Progeny sex ratio variation in ungulates: maternal age meets environmental perturbation of demograpgy

FORUM FORUM FORUM

FORUM is intended for new ideas or new ways of interpreting existing information. It provides a chance for suggesting hypotheses and for challenging current thinking on ecological issues. A lighter prose, designed to attract readers, will be permitted. Formal research reports, albeit short, will not be accepted, and all contributions should be concise with a relatively short list of references. A summary is not required.

Progeny sex ratio 7ariation in ungulates: maternal age meets en7ironmental perturbation of demograpgy Da7id Saltz, Mitrani Dept of Desert Ecology, Jacob Blaustein Inst. for Desert Research, Ben Gurion Uni7. of the Nege7, Sede Boqer Campus, 84990 Israel ([email protected]).

Many adaptive theories have been proposed for explaining variation in progeny sex ratio (PSR) of vertebrates (Hardy 1997). Of these, two opposing hypotheses are commonly used for ungulates: the Trivers and Willard hypothesis (TWH, Trivers and Willard 1973) and the local resource competition hypothesis (LRC, Silk 1983, Caley and Nudds 1987). Trivers and Willard (1973) proposed that when adaptive variance in reproductive success differs between the sexes, mothers should invest more in the sex with the higher variance and, therefore, opt to produce this sex when conditions enable the extra investment. Because most ungulates are polygynous with reproductive success varying considerably between males, this hypothesis is invoked if a male-biased PSR is associated with favorable conditions (e.g. Clutton-Brock et al. 1988 and others mentioned in Table 1). LRC predicts that the gender less apt to disperse imposes overall greater competition on the mother. To minimize this competition, mothers should produce the non-dispersing gender when conditions are favorable. Because in ungulates sons tend to disperse earlier than daughters, the LRC hypothesis is invoked if favorable conditions are associated with a female biased PSR (e.g. Hewison and Gaillard 1996 and others mentioned in Table 1). A separate adaptive theory for each of two opposing situations makes any results explainable on an ad hoc basis. This has caused several critics to question the findings (Williams 1979, Festa-Bianchet 1996, and others) and attribute them to random occurrences. However, as long as the pattern remains consistent within a species, the findings cannot be challenged and could be explained by differences in the biology of the various species (Kojola 1998). Nonetheless, I documented nine cases where the pattern was inconsistent within a species (Table 1). OIKOS 94:2 (2001)

As pointed out by Hewison and Gaillard (1999), the abundance of intraspecific inconsistencies requires some explanation. While Type-I-error (Festa-Bianchet 1996) or heterogeneity within species (Hewison et al. 1999) are possible explanations for these inconsistencies, they appear improbable to me. It is my opinion that in many cases PSR is determined by the age of the mother and if the age of the mother is unaccounted for spurious correlations may be produced between PSR and other factors that have an age-dependent impact on the population. I demonstrate this effect using a Leslie matrix model (Appendix I) based on the dynamics of a reintroduced population of Asiatic wild ass exhibiting PSR that is dependent on maternal age (Saltz and Rubenstein 1995).

PSR in reintroduced Asiatic wild ass The Asiatic wild ass is an endangered equid that was reintroduced in Israel in 1983. I studied this reintroduced population from 1988 to 1996 (Saltz and Rubenstein 1995, Saltz et al. 2000). Until 1992, PSR was significantly male-biased (p= 0.012, N= 52, x2 = 6.23, df = 1) and was attributed, in accordance with the TWH, to the low population density. In 1993, PSR became female skewed (Fig. 1), and since the number of adult females in the population was consistently increasing the negative trend fit well the TWH. However, as data accumulated we found that PSR was in fact significantly dependent on maternal age (Saltz and Rubenstein 1995), with prime-aged females producing more males and young primiparous and old females producing mostly females. A significant pattern was evident in the world studbook as well and was still evident in the field in later years. Furthermore, data 377

from the studbook suggested that females that were primiparous at older ages tended to produce males (Saltz and Rubenstein 1995) suggesting that young age rather than primiparity is the cause of a female skewed PSR. Consequently, Saltz and Rubenstein (1995) concluded that male skewed PSR in the early years was due to the abnormal age structure of the population caused by the reintroduction process (most reintroduced female were between three and five years old when released). The aging of the reintroduced females caused the negative trend in PSR. From 1993 to 1996, although the number of adult females in the population continued to increase PSR began shifting back towards being male-skewed (Fig.

