.
Brusick D. (1986). Molecular Toxicology Newsletter, 5 (3).
Cooper J., Saracci H. & Cole P. (1979). Describing
the validity of carcinogen screening tests. Br. J. Cancer
39, 87.
Darroudi F., Sobels F. H. & Natarajan A. T. (1983).
Evaluation of the mutagenic potential of the hair
dye N-methyl-amino-2-nitro-4-N,N-bis(2-hydroxyethyl)-
aminobenzene in a battery consisting of Drosophila and
mammalian cytogenetic assays. Mutation Res. 116, 169.
de Serres F. J. & Ashby J. (Editors) (1980). Short-term Tests
for Carcinogens: Report on the International Collaborative
Program. Elsevier, New York.
707
Imaida K., Ishihara Y., Nishio O., Nakaniski K. & Ito N.
(1983). Carcinogenicity and toxicity tests on p-phenylene-
diamine in F344 rats. Toxicology Lett. 16, 259.
International Commission for Protection Against Environ-
mental Mutagens and Carcinogens (1983). Committee 3
Final Report. Mutation Res. 114, 179.
Loprieno N., Mariani L. & Rusciano D. (1983). Lack
of genotoxic properties of the hair-dye component, N-
methyl-amino-2-nitro-4-N,N-bis(2-hydroxyethyl)amino-
benzene, in mammalian cells in vitro, and in yeasts.
Mutation Res. 116, 161.
McCann J., Choi E.; Yamasaki E. & Ames B. (1975).
Detection of carcinogens as mutagens in the Salmonella/
microsome test: assay of 300 chemicals. Proc. natn. Acad.
Sci. U.S.A. 72, 5135.
National Toxicology Program (1979). Bioassay of p-
Phenylenediamine Dihydrochloride for Possible Carcino-
genicity. Technical Report No. 174.
National Toxicology Program (1985). Toxicology and Car-
cinogenesis Studies of HC Blue 1. Technical Report
No. 271.
Pienta R. (1979). A hamster embryo cell model system for
identifying carcinogens. In Carcinogens: Identification and
Mechanisms of Action. p. 121. Edited by C. Griffin &
C. Shaw. Raven Press, New York.
Shahin M. M. & Bugaut A. (1983). Mutagenic evaluation
of the monocyclic aromatic amine N-methylamino-
2-nitro-4-N,N-bis(2-hydroxyethyl)aminobenzene in the
Salmonella typhimiurium/mammalian microsome test.
Mutation Res. 116, 155.
Shelby M. & Stasiewicz S. (1984). Chemicals showing no
evidence of carcinogenicity in long-term, two-species
rodent studies: the need for short-term test data. Envir.
Mutagen. 6, 871.
Zeiger E. (1987). Carcinogenicity of mutagens: predictive
capability of the Salmonella mutagenesis assay for rodent
carcinogenicity. Cancer Res. 47, 1287.
Zeiger E. & Tennant R. (1985). Mutagenesis, clastogenesis,
carcinogenesis: expectations, correlations and relations.
Proceedings of the 4th International Conference on En-
vironmental Mutagens. Stockholm, July. In press.