Introduction

Fisher’s fundamental sex ratio theorem (Fisher 1930) offers a general explanation for the prevalence of 1:1 sex ratios among animals. Trivers and Willard (1973) noted, however, that the litter sex ratio of a given female is expected to depart from unity where mothers bias offspring production towards the sex that will maximize their reproductive payoff with the investment of additional resources. In polygynous species, male-biased litters prove advantageous, as Darwin (1871) recognized in suggesting that males potentially have multiple reproductive partners in any given breeding season, and thus, pay a higher fitness dividend than female offspring of similar quality. Williams (1979) subsequently refined Trivers and Willard’s hypothesis, noting that where optimum parental investment in male versus female offspring differs, and females are capable of producing more than one offspring at a time, both sex ratio and litter size must be regulated to match maternal capability with litter cost. Thus, maternal age, physiology and access to resources should affect both litter size and sex ratio where adaptive sex allocation occurs (Williams 1979).

The basic predictions of the Trivers-Willard hypothesis enjoy considerable empirical support (Austad and Sunquist 1986; Berube et al. 1996; Cameron 2004; Clark et al. 1990, 1991). In Red Deer, males have up to three times the reproductive success of females over their lifetime (Pemberton et al. 1992; Rose et al. 1998). Mothers able to invest more resources in their offspring tend to invest more in male offspring than females, and males are more affected by additional investment than females (Austad and Sunquist 1986; Clutton-Brock et al. 1984). Birth weight is correlated with lifetime reproductive success in sexually dimorphic species such as Red Deer, with heavier males siring more offspring (Kruuk et al. 1999). In social species, males in better body condition may also be better able to compete for dominance and secure mates (Alberts et al. 2006).

Alternatively, the local resource competition hypothesis suggests that it would prove beneficial for mothers in poor condition to invest in male offspring (Clark 1978). This hypothesis posits that in order to avoid competition with progeny for limited resources, mothers should produce litters biased towards the dispersing sex. In many mammals, dispersal is male–biased (Michener and Michener 1977), and local resource competition has been implicated as a factor promoting biased sex ratios (Caley and Nudds 1987; Clark 1978; Cockburn et al. 1985; Johnson 1986). Consistent with this notion, brushtail possums (Trichosurus vulpecula) in food-rich areas with high population density experience a relatively low per capita availability of female den sites, and produce male-biased offspring sex ratios (Johnson et al. 2001). The potential for competition with philopatric daughters, however, diminishes as dams age, resulting in a shift in offspring sex ratio toward the production of females as breeding females grow older (Isaac et al. 2005). That effect of maternal age is offset, however, by maternal body condition, as females in good body condition, or that showed improved body condition between years, were more likely to produce male offspring (Isaac et al. 2005). It is evident then, that independent forces interact in selecting for optimal female reproductive investment, and that the expression of adaptive sex allocation will be contingent not only on the availability of resources, but on variation in body condition over a female’s lifetime.

Manipulation of offspring sex ratio can occur prior to conception (Oddie 1998; Pike 2005; Roche et al. 2006), during gestation (Austad and Sunquist 1986; Cameron 2004; Myers et al. 1985; Owusu et al. 2010) or after gestation, via differential provisioning, neglect or culling of the non-preferred sex (Fairbanks and McGuire 1995; McClure 1981; Moses et al. 1998; Tait 1980; Voland 1984). Post-conception adjustment of offspring sex ratio requires reduction of the litter or clutch size (Gedir and Michener 2014; Helmreich 1960; Krackow 1997), resulting in a litter size–sex ratio trade-off.

Field studies of Richardson’s ground squirrels (Urocitellus richardsonii) have revealed that increasingly male-biased offspring sex ratios are accompanied by reduced litter size at juvenile emergence, and increased fecal glucocorticoid metabolite concentrations (Ryan et al. 2012, 2014) as predicted by Cameron’s (2004) glucose metabolism hypothesis. Gedir and Michener (2014), however, found no relationship between litter size and sex ratio in their analysis of 24 years of litter data from Richardson’s ground squirrels, suggesting a Mendelian mechanism of sex allocation (Williams 1979) rather than adaptive modification of offspring sex by breeding females. Gedir and Michener (2014) attribute the findings from the single-year studies of Ryan et al. (2012, 2014) to sampling bias implicit in short-term ecological studies (O’Neill et al. 1986; Maxwell and Jennings 2005; Sergeant et al. 2012; Sullivan et al. 2013), where litter size–sex ratio trade-offs may occur by chance alone (Cameron 2004; Clutton-Brock and Sheldon 2010). It is equally plausible, however, that any difference in the expression of adaptive sex allocation among female Richardson’s ground squirrels is accounted for by variation in female body condition between study sites as affected by variation in resource availability or differences in the life history of litter-producing females.

