The weight of an organism at any point in its life can be considered a function of its genetic and environmental histories. Sexual dimorphism for body weight varies among gallinaceous birds, with male birds being heavier than female birds (e.g., chickens) or female birds being heavier than male birds (e.g., Japanese quail). In both cases, the degree of dimorphism at a specific age differs depending on the developmental and physiological aspects at the time of comparison. Furthermore, Rensch’s rule for sexual dimorphism for body weight in wild birds has been compromised in the case of the domestic chicken (Remeš and Székely, 2010).

Chickens were domesticated during the Neolithic period and have been amenable to artificial selection for various traits, including those focusing on sport, show, ceremony, and foods (Smith and Daniel, 1975). Chicken meat as food was primarily a result of traits selected for other purposes until the past century, when specialized breeding programs were designed for rapid juvenile growth (Siegel, 2014, 2023). This development was genetically facilitated by the availability of foundation stocks (Hanke et al., 1974), short generation intervals, and moderate heritability for juvenile body weights. Crosses between breeds and lines within the breeds allowed the breeders to protect their commercial products. The number of days to market weight has reduced dramatically (Havenstein et al., 2003; Zuidhof et al., 2014), and the age at market weight is currently measured in days instead of weeks. Because of sexual dimorphism in juvenile body weights, male and female birds attain their market weights at different ages; hence, sex-separate rearing is not uncommon and has become more popular. High-speed imaging and artificial intelligence tools have allowed autosexing of chicks at hatching through differences in the lengths of the primary and covert wing feathers, thereby replacing tedious human labor. Specific matings are needed, which are the result of sex-linked genetic effects (Chambers et al., 1994). Correlated responses, including increased feed intake (Carney et al., 2022), are associated with rapid juvenile growth and heavier body weights at younger ages. To address the potential cardiac, skeletal, and metabolic issues associated with excessive feed intake in the body weights of the breeders (adults), feeding programs to achieve the target body weights are readily available in management manuals provided by commercial distributors for their particular stocks.

Although the heritabilities for juvenile body weight in chickens can date back to several decades, as seen in the 176 summarized cases (Siegel, 1962), reports on the inheritance of sexual dimorphism per se are sparse. Sexual dimorphism for body weight in chickens can be viewed as the absolute or proportional differences between male and female birds. Thus, while the former may increase or decrease, the latter can remain constant or change. The inheritance of differences between the sexes for a common trait is confounded by multiple factors. Merritt (1966) and Mignon-Grasteau et al. (1998) provided valuable data and insights into the genetics involved as well as low heritability. Subsequently, Adedibu and Ayorinde (2011) reported the genetic differences in male to female body weight ratios at 4 and 8 weeks in one stock but only at 8 weeks in another stock. Maniatis et al. (2013) reported a genetic correlation of 0.91 for body weight between the sexes at 35 days, with heritabilities of 0.04 for the body weight differences and ratios. Although these results allow inference of little genetic variation for sexual dimorphism for juvenile body weights, differences among various breeds are well documented (American Poultry Association, 1947).

Starting from a common founder population of White Plymouth Rock (WPR) chickens, we developed two lines by divergent selection for a single trait, namely, body weight at 8 weeks (56 days) of age. After 67 generations of divergent selection, the two lines differed by more than 10-fold without any overlap in the body weight at selection age. Herein, we report the sexual dimorphisms for body weights at 4 and 8 weeks of age across the 67 selected generations as well as the founder population of these two lines.

In both selected lines, the male chickens were consistently heavier than the female chickens at 4 and 8 weeks of age (Figures 1, 2). At 4 weeks, although the generational pattern was consistent for all generations of HW chickens, the degree of difference varied. For LW chickens, although the dimorphism followed the same pattern (males > females), some generations had slight differences between the male and female chickens. By 8 weeks, the magnitudes of dimorphism were more accentuated, with the male and female chickens consistently reflecting the within-generation environments.

On a relative basis (with lower percentage indicating greater dimorphism), the greater dimorphism in the founder populations at 8 weeks (80.8%) than 4 weeks (88.9%) continued throughout the 67 generations in both HW and LW chickens (Table 1; Figure 3). When viewed across the 67 generations, there were generations when the relative sexual dimorphism for body weight appeared to increase or decrease. This “self-correction” pattern was observed at both ages, with the variation being greater in LW than HW chickens and more evident in the later than earlier generations (Table 1). The magnitude of variation in the HW chickens was consistently greater at 4 weeks than 8 weeks. In the LW chickens, the variation at 8 weeks was greater than that at 4 weeks, except for the F₁₃–F₂₃ and F₄₆–F₅₆ periods.

The same data were used in both the correlation and regression analyses because the evidence was unclear for cause–effect associations; the results obtained from these analyses were also similar. At 4 weeks, the regressions for both lines across all generations were essentially superimposed, while they were different at 8 weeks (Figure 3). Quantifications of these graphic patterns are provided in Tables 2, 3 for 4 and 8 weeks of age, respectively. At 4 weeks, there were generational periods wherein the dimorphism increased or decreased significantly only during the first half of this long-term selection experiment (Table 2). Although the generational interval differences were evident in both lines at 8 weeks, the patterns between them were different (Table 3). Overall, there was a significant reduction in dimorphism in the HW group but not in the LW chickens. Moreover, with the exception of the F₂–F₁₂ period, the standard errors of the regressions were one order of magnitude greater in the LW than HW chickens. These contrasting results between the two lines and ages are also seen in Figure 1.

