Selection for Egg Mass in the Domestic Fowl. 1. Response to Selection H. L. MARKS US Department of Agriculture, Science & Education Administration, Agricultural Research, uthern Regional Poultry Breeding Project, University of Georgia, Athen Georgia 362 (Received for publication September 1, 198) ABSTRACT A study was conducted to measure selection response for daily egg mass from 154 to 294 days of age in four s of White Leghorns and to determine the possibility of selecting males with high packed-sperm volume (PSV) as a corollary to egg mass. After three generations of selection, deviation of the selected s from the control in total egg mass per bird for 28 days of production varied from 137 to 2173 g. Data failed to support the hypothesis that selection for PSV in males is associated with selection for increasing egg mass. Realized heritability for daily egg mas obtained by dividing the cumulative corrected response by the cumulative selection differential ranged from.23 to.66 with a mean across s of.42. Increases in egg mass equal or superior to those observed during the period of selection (154 to 294 days of age) were also observed during the second half of the test period (294 to 434 days). (Key words: chicken genetic egg production, selection, egg mass) 1981 Poultry Science 6:1115-1122 INTRODUCTION The negative genetic correlation between egg number and egg weight in the domestic fowl is well documented (Abplanalp, 1956; Jerome et al., 1956; Hick 1958; Clayton and Robertson, 1966; Craig et al, 1969; Kinney, 1969; Quadeer et a/., 1977). The economic importance of this negative relationship between egg weight and egg production has stimulated recent studies of the potential value of egg mass as a selection tool. The concept of using egg mass as a selection tool to improve egg production efficiency was suggested by Waring et al. (1962) and Hicks (1963). Bohren (197) suggested that selection for egg mass might be accomplished through the use of an index formed from a ar combination of the logarithmic functions of the component traits but restricted so as to allow no change in age at first egg. Quadeer et al. (1977) reported moderately low heritabilities ranging from.5 to.16 from sire component analyses when selection for egg mass (measured from 3 to 4 weeks of age) was practiced in different social environments. Garwood et al. (1978) reported a realized heritability of.6 ±.8 for a restricted index designed to increase egg mass and keep age at maturity constant. me of the more important economic traits in livestock production such as egg production in the domestic fowl and milk yield in dairy cattle are sex limited. This limitation precludes reliable mean other than progeny and sib test of estimating the corresponding genotype in the opposite sex. Consequently, the genetic progress for the improvement of these traits is' less than that for traits measurable in both sexes. This situation often applies to reproductive traits and results in a lower selection intensity for the sex-limited trait. Although little attention has been devoted to the male's role in improving egg production of the domestic fowl (with the exception of progeny testing and selection of males on the basis of sib performance), semen production and egg production appear to be related in White Leghorn as evidenced by selection for high and low fecundity (Jones and Lamoreux, 1942). However, Frankham and Doornenbal (1972) were unable to demonstrate any differences in either semen volume or sperm numbers between males of s selected for increased egg production and males of a control. More recently Nestor (1976; 1977) suggested a possible relationship between semen yield and egg production in the turkey. Total semen volume, however, may not be an accurate measure of total cell mas because about 75% of its volume is a supernatant fluid (Munro, 1938). The establishment of a trait such as egg mass as the criterion for selection to improve egg production efficiency may allow for the identification of a possible direct corollary in the male. If egg mass is defined as the total repro- 1115
1116 MARKS TABLE 1. Means and standard errors for daily egg mass 1 (Periods 1 5) by selection treatment and generation Selected 2 s s 3 41.3 a ±.7 43.5 b ±.5 45.3 a ±.5 47.5 a ±.4 4.5a ±.6 46.a±.5 45.5a ±.6 47.3 a ±.6 4.9 44.8 45.4 47.4 4.9 44.5 43.2 42.7.3 2.2* 4.7** a ' In each generation, means within s with different superscripts are significantly different at P<.1. 