1). This is in contrast to the trend from 1988 to 1992. Had maternal age been unknown and the study been restricted to just these years (1993–1996), LRC would be a reasonable explanation. However, the pattern was in fact caused by a combination of two processes: (1) females born in the wild reaching the male producing age groups, and (2) delayed primiparity (probably associated with the increase in population density), expressed as a decline in the annual natality rate of three-year-old females from 1.00 (n= 6) in the years 1988–1991 to 0.3 (n= 10) in the years 1992– 1995 (P = 0.006, Fisher’s exact). Annual natality rate of females four years old and older remained unchanged during this period (0.65, n= 40 and 0.69, n =67, respectively,

Table 1. Studies on ungulates documenting deviations of offspring sex ratio from the expected 1:1 classified according to the theory they support. Species

Deviations of offspring sex ratio from unity in agreement with Trivers and Willard’s hypothesis

Deviations of offspring sex ratio in agreement with the local resource competition hypothesis

Correlate

Reference

Correlate

Reference Watson 1982

Nutrition

Gazella gazella Gazella cu6ieri

Prime age Prime age

Yang et al. 1989 Shy et al. 1998

O6is canadensis

Past reproduction

Bison bison

Past reproduction Prime age

Capreolus capreolus Odocoileus 6iginianus

Prime age Reproductive success

Kent 1995 Kent 1992 Meilke et al. 1993 Baharav 1989 Alados and Escos 1994 Berube et al. 1996 Rutberg 1986 Green and Rothstein 1991 Wauters et al. 1995 Woolf and Harder 1979 Mansell 1974

Habitat condition

Sus scrofa

Prime age Litter size Dominance

Odocoleus hemionus

Litter size and body

Kucera 1991,

mass

Burke and Birch 1995

O6is aries

Dama mesopotamica Cer6us elaphus

Rangifer tarandus

Prime age and condition Prime age Dominance Condition Density and weather Prime age Body mass

Equus hemionus

Prime age

Equus zebra

Prime age (non-significant trend)

Equus gre6yii Equus caballus

Prime age Condition

Ceratotherium simum

Prime age and past reproduction

378

Saltz 1996, this paper Clutton-Brock et al. 1988 Kohlmann 1999 Kruuk et al. 1999 Kojola and Eloranta 1989 Kojola 1997, Thomas et al. 1989 Saltz and Rubenstein 1995 Novellie et al. 1996 Hayward 1987 Cameron et al. 1999 Owen-Smith 1988

Population density Dominance

Festa-Bianchet 1991

Body mass

Reproductive success

Hewison and Gaillard 1996 Degayner and Jordan 1987, Verme 1983 Verme 1985 Robinette et al. 1977, Pederson and Harper 1984 Verme 1983

Winters severity

Lowe 1969

Prime age and reproductive success Nutrition Habitat

Post et al. 1999 Habitat

Skogland 1986

Dominance

Lloyd and Rasa 1989

OIKOS 94:2 (2001)

Fig. 1. Proportion of males born by year in a reintroduced Asiatic wild ass population (solid line) and the number of adult females in the population (broken line). Numbers below solid line are the number of adult females giving birth (sample size). Most females were three- or four-years-old at the time of reintroduction resulting in an abnormal age distribution.

P= 0.696, Fisher’s exact). As mentioned, females that were primiparous at older ages did not exhibit a tendency towards female-biased PSR when primiparous. To summarize: prime-aged female Asiatic wild assess tend to produce male offspring in agreement with TWH. As the population reintroduced in Israel increased, PSR in the population shifted from male skewed to female skewed and back again. This was caused mainly by the abnormal age class structure dictated by the reintroduction protocol and a density dependent delay in primiparity. However, if the age of individuals was unknown, the changes over time in PSR would support either TWH or LRC, depending on the time-frame of the study.

