Research Notes
A Method for Separating Sex-Linked Imperfect Albino (S*ALS)
and Nonalbino Embryos Before Hatch
G. A. SANTOS and F. G. SILVERSIDESl
Departement des sciences animates, Universite Laval, Ste-Foy, Quebec, Canada G1K 7P4
ABSTRACT The feasibility of using the sex-linked
gene for imperfect albinism (S*ALS) to sex chicks during
incubation by candling was studied. With this technique,
the dark eye of nonalbino embryos can be positively
identified. Two trials were performed. In a first trial,
66.5 and 89.5% of the 254 nonalbino and 210 albino
chicks produced in four hatches were correctly identified by candling at 7 d of incubation. Of 191 eggs
predicted to be nonalbinos, 22 were albinos, resulting in
an overall accuracy of 88.5% for identification of
nonalbino embryos. In a second trial, the accuracy of the
technique from 7 to 10 d of incubation was evaluated.
Increased age resulted in a tendency for lower accuracy,
but candling at 8, 9, or 10 d of incubation allowed
identification of a greater (P < 0.05) proportion of the
nonalbino population than at 7 d of incubation.
Candling at 8 d of incubation allowed identification of
nonalbinos and albinos with an accuracy of 81.3 and
84.9%, respectively, suggesting that the *ALS gene could
be used to sex chicks during incubation when used in a
sex-linked cross. This technique may prove advantageous to the laying industry because of savings of
incubator and hatcher space. The males of commercial
layer lines are normally killed at hatch. Reducing the
number that hatch by eliminating them before 10 d of
incubation could diminish animal welfare concerns.
(Key words: sexing, candling, albinism, chicken, sex-linked gene)
1996 Poultry Science 75:585-588
associated with reduced growth before 1 wk of age in
some environments (Silversides and Crawford, 1991)
and an increase in egg production late in the laying
period (Su et al, 1994). In a mating scheme in which
albino males are mated to nonalbino hens, the albino
female chick is characterized by its red eyes and the
nonalbino male can be recognized by its black eyes. This
scheme does not require the presence of other specific
genes for color with the probable exception of the
absence of autosomal albinism.
In the layer industry, male chicks are killed immediately after hatch, and this practice could generate animal
welfare concerns. A method for sexing the developing
chick embryo during incubation would significantly
reduce the number of males hatching and therefore
reduce these concerns by reducing the number of chicks
to be killed. In addition, sex separation during incubation could provide significant savings in incubator space
and in energy consumption. This paper describes
research on the use of the *ALS gene to determine the
chick genotype before 10 d of incubation.
INTRODUCTION
Sex separation at hatch is an important aspect of the
poultry industry because only the female is kept in the
laying industry and the broiler industry often raises the
sexes separately. Chicks can be sexed by observing
differences in the rudimentary copulatory organs (Jull,
1934), but this method of sexing, known as vent sexing,
is labor-intensive and therefore expensive. The K gene
for slow feathering is used extensively in the chicken for
sex segregation (Warren, 1976), but its close linkage with
an endogenous virus (Bacon et ah, 1988), which can
cause immunological tolerance to lymphoid leukosis
(Crittenden et ah, 1987), is a disadvantage. Neither ventsexing nor feather sexing can be used to sex chicks prior
to hatch, nor do they provide 100% sexing accuracy.
The sex-linked gene for imperfect albinism (S*ALS,
using the nomenclature described by Crittenden et al.,
1995), which was previously known as sal"s, can also be
used to sex chicks at 1 d of age (Hutt and Cole, 1973) on
the basis of eye color with 100% accuracy (Silversides
and Crawford, 1991). This method of sex segregation has
the advantages of simplicity and that the gene is not
associated with endogenous viral loci of the same type
as is K (Silversides et ah, 1993). This gene has been
MATERIALS AND METHODS
Albino (S*ALS) hens of a line laying white-shelled
eggs (Su et al, 1994) were inseminated with semen from
heterozygous albino males. Two experiments were
performed. In the first, eggs were set in four hatches to
produce a total of 464 chicks in approximately equal
Received for publication September 18, 1995.
Accepted for publication January 9, 1996.
'To whom correspondence should be addressed.
585
586
SANTOS AND SILVERSIDES
TABLE 1. Identification of embryo genotype by candling at 7 d of incubation
Error1
Actual
no.
Nonalbino
1
2
3
4
Total
66
77
82
29
254
Population identified
Albino
Type I
Type II
76
50
64
20
210
34.2
25.0
27.9
39.3
31.1
9.7bc
17.7»"
3.3'
14.3"
11.5
(n).
Nonalbino
Albino
42.4b
84.4"
70.7*
62. lab
96.1=
72.0b
96.9^
89.5*b
89.5
- (%)
66.5
'-^Percentages in the same column with no common superscript differ significantly (P < 0.05).