1 Egg mass expressed as grams per bird per day. 2 Mean of all selected s. *P«.5. * *P=fi.1 for comparison of mean of selected s vs. mean of control within generation. ductive cell mass produced by the female (composite of both egg size and egg number), then the packed sperm volume (PSV) of semen could represent a corresponding unit in the male, assuming that gonads of both males and females are subject to control by common neuroendocrine systems. The objectives of this study were: a) to determine the actual selection response (both part and whole record) for egg mass when the response is measured for a continuous rather than a limited observation period, and b) to determine the possibility of selecting males for PSV as a corollary to egg mass in the female. c o O E o > Q s EM 1 - Rep 1 EM 2 - Rep 1 EM 1 - Rep 2 EM 2 Rep 2 FIG. 1. Percent deviation from control in daily egg mass (periods 1-5). MATERIALS AND METHODS The NCR Leghorn population maintained at the North Central Regional Poultry Genetics Laboratory served as the base population for this study. In the base generation, 4 females were assigned to individual laying cages at 154 days of age. Birds were housed in mid-september 1975 after floor rearing under natural daylength. The feeding regimen consisted of a 22% protein starter diet from to 6 weeks of age, a 15% protein grower diet from 6 to 22 weeks of age, and a 16% laying diet from 22 to 62 weeks of age. After housing, birds were subjected to a 17 hr photoperiod during the laying cycle. Age at first egg (in days) was recorded with individual body weights at 8, 22, and 62 weeks of age. Management procedure diet and hatching schedules for generation 1 to 3 were the same as those in the base generation. Individual daily egg records were recorded for ten 28-day periods starting at 22 weeks of age with egg weights obtained during the second week of each period. Egg mass for each period was obtained by multiplying the number of eggs per bird times the mean egg weight for that period. Selection was practiced for daily egg mas defined as the total egg mass produced in the first 5 periods (14 days) divided by the days in these periods after females reached sexual maturity (this procedure was followed in an attempt to minimize any influence that sexual maturity may have on egg mass). Daily egg mass was recorded in a similar manner for periods 6 to 1 to estimate changes in egg mass
SELECTION FOR EGG MASS 1117 S, S2 s EM 1 Rep 1 EM 2 Rep 1 EM 1 Rep 2 EM 2 Rep 2 FIG. 2. Deviation of total egg mass from control value by and generation (periods 1 5); x = generation mean. during the second half of the laying year, when selection was practiced for daily egg mass from 154 to 294 days of age. Semen samples were collected from males during the third and fifth laying periods. In the and Sj generation triplicate samples of each male's semen were placed in microhematocrit tubes and centrifuged for 3 min period; packed sperm volume (PSV) percentages were then read on a microcapillary reader. In the S2 and S3 generations duplicate samples were evaluated. The 4 females were randomly allocated to 4 subgroups to establish 4 selected s of 1 birds per. Two replicate s (, Rep 1 and, Rep 2) were established by mating 6 males per with high PSV values with 24 females per with high egg-mass values and 2 replicate s (, Rep 1 and, Rep 2) were established by mating 6 males per with intermediate PSV values to 24 females with high egg-mass values. Hereinafter, denotes the selection treatment in which females were selected on the basis of high egg mas while denotes the selection treatment in which males were selected on the basis of high PSV and females on the basis of high egg mass. The same procedure was used to produce three subsequent generations of these s. Eggs were obtained from the NCR randombred population each generation to produce the nonselected control in order to measure environmental variation across generations. Analyses of variance to compare the effect of selection treatment were based on a mixed model with replications considered as a random effect and selection treatments as a fixed effect. The following model describes the effect of each of the variables studied. Y ijk = ju + R; + Tj + RTy + Eijk where Y ijk is the observation on the k*h individual in the i th replicate (R), and j t h treatment (T). The mean is represented by ji and Ejjk represents random variation among individuals. To measure the effectiveness of selection for egg mas the GLM procedure (Barr et al., 1979) was used TABLE 2. Selection differentials for egg mass by generation Selection differential (g of egg mass/bird/day) Rep 1 Rep 2 Rep 1 Rep 2 Expected Effective Expected Effective Expected Effective Expected Effective S 9.16 8.42 9.5 1.1 8.89 8.68 9.8 8.81 S, 6.13 6.33 6.63 6.75 7.47 7.34 5.44 5.3 S 2 6.79 6.39 6.4 6.94 6.64 6.68 7.49 7.91