To explore whether evidence of adaptive sex allocation is conditional upon variation in environmental conditions impacting resource availability or life history, we tested for a relationship between litter size and sex ratio for all dams with known litters from long-term, multigenerational data within the same Winnipeg, MB zoo-based population of Richardson’s ground squirrels where trade-offs between litter size and sex ratio were reported based on independent single–year studies (Ryan et al. 2012, 2014). Failure to detect a relationship between litter size and sex ratio in our multigenerational data would support Gedir and Michener’s (2014) contention that the litter size–sex ratio trade-offs reported by Ryan et al. (2012, 2014) represent an artefact of sampling bias within a species where random sex allocation prevails. Further, by explicitly testing for effects of female age, parity and variation in resource availability to the expression of litter size–sex ratio trade-offs, we test whether the inconsistent documentation of these trade-offs reflect conditional adaptive sex allocation.

Materials and methods

Ten-year data set

Data were obtained from a Richardson’s ground squirrel population at the Assiniboine Park Zoo in Winnipeg, MB (49° 52′ N, 97° 14′ W), which has been the subject of intensive study from 2004 to 2016. Each year, live-trapping occurred during the three–week breeding season from mid-March through early April to the immergence of adults into hibernation by the end of July. Squirrels were captured using live traps (Tomahawk Live Trap Co., Tomahawk, WI) baited with peanut butter (No Name™ Brand, Loblaw Companies Ltd., Toronto, ON). Trapped squirrels were transferred into a cloth bag for weighing to the nearest 5 g with a Pesola® spring balance (Pesola AG, Schindellegi, Switzerland), then handled by gripping the body around the scapulae using a leather-gloved hand. Upon first capture, squirrels were marked for permanent identification by affixing a numbered metal ear tag (National Band and Tag Company, Monel no. 1, Newport, KY) through the right, left or both pinnae. Additionally, a unique identification mark was applied using a paint brush to the squirrel’s dorsal pelage using black hair dye (Clairol Hydrience 52S; Pearl Black, Stamford, CT) for identification of individuals from a distance and to minimize the need to handle the squirrels every time they were live-trapped. Newly emerged juveniles were assigned a sequential number, yearling squirrels were assigned a dot above their number from the previous season, and adults of 2+ years or unknown age were assigned a unique mark.

Within the 13 years, 10 years of data documenting litter size, sex ratio, dam identity, and the number of prior litters produced by each dam were available. This resulted in a sample of 420 litters from 301 distinct dams, ranging in age from 1 to 6 years old and producing between 1 and 4 litters over the years they were trapped. All statistical tests were conducted using R statistical software (R version 3.4.0, R Core Team 2013). Data were tested for normality using the Anderson–Darling normality test within the R package kSamples (Scholz and Zhu 2017). Model II major axis regressions (Ricker 1973; Laws and Archie 1981) were employed to test for potential relationships between litter size and sex ratio within the first, second, and third litters (parity) of each mother pooled across all years, as well as on an annual basis, to test for any apparent relationships within each calendar year. Additionally, regressions were run on litter size and sex ratio of litters from 1-, 2-, and 3-year-old mothers known to have had at least three observed litters over the sample period, to test for any apparent relationship within specific dam age cohorts. Both parity and maternal age are expected to influence the expression of adaptive sex allocation as yearling females complete their growth to adulthood during their initial gestation and lactation (Michener 1989a; Dobson and Michener 1995), and thus may be energetically stressed relative to adult females. Further, the reduced residual reproductive value of older females that have produced more litters could modify both the ability to produce, and relative value of male versus female offspring (Williams 1966; Stearns 1989). Further regressions were performed to explore relationships between litter size and litter sex ratio before versus after the construction of the zoo’s Australian Walkabout exhibit in 2013 for dams in that area, with an additional two-sample Wilcoxon rank-sum test comparing annual litter sex ratios before versus after construction. These contrasts were deemed important to consider as the establishment of that exhibit introduced foods including carrots, beets, and various greens for the enclosure’s red kangaroos (Macropus rufus) and emus (Dromaius novaehollandiae) but that were accessed ad libitum by squirrels in the area, potentially affecting dam body condition.