J
Type I error: predicted as albino when the real genotype is nonalbino; Type II error: predicted as
nonalbino when the real genotype is albino.
proportions of albino and nonalbino siblings. In this
experiment, the eggs were candled after 7 d of
incubation to attempt to identify the nonalbino embryos
on the basis of their eye color. During candling, the eggs
were rotated two or three times to cause the eye of the
7-d embryo to make contact with the shell. At this age,
the dark eye of the nonalbino embryo can be seen
through the shell of white eggs and the red eye of the
albino embryo cannot be differentiated from the blood
vessels of the embryonic membranes. When the dark eye
could be observed through the shell at candling, the
embryo was classified as nonalbino, and when no dark
pigment was seen, the egg was classified as albino.
Because in a sexing scheme for laying hens the females
are kept, all borderline eggs were classified as albinos.
The same procedure was used in a second experiment to
classify the same 81 embryos by candling after 7, 8, 9,
and 10 d of incubation. In both experiments, the actual
chick genotype was identified at hatch, and eggs that
did not hatch were opened to identify the embryo
genotype. All observations were performed by a single
investigator.
Statistical analyses were performed by application of
contingency chi-square with Yate's correction for continuity (Steel and Torrie, 1980). The method described
allows the positive identification of only nonalbinos, so
the Type I error was defined as the classification of
nonalbino embryos as albinos (exclusion of individuals
that should be included), and the Type II error was
defined as the classification of albino embryos as
nonalbinos (inclusion of individuals that should have
been excluded). The results obtained in Experiment 1
were also compared to those expected by chance.
Expected values for a prediction by chance were
calculated on the basis of the actual proportions of each
genotype in the population.
RESULTS
For the four hatches in Experiment 1, a total of 210
albino and 254 nonalbino chicks were obtained (Table 1).
At candling, 191 eggs were predicted to be nonalbinos,
and of these 22 were actually albinos, resulting in a Type
II error of 11.5% and an accuracy of 88.5%. The use of
the S*ALS gene to segregate albino from nonalbino
embryos by candling at 7 d of incubation allowed
correct identification of 66.5 and 89.5% of the nonalbino
and albino populations, respectively (Table 1).
Several differences in the efficiency of the technique
were observed between hatches (Table 1). In Hatch 2,
84.4% of the nonalbino population was correctly identified, but a much lower (P < 0.05) proportion (72.0%) of
the albino population was correctly identified than in
the other three hatches (89.5 to 96.9%). More borderline
embryos were classified as nonalbinos in Hatch 2, and
this is reflected in the higher Type II error (17.7%) and
the lower Type I error (25%) than in the other hatches.
In Hatches 1, 3, and 4 over 89.5% of the albino
population was correctly identified. For all hatches, the
Type I error was consistently higher than the Type II
error (Table 1).
The accuracy of prediction of the genotype by
candling of the eggs in Experiment 1 was compared to
that expected if the predictions had occurred by chance
(Table 2). Of the 191 embryos predicted to be nonalbinos, chance would have dictated that only 105 of them
TABLE 2. Comparison of predictions of the genotype obtained
by candling with those expected by chance alone
Expected 2
Actual
genotype
(n)i
Nonalbino
Albino
Nonalbino
Albino
X2
P
Nonalbinos
Albinos
Total
191
273
464
169
85
254
22
188
210
104.6
149.4
254
86.4
123.6
210
51.8
29.9
<0.01
<0.01
dumber of each genotype predicted by candling.
2
Expected prediction = n x (actual proportions of a given genotype in the observed population).
RESEARCH NOTE
587
TABLE 3. Identification of embryo genotype by candling at 7, 8, 9, and 10 d of incubation
Error 1
Actual
Age
Nonalbino
(d)
7
8
9
10
48
48
48
48
Population identified
Albino
Type I
Type II
33
33
33
33
36.2
24.3
18.8
27.3
8.8
11.4
14.3
18.8
(n)
Nonalbino
Albino
64.6b
81.3»>>
87.5*
81.3 ab
90.9
84.8
78.8
72.7
(%)
a b
' Percentages in the same column with no common superscript differ significantly (P < 0.05).
Type I error: predicted as albino when the real genotype is nonalbino; Type II error: predicted as
nonalbino when the real genotype is albino.
:
should have been nonalbinos, but the candling technique allowed the correct identification of 169 of them.
Similar results were observed for the embryos predicted
to be albinos. The number of correct predictions of
genotype with the candling technique using *ALS was
significantly greater (P < 0.01) than those that would
have been expected by chance alone.
In Experiment 2 (Table 3), at 7 d of incubation the
Type I and Type II errors (36.2 and 8.8%, respectively)
and the proportions of embryos of each genotype
correctly identified (64.6 and 90.9% for nonalbinos and
albinos, respectively) were similar to those observed in
Experiment 1. Candling at 8, 9, and 10 d of incubation
resulted in the correct prediction of a higher (P < 0.01)
proportion of the nonalbino population (81.3 to 87.5%)
than at 7 d of incubation (64.6%). As the embryos
became older, the Type II error rate increased and the
proportion of the albino population correctly identified
decreased (neither of them significantly).