1118 MARKS TABLE 3. Heritability estimates for egg mass by generation treatment Rep Regression (daughter on dam) s 3 S -S 3 1 2 1 2.14 ±.27 -.6 ±.23 1.1 ±.26 -.44 ±.24.19.37 ±.23.37 ±.25.86 ±.36.71 ±.22.58 1.2 ±.3.32 ±.18.16 ±.36.36 ±.38.47.51.21.71.21.41.42.38.66.23.42 1 Two times cumulative response divided by cumulative selection differential. to test differences between the pooled mean of the selected s and the mean of the nonselected. RESULTS AND DISCUSSION Comparison of Selection Treatments. Mean daily egg mass during periods 1 to 5 (154 to 294 days of age) increased markedly in both selection treatments (Table 1). Although there was variation among the replicate s (Fig. 1), the only significant difference between selection treatments was in the Si generation in which the egg mass of the treatment was significantly (P«.1) greater than the egg mass of the treatment. As a result of the lack of significance between treatment treatments were pooled for comparisons with the nonselected control (Table 1). Egg mass of the control demonstrated a large environmental increase in the Si generation, which was followed by a trend toward the base generation level in subsequent generations. The egg mass mean of the pooled selected treatments in the S3 generation was greater than Si and S2 generation means (Table 1), indicating a continued response to selection. In the S3 generation mean egg mass of the selected treatments was significantly (P<.1) larger than the mean egg mass of the control. When plotted as deviation from control (Fig. 1), the S3 egg mass of the individual selected s ranged from 5 to 15% above the control level. In the first generation the progress of replicate s in the treatment was greater than the progress of replicate s in the treatment (Fig. 1). Although egg mass of the selection treatment was significantly higher than egg mass of the treatment in generation 1, subsequent data in generation and 3 did not support this observation. Therefore, these data do not support the hypothesis that selection for PSV in males results in greater progress in increasing egg mass than selection for egg mass only in the female. Realized Selection Responses. The increase TABLE 4. Means and standard errors for daily egg mass 1 (Periods 6 1) by generation Selected* s Sa s 3 37.2a ± 1.5 43.5a ±.9 42.9 a ±.8 44.93 ±.8 3 7.4 a ± 1.2 43.9 a ± 1.1 44.1 a ±.9 44.1 a ±.9 37.3 43.7 43.5 44.5 37.3 4.8 39.3 37.5 2.9 4.2* 7.** a In each generation, means within s with same superscripts are not significantly different at P>.1. ' Egg mass expressed as grams per bird per day. 2 Mean of all selected s. *P«.5. * *P<.1 for comparison of mean of selected s vs. mean of control within generation.
SELECTION FOR EGG MASS 1119 TABLE 5. Comparison of egg mass/bird by period and generation Period Rep 1 Rep 2 Egg mass(g) per bird Rep 1 Rep 2 S3 1-5 6-1 1-5 6-1 1-5 6-1 1-5 6-1 5185 521 5842 6183 659 626 6345 6346 in total egg mass per bird during periods 1 to 5 of the individual selected s as a deviation from the control varied from 544 to 956 g in the S3 generation (Fig. 2). The mean deviation of the selected s was 171 g in generation Si, 364 g in generation S 2, and 763 g in generation S3. The selection response in these s appears ar, without evidence of reduced progress accompanying increased generation number. The response to selection for egg mass in this study is in general agreement with the results of Garwood et al. (1978). Although our response after a single generation of selection was not as great as that reported by Garwood et al, (1978), both studies indicate a rapid and large initial response to selection for egg mass. Our result however, do not support those of Quadeer et al. (1977). It must be noted, however, that selection in the current study was based on a continuous scale (egg mass from 154 to 294 days) while selection for egg mass by Quadeer et al. (1977) was confined to measurements from 3 to 4 weeks of age; therefore, the selection variables were not entirely the same. Also, the base population used in this study and the population used by Garwood et al. (1978) originated from the same control population, whereas the base population used by Quadeer et al. (1977) was of a different origin. Selection differentials (expected and effective) for daily egg mass were similar across s (Table 2). Differences between the expected (unweighted) and effective (weighted) selection differentials were small and not significantly different. Therefore, natural selection has apparently exerted little influence on the 498 4424 5574 5895 5887 5742 617 6184 4984 578 656 6254 6142 626 636 6379 595 521 5919 5913 5821 637 5948 5924 5677 5671 5613 5449 543 5162 response to artificial selection in these s. The size of the selection differentials was reduced after the first generation in all s (Table 2). Realized heritabilities for daily egg mas obtained by dividing the total corrected responses (controls used to correct for environmental variation) by the cumulative selection differentials of the selected ranged from.23 to.66 with a mean of.42 (Table 3). Regression of daughter on dam heritabilities for egg mass by and generation demonstrated considerable variation across generations. This variation may have been due to the small EM 1 - Rep 1 EM 2-Rep 1 EM 1 Rep 2 EM 2 Rep 2 s FIG. 3. Deviation of total egg mass from control value by and generation (periods 1 1); x = generation mean.