PSR, maternal investment, and age in ungulates Maternal investment in mammals extends over a period of time and may vary considerably from conception to the end of lactation. In species where maternal investment should vary according to offspring gender this would have consequences for the choice of cue on which to base gender determination. In particular, an effective cue must correlate with the mother’s ability to invest not only at time of conception, but also later on during gestation and especially during lactation (the most costly part of rearing an offspring – Oftedal 1985). Thus, the duration of maternal investment, its distribution over time, and the predictability of the environment will affect which cues are better determinants of offspring gender. Ungulates typically have long gestation and lactation terms that span several seasons. These seasons are often poorly correlated and vary greatly in food availability. Usually, conception occurs before the winter and parturition after the winOIKOS 94:2 (2001)

ter. The winter is probably the most important determinant of the mother’s condition and habitat quality at parturition (in both arid and temperate zones). Because mothers can not foresee winter severity and because the mothers’ condition and habitat quality at the time of conception may have little or no correlation to these factors at the time of parturition and lactation, environmental conditions at conception are not likely to be a reliable cue upon which females should base their ‘decision’ concerning the gender of the offspring. By contrast, maternal age is only weakly correlated with the mother’s present condition, but is a better predictor of her future status. Thus, in many ungulates species, age should be a more reliable cue upon which to base gender determination than environmental or body conditions at time of conception. Interestingly, age is, in fact, the more common correlate of PSR in ungulates (Table 1). Specifically, young primiparous females that are inexperienced and have not reached full adult size and old females that are physically weaker should produce the gender requiring less investment and vice versa.

Consequences If age is a major determinant of PSR in ungulates and if it is unaccounted for in any analysis, PSR may appear to be affected by other factors that are correlated to age but not necessarily in the way PSR is, thus generating spurious correlations. I presented two specific examples from the Asiatic wild ass. These two examples can be generalized as: (1) deviation from a normal age structure and (2) density dependent responses varying with age. Are these two phenomena typical of ungulate populations? If so how would they impact PSR at the population level in species where maternal age determines PSR?

Deviations in age structure Because regulation of ungulate populations occurs mostly through reproductive success and juvenile mortality (Gaillard et al. 1998), environmental fluctuations affect the age structure of ungulate populations (Sæther 1997). Periodically, annual natality and recruitment in ungulates, following harsh winters or droughts, can approach or even equal zero (Mason 1990, Owen-Smith 1990, Bartmann et al. 1992). This will cause missing cohorts. As the missing cohort advances through the ages, PSR for the population will change accordingly. When the missing cohort passes through male producing ages, PSR will be female skewed and vice versa. The effect of a single missing cohort on PSR, when PSR is determined by maternal age, can be demonstrated with a simple deterministic Leslie matrix model 379

(Appendix I) based on the Asiatic wild ass life table presented by Saltz and Rubenstein (1995), assuming a logistic growth curve, normal age distribution, and keeping carrying capacity (K) constant. The missing cohort was created simply by setting annual natality= 0 in one pulse (year) of the simulation. Following this infraction, a declining oscillation occurs in PSR with waves of 2 –4 years (Fig. 2). When K was allowed to fluctuate, the frequency at which population size exceeded K was dependent on the inflection point of the growth curve. The inflection point was predetermined by the exponent in the generalized logistic equation (Gilpin et al. 1976). When the inflection point was set at 0.5K or 0.6K, there were no significant deviations of PSR from 1:1 (P\ 0.05 in a random sample of 100 offspring/pulse) in any of the 50 pulses analyzed (Appendix I) (Fig. 3a, b). However, the data for consecutive years were autocorrelated, i.e. PSRs were consistently skewed in one direction over a period of several pulses (years). For an inflection point occurring at 0.7K, 14 of 50 pulses had a PSR significantly different from unity for N= 100 (PB 0.05), ten of them male-skewed (Fig. 3c). For an inflection point occurring at 0.8K, 20 of 50 pulses had a PSR significantly different from unity ranging from 0.4 to 0.6, 14 of them male-skewed (Fig. 3d). When population size exceeded K there was no recruitment (as defined by the model). Note, in Fig. 3d, that the proportion of males born always declined from the year before the population exceeded K to the year after the population exceeded K. Since density is expected to decline following years with no recruitment, the apparently associated decline in PSR could be explained within the framework of LRC. This, in spite of the underlying assumption of TWH dictated in the model, i.e. that prime age females tend to produce sons.

Fig. 3. Proportion of males born in a hypothetical stochastic population growth model based on the life table of Asiatic wild ass and using the generalized logistic growth curve with inflection points at 0.5K (a), 0.6K (b), 0.7K (c), and 0.8K (d). Stochasticity is induced by allowing K to fluctuate 9 20% in a random uniform distribution around the mean. Thick solid line with circles is for an un-harvested population near carrying capacity. Thin broken line is for a harvested population kept at approximately 0.5K. Asterisks indicate a significant deviation in the un-harvested population from the expected 0.5 for a random sample of 100 offspring. Gaps in the line indicate reproductive pauses caused by the population exceeding carrying capacity.