Beyond providing supplementary food, the construction of the Australian Walkabout exhibit constituted one of four major construction events directly on, or adjacent to areas occupied by ground squirrels in 2012, 2013, 2014 and 2016. In that anthropogenic activity may induce stress (O’Regan et al. 2010; Schell et al. 2013; Van Meter et al. 2009), we contrasted litter sex ratios of dams in years with versus without construction to test whether presumptive stress associated with construction activity affects the expression of adaptive sex allocation.

Major axis regression analyses were performed using the R package lmodel2 (Legendre 2014) and the two-sample Wilcoxon rank-sum test was run using the wilcox.exact function in the R package exactRankTests (Hothorn and Hornik 2017) as well as the qnorm function to calculate the associated Z value. A Binomial test was performed using the R (2013) binom.test function to test for alternation of male- versus female-biased average litter sex ratios within the population between years of study to determine if dams alternate sex biases annually, producing more of what had been the rarer sex in the previous year. For all tests, results were deemed significant at an alpha value of 0.05, and reported as mean ± SE unless otherwise stated. All plots were done using the ggplot2 package in R (Wickham 2009) and labeled in CorelDRAW Home & Student X6 (CorelDRAW 2012).

Placental scar counts

In addition to the analysis of 10 years of litter data, 64 dams were humanely euthanized by Assiniboine Park Zoo staff as part of a ground squirrel control program and their uteruses were dissected out for placental scar or embryo counts in 2016. Fifty of these squirrels were euthanized by the pest control officer at the Assiniboine Park during early gestation, precluding the collection of data on litter size or sex ratio at juvenile emergence for those females. Necropsy of these individuals allowed us to assess whether placental scars appeared alongside viable embryos, as would be expected if females manipulate litter sex ratio via the selective abortion or resorption of embryos of the non-preferred sex. The remaining 14 dams were from the study population within the zoo that had been manipulated as part of our failed attempt to experimentally manipulate circulating blood glucose levels during gestation, and for which post-weaning litter size and sex ratios were known. Following euthanasia, the uterine horns were dissected out of each female under veterinary supervision, and placental scar counts were obtained. Scar counts from the 14 dams were compared to litter counts at emergence from the natal burrow via a Wilcoxon signed-rank test using the wilcox.exact function from the exactRankTests package (Hothorn and Hornik 2017) and the qnorm function for the associated Z value to see if the count at emergence was representative of the litter size during gestation.

Results

Ten-year data set

From 2006 through 2016, the average percentage of males produced by dams was 49.1%, ranging from a minimum of 43.2% to a maximum of 55.8% in any given year (Table 1). Of 409 known litters trapped at juvenile emergence, seven had only a single offspring. Population-wide offspring sex ratios alternated between male-versus female-biased in successive years 5 times over the 10-year period, which does not depart from the number of alternations expected by chance (Binomial test, P = 1.0). Average litter sex ratios of dams in their first and second breeding seasons approximated parity, while the average percentage of males produced in a dam’s third breeding season dropped to 36.8% (Table 2).

Table 1 Annual Richardson’s ground squirrel litter sex ratios as proportion male
Full size table
Table 2 Average Richardson’s ground squirrel litter size, litter sex ratio, and number of litters by dam parity from 2006 to 2016
Full size table

Major axis regressions of litter size versus sex ratio for first, second, and third litters pooled across years detected a statistically significant relationship for initial litters only (R2 = 0.011, P = 0.033; R2 = 0.003, P = 0.301; R2 = 0.023, P = 0.253 for first, second, and third litters respectively; Fig. 1). The relationship between litter size and sex ratio varied among subsequent litters born to dams, with first litters becoming more female-biased with increasing litter size (Fig. 1a), second litters showing a similar trend (Fig. 1b), and third litters becoming more male-biased with increasing litter size (Fig. 1c). No significant relationships between litter size and sex ratio were evident, however, among the smaller sample of known-age dams that were considered separately at one-, two-, or three-years of age (R2 = 0.001, P = 0.441; R2 = 0.028, P = 0.253; R2 = 1.460e−4, P = 0.492 respectively; Fig. 2).