The body weight of an organism at any point in time is a function of its genetic and nongenetic factors as well as past and present interactions. In multigenerational studies such as the one reported herein, epigenetic issues may also be relevant. The process across generations is dynamic, and social issues can be a factor when both male and female chickens are involved; however, these were precluded by our reproductive design. The chronological and physiological ages may differ, and these were documented early for the HW and LW lines (Katanbaf et al., 1988); prior to and since then, we have been aware of this dichotomy, as evidenced in the developmental stages for numerous traits (Washburn and Siegel, 1963; Siegel and Dunnington, 1985, 1987; Zelenka et al., 1988; Jambui et al., 2017). Thus, studying the inheritance of sexual dimorphism for body weight in chickens is a complex process.

As noted in the introduction, there is a vast amount of literature on the heritabilities for juvenile body weights for male and female chickens. Although these values are population specific, they may be similar (such as 0.28 for male and 0.29 for female chickens at 35 days) (Maniatis et al., 2013) or dissimilar (0.28 for male and 0.43 for female chickens at 8 weeks of age) (Mignon-Grasteau et al., 1998). The realized heritabilities at selection age for the first four generations of our experiment (male: female) were 0.23:0.21, 0.33:0.32, 0.35:0.27, and 0.31:0.28, respectively, with a small but consistent pattern and an unweighted mean of 0.30:0.27 (Siegel, 1962). These estimates are consistent with molecular analyses that showed that the 8-week body weights in these lines are influenced by many genes, each with small effects (Lillie et al., 2018). However, these findings do not address dimorphism per se, as such estimates involve two sets of individuals (males and females). The fact that the published heritability estimates of dimorphism per se of 0.08 (Mignon-Grasteau et al., 1998) and 0.04 (Maniatis et al., 2013) for juvenile body weights were low is not surprising, nor is the fact that the genetic correlation for body weight between the sexes was 0.91 (Maniatis et al., 2013). As noted by Merritt (1966), the magnitude of difference in heritabilities for male and female chickens is based on the contributions of the sire versus those of the dam per the sex chromosome, i.e., the sire is homogametic and dam is heterogametic. This means that the sex-linked genes are paired in male but not female progeny. Hence, the genetics of the sex chromosome is especially relevant when considering the inheritance of sexual dimorphism, particularly when comparisons involve populations selected in opposite directions from a common founder (such as the Virginia high and low weight lines). This is especially relevant when the genes at certain loci are at different frequencies and may be lost over time. Moreover, as the lines became more diverse over generations, the maternal and linkage effects may have become more relevant to the complexity of the differences in the 4- and 8-week body weights.

The Virginia lines are the WPR chickens, which is a meat-type breed (Hanke et al., 1974) first recognized by the American Standard of Perfection in 1888 (American Poultry Association, 1947). The WPR breed is a composite of the Cochin, Light Brahma, Black Minorca, Black Java, Langshan, and Dominique breeds (American Poultry Association, 1947; Guo et al., 2019). An analysis of the genealogy of the Virginia lines (Guo et al., 2019) showed that the overall proportions of autosomal contributions from the contributing breeds were similar. These contributions (HW:LW) were Cochins (30:32), Dominique (33:30), Black Java (26:27), Langshan (7:4), Light Brahma (4:7), and Black Minorca (0:0). However, when viewed at the individual chromosome level, there were major differences; this was evidenced by the contributions of the founder WPR breeds toward the sex chromosomes. For the W chromosome, the contributions came from just two of the founders, i.e., 49% Black Java and 51% Cochin, whose sexual dimorphism for juvenile body weights were 81% and 78%, respectively. In contrast, the percentage contributions to the Z chromosome (HW:LW) were from the Dominique (46:52), Black Java (28:21), Cochin (10:13), Light Brahma (11:9), Black Minorca (4:3), and Langshan (1:2) breeds, whose contributions to the dimorphism of juvenile body weights were 67%, 81%, 78%, 80%, 87%, and 81%, respectively (American Poultry Association, 1947). The changes in the contributions of these breeds toward the sex chromosomes occurred over many generations and were likely responsible for the cross-generational fluctuations that ultimately resulted in two and five of the original breeds contributing to the sex chromosomes.

Over the course of the 67 generations of selection for the high and low 8-week body weights, dynamic changes were noted to occur at the molecular and phenotypic levels for the males and females, with reductions in relative dimorphism at 4 and 8 weeks with selection for the HW line but not the LW line. Although the relative contributions from the founder WPR breeds were similar in both lines, there were large differences in the contributing stocks between the HW and LW lines in terms of the sex chromosomes. Across multiple generations, homeostatic mechanisms will often come into play, and the degree of dimorphism at a fixed age will be population-specific. It will be very difficult to genetically modify sexual dimorphism within a particular population using the conventional tools available at present. In breeding programs where the final product involves several closed populations, combining abilities may differ among the evaluated populations. The long-term divergent selection for 8-week body weight in the present study revealed the dynamic biological complexity of a trait that is ironically easy to measure.