Density dependent responses varying with age

Fig. 2. The impact of a single missing cohort over time on the proportion of males born in a hypothetical deterministic population growth model based on the life table of Asiatic wild ass. The missing cohort is caused by a reproductive pause in a single pulse (marked by arrow).

380

In many species of ungulates the reduction in population reproductive success due to increased population densities is often manifested in the form of delayed primiparity (Owen-Smith 1990, Jorgenson et al. 1993). However, females that are primiparous at older ages may produce a PSR that is no different from that of non-primiparous females of the same age (Saltz and Rubenstein 1995, Kojola 1997) suggesting that age, in OIKOS 94:2 (2001)

the form of achieving adult size, is important in determining PSR in primiparous mothers. Thus, if younger females tend to produce a specific gender, delayed primiparity will cause a shift in PSR at the population level. I found such a pattern also in Persian fallow deer (Dama dama mesopotamica) that, based on studbook data, exhibit an age-dependent sex ratio (Saltz 1996). Twenty-five percent of females primiparous at age 2 gave birth to males (n=16) while 69% of females primiparous at age 3 or older gave birth to males (n= 13) (P=0.027, Fisher’s exact). In terms of the Leslie matrix model described above, the decline in annual natality rate in response to increased population densities was stronger in three-yearold female wild asses (the female-producing age group) than older females. I included this in the model by setting the inflection point for three-year-old females to 0.5K and for older females to 0.8K. I then tested the impact of density by simulating two harvest regimes: one reducing the population to approximately 0.7K (low harvest) and the other reducing the population to 0.4K (high harvest). In both cases PSR was male skewed, but under the low harvest, PSR was more male-skewed and more variable than the high-harvest PSR (Fig. 4). Thus, at low harvest, when resources are presumed to be limited, more males are produced, again, falsely supporting the LRC hypothesis.

Discussion In ungulate species where PSR is determined by maternal age, environmental conditions may appear to impact PSR if age is unaccounted for. The shifts in PSR may result from missing cohorts caused by periodic slumps in reproduction following bad winters or a delay in primiparity caused by increased population

Fig. 4. Proportion of males born in a population where density dependence of female-producing three-year-old mothers is stronger (inflection point at 0.5K) than in older mothers (inflection point at 0.8K) for two populations kept at approximately 0.7K (low harvest) and 0.4K (high harvest), respectively. OIKOS 94:2 (2001)

density. The latter requires that annual natality rate and PSR be independent of each other. The claim of independence between annual natality rate and PSR is supported by the fact that age of primiparity is usually affected by population density but it is the age, not primiparity, which determines PSR. Although when measured at the population level PSR may significantly differ from par due to changes in age structure or delayed primiparity, these differences are expected to be minor because the true factor affecting PSR is not accounted for. Thus, random noise would have a dampening effect. For the Asiatic wild ass the model predicts PSRs ranging from 0.4 to 0.6. These results are in agreement with studies where PSR was linked to population, rather than individual animal, characteristics (i.e. habitat, population density, weather, etc., Pederson and Harper 1984, Clutton-Brock and Iason 1986, Skogland 1986, Hewison and Gaillard 1996, Shy et al. 1998, Post et al. 1999) Given the long gestation period and the cost of lactation, environmental conditions, that determine maternal nutritional status at the time of conception, are not expected to be a good cue for deciding the sex of progeny in many ungulate species. The dangers of correlating PSR at the population level to environmental conditions have been previously pointed out (Burke and Birch 1995) and such correlations must be treated with caution. Interestingly, of 15 studies that I documented supporting LRC (Table 1) PSR was associated with environmental conditions (nutrition, habitat, winter severity, etc.) in seven cases. By contrast, in none of 25 papers supporting the Trivers and Willard hypothesis were environmental conditions associated with PSR (P= 0.002, Fisher’s exact), with one possible exception (Kruuk et al. 1999). Kruuk et al. (1999) found that in red deer fetal sex ratio was determined at conception by maternal dominance, but harsh winters and population density caused an increased loss of male fetuses. The cases in which PSR was correlated with natality rate could be a direct outcome of young females delaying primiparity rather than an actual change in the tendency to produce a certain gender. Furthermore, although not addressed in the model presented in this paper, poor environmental conditions may also have a stronger effect on the natality rate of older age groups that tend to produce females too. In seven of the nine species where data supported both TWH and LRC, age was one of the correlates but was never found to support both hypotheses in a single species. Finally, of 11 species where PSR was correlated with maternal age, only one (white-tail deer) supported LRC. Because the relationship between PSR and age may be bell-shaped, rather than linear, field data may support either the TWH or LRC, depending on the age of the individual animals being studied. If the natality rate of young females is low due to high densities, the relationship between PSR and age will appear linearly 381