Fig. 1
figure 1

Major axis regression comparing litter sex ratio, expressed as the percentage of male offspring, and litter size in a first litters (n = 301), b second litters (n = 83), and, c third litters (20 litters total) of all dams. The red line represents the line of best fit, data points are scaled by size to reflect the number of tied observations

Full size image
Fig. 2
figure 2

Major axis regression comparing litter size and sex ratio, expressed as percentage of the litter that was male, in litters of a 1-year-old (n = 18), b 2-year-old (n = 19), and, c 3-year-old dams (n = 18) for which at least three litters were observed on-site across all years. Red line represents the line of best fit, data points are scaled by size to reflect the number of tied observations

Full size image

Major axis regression of all litters by year for each of the 10 years of litter sex ratio data revealed statistically significant relationships between litter size and sex ratio in 2013 and 2016, where increasing litter size was associated with a decreasing proportion of males within litters (Table 3). While not significant, that same trend toward a decreasing proportionate representation of males with increasing litter size was evident in 7 of the 10 years for which data were available (Table 3). Considering only litters of primiparous dams, a significant relationship between litter size and sex ratio was evident only in 2016, though that relationship approached statistical significance in 2007 as well (Table 4). Further, in all years except 2015, litters of primiparous dams became increasingly female-biased as litter size increased (Table 4).

Table 3 Major axis regression line equations comparing litter size and sex ratio as percentage of the litter that is male with corresponding R, R-squared, and significance values for all Richardson’s ground squirrel litters by year in which they were observed
Full size table
Table 4 Major axis regression line equations comparing litter size and sex ratio as percentage of the litter that is male with corresponding R, R-squared, and significance values for litters of primiparous dams by year of trapping
Full size table

We detected no association between years with major construction events and sex-biased litters, with litters in 2012 averaging 55.6% male and the remaining three years remaining around parity (Table 1). Considering only litters of primiparous dams, however, two of the four years having the greatest physical disruption also had the most male-biased sex ratios (60.7% in 2012 and 56.2% in 2016; Table 1). Of all years studied, only 2013 and 2016 showed statistically significant litter size–sex ratio trade-offs (Tables 3 and 4), coincident with major on-site disturbances.

Trade-offs between litter size and sex ratio among dams occupying the Australian Walkabout enclosure escaped statistical significance both before (R2 = 2.896e−4, P = 0.444; Fig. 3a) and after its construction (R2 = 0.009, P = 0.169; Fig. 3b), though the proportionate representation of females within litters increased with increasing litter size after construction (Fig. 3b). Despite that trend, litter sex ratios of dams in the walkabout did not differ between the pre- and post-construction periods (W = 13, Z = 1.368, P = 0.914).

Fig. 3
figure 3

Major axis regression comparing sex ratio, expressed as the percentage of male offspring, and litter size for all litters occurring in the picnic area a prior to construction of the Australian Walkabout (n = 70), and, b after construction of the Australian Walkabout (n = 105). The red line represents the line of best fit, data points are scaled by size to reflect the number of tied observations

Full size image

Placental scar counts

A total of 481 implantations were counted, including 233 embryos and 248 scars. The average number of scars or embryos observed was 8.017/dam, and ranged from 5 to 14 total implantations. In the 50 females euthanized during early gestation, scars and embryos were never observed together within dams. Among the 14 females euthanized subsequent to juvenile emergence, the number of placental scars (7.643 ± 0.589) exceeded the number of juveniles at emergence (6.429 ± 0.581), but not significantly (W14 = 60.5, Z = − 1.254, P = 0.105).

Discussion

Our 10-year Richardson’s ground squirrel dataset reveals the predicted litter-size–sex ratio trade-offs where adaptive sex allocation involves the elimination of offspring of the non-favored sex. The proportionate representation of males within litters decreased significantly with increasing litter size for primiparous dams pooled across years and was evident in all individual years except 2015. The consistency of this trend among primiparous females suggests that selection has favored the evolution of a mechanism promoting adaptive sex allocation via elimination of developing females, resulting in the production of smaller male-biased litters where conditions favor the production of males by a given dam.