112 MARKS TABLE 6. Mean age (days) and standard errors for age at first egg Selected 1 s S, S3 1623 ± 1 158 a ± 1 161 a ± 1 167a ± 1 161 a ±1 16 l a ±1 16 a ±1 162b ±1 162 16 161 165 162 164 165 167-4* -4* -2 a b ' In each generation, means within s with different superscripts are significantly different at P<.1. 1 Mean of all selected s. *P<.5. population size of the selected s. Mean values within s across generation were, however, not greatly different from realized heritability estimates for egg mass (Table 3). Drift variances were not calculated since it is unlikely that meaningful estimates would be obtained as a result of the limited replications (2) and small population sizes involved in this study. Correlated Traits. Mean daily egg mass of the selected s during periods 6 to 1 (294 to 434 days of age) also demonstrated a definite increase across generations (Table 4). In the S3 generation the mean egg mass of the individual selected s was significantly (P<.1) higher than the egg mass of the control. The standard errors of mean egg mass (Table 4) were considerably larger than the standard errors associated with egg mass values during the first 5 periods (Table 1). Deviations in egg mass of the mean of selected s from that of the control were greater during periods 6 to 1 than during periods 1 to 5. Although mean daily egg-mass values were larger during periods 1 to 5 (Table 1) than periods 6 to 1 (Table 4), the total egg mass produced on a per-bird basis was generally similar for these two observation periods (Table 5). This similarity may be due to the large egg size accompanying the dec in hen-day egg production during periods 6 to 1. It is also possible that more total "laying days" were available for periods 6 to 1 than for periods 1 to 5 because all birds were not sexually mature at the beginning of period 1. Total egg mass per bird in the selected s as a deviation of the control varied from 137 to 2173 g, with a mean of 181 g, in the S 3 generation (Fig. 3). These data indicate that when selection was practiced for part-record daily egg mas an additional increase equal to or superior to the part-record increase was exhibited during the second half of the laying year (Fig. 2 and 3). This favorable influence is TABLE 7. Mean 62-week body weights (g) and standard errors by generation Selected 1 s s S3 199 a ± 22 1914 a ± 24 1934a + 19 192lb ± 19 1969 a ± 22 1988 a ± 21 1989 a ± 22 287 a ± 23 1939 1952 1962 24 1939 25 1912 1853 a b ' In each generation, means within s with different superscripts are significantly different at P<.1. 1 Mean of all selected s. **P<.1 for comparison of mean of selected s vs. mean of control within generation. -53 5 151**
SELECTION FOR EGG MASS 1121 g Prod. UJ Day!R 85 1 82 81-76 -1 74 - ~- $/ en- 84-83- 8-79- 78-77- 75- z//^u / // /// //// / / / / EM 1 Rep 1 EM 2 Rep 1 EM 1 Rep 2 EM 2 Rep 2 / 1 1 s 3 s FIG. 4. Percent hen-day egg production by and generation (periods 1 5). apparently due to combined selection pressure on both egg size and egg number. Although age at sexual maturity was similar between the and treatment with the exception of the S3 generation, the mean age at sexual maturity of the selected s was 4 days earlier than the control in the Si and S2 generations (Table 6). However, the mean age at sexual maturity of the control was not significantly different from that of the mean of the selected s in the S3 generation. These data indicate that selection for daily egg mass did not greatly alter the age at sexual maturity. Mean body weights at 62 weeks of age appeared to be similar between the and treatments in all generations except the S3 generation. The S3 generation mean body weights of the selected s were significantly (P<.1) higher than body weights of the control (Table 7). Additional data are necessary to confirm this apparent trend and determine whether selection for daily egg mass results in an increase in adult body size. Percent hen-day egg production (periods 1 to 5) of replicate s in the and treatments increased substantially across generations (Fig. 4). Egg production of selected s varied from 4.6 to 6.5 percentage points above that of the control in the S3 generation. The mean egg production of selected s during periods 6 to 1 was also superior to that of the control but the differences were not significant (Table 8). Although and mean egg weights at 4 weeks of age were not significantly different (Table 9),, Rep 1 and, Rep 1 egg weights were approximately 4 g larger than those of, Rep 2 and, Rep 2 s in the S 3 generation. The mean egg weights of the selected s were larger than the mean control egg weight in all generations (Si to S3); however, the difference was significant only in the S 3 generation. These data indicate that age at sexual maturity, 62-week body weight, hen-day egg production, and 4-week egg weight responses were similar for the and selection treatments. Results of this study indicate that the negative relationship between egg weight and egg production can be circumvented, at least to some degree, by selection for egg mass. The realized heritabilities observed for egg mass indicate that selection for this trait may be advantageous for improving the reproductive efficiency of stocks of low productivity prior to TABLE 8. Mean percentage hen-day egg production and standard errors (275 to 494 days) by generation Selected 1 s s. s 3 72.* ± 2. 72.9 a ± 1. 72.3 a ± 1. 75.4 a +.8 69.6 a ± 1.3 76.7*± 1.1 74.5* ±.8 74.8* ± 1. 7.8 74.9 73.4 75.1 7.8 73.9 72.8 71.9 1..6 3.2 In each generation, means within s with same superscripts are not significantly different at P>.1. 1 Mean of all selected s.