DISCUSSION
The proposed utilization of the S*ALS gene for sex
determination during incubation constitutes a new
application for the gene. The data presented show
clearly that it is possible to determine the embryo's eye
color and thus its genotype by candling from 7 to 10 d of
incubation. In a sexing scheme in which the male parent
is albino and the female parent is nonalbino, use of the
*ALS gene would allow separation of male and female
embryos at 7 to 10 d of incubation.
Prior to the experiments described, the investigator
had no experience in segregating albino and nonalbino
chicks by candling. Although the results obtained were
relatively consistent between hatches, it should be
expected that accuracy would improve with practice.
Accuracies rivalling those of vent sexing (full, 1934) or of
the use of the gene K (Warren, 1976) may be difficult to
obtain, but neither of these methods allows sexing prior
to hatch.
The results presented also show that the accuracy of
the method can be manipulated in the direction desired
by the technician. A greater proportion of the nonalbino
population can be identified if desired, but this results in
greater Type II errors and a lower proportion of the
albino population being correctly predicted. Alternately,
it is possible to correctly identify a large proportion of
the albino population but a lower proportion of the
nonalbino population will be identified and higher Type
I errors will result.
Comparison of the efficiency of the candling technique from 7 to 10 d of incubation shows that the dark
eye of nonalbinos becomes increasingly easy to detect
with increasing age. This is reflected in the larger
percentage of the nonalbino population that was detected. However, the increase in age also results in
reduced mobility of the embryo and thicker embryonic
membranes, making the observation of the eye more
difficult and producing a larger Type II error rate. The
consequences of Type I and Type II errors will
determine which is most desirable and therefore the best
age for candling.
Sexing of embryos during incubation using the *ALS
gene presents several advantages over current sexing
techniques. The technique could be mechanized with a
spectrophotometric method similar to that described by
Brant et al. (1953) to detect blood spots in white-shelled
eggs. In addition, in the laying industry the technique
would allow the identification and elimination of the
majority of the nonalbino (male) population while
conserving nearly all of the albino (female) population.
At hatch, the rest of the females could be identified on
the basis of eye color with 100% accuracy, as suggested
by Silversides and Crawford (1991).
The elimination of the males before 10 d of incubation
would result in significant economic savings because of
lower incubator and hatcher space requirements. This
technique would also significantly reduce the number of
males to be killed at hatch, which may prove important
because of increasing animal welfare concerns in the
chicken industry.
REFERENCES
Bacon, L. D., E. Smith, L. B. Crittenden, and G. B. Havenstein,
1988. Association of slow feathering (K) and an endogenous viral (ei>21) gene on the Z chromosome of
chickens. Poultry Sci. 67:191-197.
Brant, A. W., K. H. Norris, and G. H. Chin, 1953. A
spectrophotometric method for detecting blood spots in
white-shelled eggs. Poultry Sci. 32:357-363.
Crittenden, L. B., S. McMahon, M. S. Halpern, and A. M.
Faldly, 1987. Embryonic infection with the endogenous
588
SANTOS AND SILVERSIDES
avian leukosis virus Rous-associated virus-0 alters
response to exogenous avian leukosis infection. J. Virol. 61:
722-725.
Crittenden, L. B., J. Bitgood, and D. Burt, 1995. Chick. Pages
33-34 in: Genetic Nomenclature Guide. Trends Genet.
Elsevier, Amsterdam, The Netherlands.
Hutt, F. B., and R. K. Cole, 1973. Identification of sex in chicks
by use of the gene sal. Poultry Sci. 52:2044.(Abstr.)
Jull, M. A., 1934. The feasibility of sex segregation in day-old
chicks. Poultry Sci. 13:250-254.
Silversides, F. G., and R. D. Crawford, 1991. Effects of
imperfect albinism (sa|-s) on growth in a heavy line of
chickens. Poultry Sci. 70:6-12.
Silversides, F. G., M. Tixier-Boichard, L. Durand, and P. Merat,
1993. Research note: The sal~c mutation in the chicken is not
linked to Rous-associated virus-2 related DNA sequences.
Poultry Sci. 72:368-372.
Steel, R.G.D., and J. H. Torrie, 1980. Principles and Procedures
of Statistics: A Biometrical Approach. McGraw-Hill Publishing Co., Montreal, QC. Canada.
Su, H., F. G. Silversides, and P. Villeneuve, 1994. Production
and quality of eggs from sex-linked imperfect albino (sal_s)
and nonalbino (s+) hens kept under commercial conditions. Poultry Sci. 73:1813-1821.
Warren, D. C, 1976. Feather sexing chicks. Poult. Trib. 82(2):30,
32-34.