In that 7 of 409 litters sampled had only a single juvenile at weaning, and fewer juveniles were weaned than there were placental scars among the small sample of dams euthanized after juvenile emergence, it appears that adaptive sex allocation occurs via litter size reduction. There is no compelling evidence, however, that such manipulation occurs early in gestation, as placental scars were not detected alongside viable embryos in the uterine horns of a larger sample of females euthanized in early gestation. This suggests either that litter size reduction occurs during the earliest stages of gestation, prior to the formation of a placental scar (Conway 1955), or that it occurs from late gestation through the end of lactation for Richardson’s ground squirrels. The correlation between elevated circulating cortisol in mothers during early gestation and increasingly male-biased litters (Ryan et al. 2012, 2014) is consistent with early litter size reduction (Ryan et al. 2012, 2014), while selective neglect or killing of nursing young seems unlikely in that infanticide is uncommon in Richardson’s ground squirrels (Michener 1973, Michener 1982). These findings, of course, do not rule out the potential involvement of pre-gestational mechanisms (Cameron 2004) affecting offspring sex.

A maternal bias toward the production of male offspring would prove advantageous where dams are capable of producing high-quality sons, who enjoy enhanced success as polygynists (Austad and Sunquist 1986; Bateman 1948; Clutton-Brock et al. 1984). The production of sons may also benefit females, however, where selection favors reduced competition with philopatric female offspring (Caley and Nudds 1987; Clark 1978; Cockburn et al. 1985; Johnson 1986), in that natal dispersal is male-biased among Richardson’s ground squirrels (Michener 1979; Michener and Michener 1977; Yeaton 1972). While the trend toward a reduced proportionate representation of males with increasing litter size persists among dams producing their second litter, that trend reversed completely with the proportionate representation of males within litters increasing with litter size among third litters produced by females. On the surface, this does not appear to be accounted for by dam age, and the concomitant effects of age on the body condition, energetic demand or competition experienced by breeding females, in that there was no apparent change in the relationship between litter size and sex ratio in contrasting 1-, 2- and 3-year-old dams. Sample sizes of known-age females are small, however, relative to the large samples of females known to be having their first, second, or third recorded litters on site, and as such, the parity effect on the trade-off between litter size and sex ratio could readily reflect energetic or physiological changes associated with aging. The statistical significance of the relationship between increasing litter size and the decreasing proportionate representation of males within litters may reflect the fact that first-time breeding yearling females that are developing to adulthood during gestation and lactation (Michener 1989a; Dobson and Michener 1995) are more energetically stressed than older females, promoting variation in female body condition that renders adaptive sex allocation commonplace in that cohort. The fact that any trend in the relationship between litter size and sex ratio reverses, with male offspring becoming more prevalent with increases in the size of third litters, suggests, however, that other factors, such as resource competition, come into play in influencing the value of producing male versus female offspring. In that the burrow systems that constitute critical resources for hibernation and female reproduction are controlled by matrilines (Michener 1979, 1981), selection may favor a shift to the production of dispersing males as dams age, and the prospect of controlling a resource-based territory ensuring one’s daughter’s reproductive success diminishes. Despite the increasing prevalence of male offspring with increasing litter size, litter sex ratio shifted to become female-biased in latter litters, with first- and second-time mothers manifesting sex ratios averaging around parity while litters of third-time mothers averaged 36.8% male. This change as dams age is consistent with observations that in philopatric iteroparous species, there is a benefit to having female-biased litters later in life, as daughters compete with their mother for resources for a shorter period of time, and do not require as pronounced an investment of resources as male offspring prior to weaning (Clark 1978; Clutton-Brock et al. 1982, 1986; Cockburn et al. 1985).

Our findings contrast with those of Gedir and Michener (2014), who found no evidence of trade-offs between litter size and sex ratio within litters of Richardson’s ground squirrels. Gedir and Michener’s results may differ from ours in that their analyses incorporated female age class as yearling versus adult, but did not explicitly account for parity. Further, differences in methods of analysis including their criterion of considering litters to have an unbiased sex ratio if they ranged anywhere between 40 and 60% male, their use of deviance information criteria (DIC), which favours models that make the least assumptions (Spiegelhalter et al. 2002), biasing model selection toward random allocation given that Mendelian segregation occurs in the background of all tested models, and overlapping predictions of the hypotheses tested, likely contributed to our disparate findings. Further, while Gedir and Michener (2014) explore litter sex ratios at the level of individual mothers, their model comparison treats the tested variants of the Trivers–Willard and the local resource competition hypotheses as mutually independent, ignoring the fact that they overlap and interact within a population (Isaac et al. 2005; Wild and West 2007).