1122 MARKS TABLE 9. Mean egg weight (g) and standard errors at 4 weeks by generation Selected 1 s s s 3 58.oa +.3 59.9 a +.4 6.2a +.3 6.9 a +.3 57.8a +.4 6.9 a ±.3 6.7a +.4 6.3a ±.4 57.9 6.4 6.5 6.6 57.9 6. 59.4 57.9.4 1. 2.7** In each generation, means within s with same superscripts are not significantly different at P>.1. 1 Mean of all selected s. * *P<.1 for comparison of mean of selected s vs. mean of control within generation. introduction into improved gene pools. It is also possible that selection for egg mass may provide a method to rapidly improve egg production traits of stocks which possess desirable characteristics such as resistance to certain diseases. Unfortunately, selection for PSV in males does not appear to be associated with selection for increasing egg mass. Additional experimentation is needed to identify male parameters that will allow for greater selection progress in sex limited traits. REFERENCES Abplanalp, H 1956. Selection procedures for poultry flocks with many hatches. Poultry Sci. 35: 1285-134. Barr, A. J., J. H. Goodnight, J. P. Sail, W. H. Blair, M. D. Chilko, 1979. SAS user's guide. Raleigh, NC 2765. Bohren, B. B., 197. Genetic gain in annual egg production from selection on early part-record. World's Poultry Sci. J. 26:647-657. Clayton, G. A., and A. Robertson, 1966. Genetics of changes in economic traits during the laying year. Brit Poultry Sci. 7:143-151. Craig, J. V., D. K. Biswa and H. K. Saadeh, 1969. Genetic variation and correlated responses in chickens selected for part-year rate of egg production. Poultry Sci. 48:1288-1296. Frankham, R., and H. Doornenbal, 1972. Semen characteristics of s selected for increased part-record egg production. Poultry Sci. 51: 1468-1469. Garwood, V. A., P. C. Lowe, and B. B. Bohren, 1978. A replicated single generation test of a restricted selection index in poultry. Theor. Appl. Genet. 52:227-231. Hick A. F., Jr., 1958. Heritability and correlation analysis of egg weight, egg shape and egg number in chickens. Poultry Sci. 37:967-975. Hick A. F., Jr., 1963. A study of egg mass and biomass and of their components in S. C. White Leghorns. Poultry Sci. 42:1277. Jerome, F. N., C. R. Henderson, S. C. King, 1956. Heritabilitie gene interactions and correlations associated with certain traits in the domestic fowl. Poultry Sci. 3 5:995-113. Jone D. G., and W. F. Lamoreux, 1942. Semen production of White Leghorn males from strains selected for high and low fecundity. Poultry Sci. 21:173-184. Kinney, T. B., Jr., 1969. A summary of reported estimates of heritabilities and of genetic and phenotypic correlations for traits of chickens. USDA Agr. Handbook No. 363. Munro, S. S., 1938. The effect of dilution and density on the fertilizing capacity of fowl sperm suspensions. Can. J. Res. Sec.D. Zool. Sci. 16:281-299. Nestor, K. E., 1976. Selection for increased semen yield in the turkey. Poultry Sci. 55:2336 2338. Nestor, K. E., 1977. The influence of a genetic change in egg production, body weight, fertility or response to cold stress on semen yield in the turkey. Poultry Sci. 56:421-425. Quadeer, M. A., J. V. Craig, K. E. Kemp, and A. D. Dayton, 1977. Selection for egg mass in different social environments. 2. Estimation of parameters in selected populations. Poultry Sci. 56:1536 1549. Waring, F. J., P. Hunton, and A. E. Maddison, 1962. Genetics of a closed poultry flock. I. Variance and covariance analysis of egg production, egg weight and egg mass. Brit. Poultry Sci. 3:151-16.