While untestable without experimental manipulation, the difference in the findings of our studies may also ultimately be attributable to differences between populations, resulting in the conditional expression of adaptive sex allocation in our eastern population but not in their western population. Litter data used in Gedir and Michener’s (2014) study were derived from Michener’s long-term study site, situated in relatively uniform grazing land surrounded by planted crop fields (see Michener 1989b and 1996 for a more complete description of site characteristics). This implies that access to food and predation pressure were more or less uniform among squirrels on the Michener study site. Such uniformity may not promote sufficient differences in body condition and stress among individuals to elicit noticeable variation in litter size and offspring sex ratio (Boonstra et al. 1998; Hik 1995; Sheriff et al. 2009). By comparison, the Richardson’s ground squirrels providing data for our study and those of Ryan et al. (2012, 2014) occupied mowed fields and berms within the grounds of the Assiniboine Park Zoo (APZ) in Winnipeg, MB, where they were exposed to extensive variation in food availability, owing to access to zoo animal feed in some areas, “hand-outs” of food from zoo visitors in others, or no supplementary food whatsoever. The APZ squirrels also likely experience considerable variation in predation pressure owing to variation in their proximity to frequent human presence, and extensive variation in habitat features characterizing the zoo setting (Wood 2018). In a general sense, the difference between our findings and those of Gedir and Michener (2014) illustrates that the expression of adaptive sex allocation is contingent upon the local environmental conditions experienced by individuals within a species.

The existence of litter size–sex ratio trade-offs in our long-term data suggests that the trends Ryan et al. (2012, 2014) reported are robust rather than a product of sampling bias (Gedir and Michener 2014). They do not, however, apply uniformly across populations or even among years within a given population. Adaptive sex allocation is a product of multiple interacting mechanisms (Dušek et al. 2011; Fisher 1999; Isaac et al. 2005; Wild and West 2007), and is expressed conditionally in response to varying intrinsic and extrinsic factors. Extrinsic factors including temperature and rainfall (Myers et al. 1985), access to supplemental food (Austad and Sunquist 1986; Gray et al. 2013, Schmidt and Hood 2012), and in particular, diets high in saturated fats (Rosenfeld et al. 2003) vary unpredictably in space and time. Intrinsic factors such as genetic constitution (Dolf et al. 2008; Gubbels et al. 2009), body condition (Austad and Sunquist 1986), circulating testosterone (Bánszegi et al. 2012; Grant 2008) or glucocorticoid levels (Cameron 2004; Ryan et al. 2012, 2014) also vary, and interact with extrinsic factors influencing optimal reproductive strategies, as revealed here by the association between biased sex ratios among primiparous females and construction events. This association also supports the assertion of Ryan et al. (2014) that variation in maternal stress modulates trade-offs between litter size and sex ratio.

Construction of the Australian Walkabout in 2013 with the provision of carrots, beets and various fresh green vegetables for red kangaroos and emus also provided squirrels with access to supplementary food from 2013 through 2016. While access to supplemental food can reduce local resource competition, differential access to food also enhances variation in body condition among dams, mirroring food supplementation studies exploring adaptive sex allocation in other species (Austad and Sunquist 1986; Koskela et al. 2004). While no significant difference in average litter sex ratios was detected before versus after Walkabout construction, post-construction litters did become increasingly female-biased as litter size increased, while no such trend was evident prior to walkabout construction and food supplementation. This difference again emphasizes that the expression of Adaptive Sex Allocation is contingent upon environmental influences, and involves interactions between benefits derived via sex-differential variance in reproductive success and local resource competition.

Differential stresses imposed by human activity including construction at our APZ site, along with extensive variation in the body condition of dams owing to variation in the availability of supplementary food, appear to modulate the conditional expression of adaptive sex allocation. Our examination of placental scars in mothers of known litters suggests that adjustment of offspring sex ratio through litter size reduction occurs during the earliest stages of gestation, leaving no evidence of placental scarring, or during the post-partum period. Further studies are necessary to document the factors contributing to the conditional expression of adaptive sex allocation in Richardson’s ground squirrels and other species.