correlated in one direction; if older age groups are not adequately sampled (due to their general scarceness in the population), PSR will appear linearly correlated with age in the opposite direction. The possible existence of a correlation between natality, age, dominance, and weight (Carranza 1988, Ellard and Crowell-Davis 1989, Kojola 1989, Hirotani 1990, Hass 1991, Locati and Lovari 1991, Zhakikh 1997, Rutberg and Greenberg 1998) further complicates the matter. For example, if subordinate females exhibit delayed primiparity when environmental conditions are poor, an age-dependent PSR will cause a male skewed PSR following bad years and dominant females will produce higher ratio of females then subordinates. Thus, unless maternal age and other individual animal characteristics are known and considered, presented results concerning PSR in ungulates should be treated with caution.

Conclusions Although age appears to be a predominant determinant of PSR, it may not be the only determinant. The main purpose of this paper is to provide a possible explanation for several intra-specific anomalies (Table 1) and demonstrate that simple, and seemingly interpretable, correlations may lead to erroneous conclusions. The latter is especially true in view of the fact that more than one factor may influence PSR and that PSR is the outcome of both gender-determination at conception and differential fetal morality (Kruuk et al. 1999). To avoid such errors, in-depth knowledge of individual animal characteristics, population status, and environmental conditions is necessary to decipher observed changes in PSR. The work of Kruuk et al. (1999) is an excellent example of such an in-depth study. In the search for a generalization of the principles governing PSR it is necessary to separate between differential fetal mortality and sex determination at time of conception. While the former is a direct response to present nutritional restrictions, the latter is a form of bet hedging and must be evaluated within the framework of the species’ ontogeny and the predictability of the environment it occupies. If the correlation between environmental conditions at conception and environmental conditions at parturition is poor (e.g. temporally variable environments) then mothers must rely on rules of thumb, such as age or dominance, for determining the sex of the offspring at conception. If the impact of a dominance hierarchy on maternal condition is low then age will influence PSR, otherwise dominance is expected to influence PSR. One possible way of isolating the various factors is comparing field data with zoo data (Owen-Smith 1988, Saltz and Rubenstein 1995). 382

Acknowledgements – I thank B. Kotler, M. Rowen, D. Rubenstein, D. Ward and U. Motro who assisted in various stages of preparing the manuscript. This is publication number 317 of the Mitrani Dept of Desert Ecology.

References Alados, C. L. and Escos, J. M. 1994. Variation in the sex ratio of a low dimorphic polygynous species with high levels of maternal reproductive effort: Cuvier’s gazelle. – Ethol. Ecol. Evol. 6: 301– 311. Baharav, D. 1989. The dynamics of mountain gazelle in Israel. – Final research report submitted to the Nature Reserves Authority, Israel. Bartmann, R. M., White, G. C. and Carpenter, L. H. 1992. Compensatory mortality in a Colorado mule deer population. – Wildl. Monogr. 121: 5– 39. Berube, C. H., Festa-Bianchet, M. and Jorgenson, J. T. 1996. Reproductive costs of sons and daughters in Rocky Mountain bighorn sheep. – Behav. Ecol. 7: 60– 68. Burke, R. L. and Birch, J. M. 1995. White-tailed deer vary offspring sex-ratio according to maternal condition and age. – Ecol. Res. 10: 351– 357. Caley, M. J. and Nudds, T. D. 1987. Sex-ratio adjustment in Odocoileus: does local resource competition play a role? – Am. Nat. 129: 452– 457. Cameron, E. Z., Linklater, W. L., Stafford, K. J. and Veltman, C. J. 1999. Birth sex ration relate to mare condition at conception in Kaimanawa horses. – Behav. Ecol. 10: 472 – 475. Carranza, J. 1988. Dominance relationships in female groups of red deer: seasonal changes. – Acta Theriol. 33: 435– 442. Clutton-Brock, T. H. and Iason, G. R. 1986. Sex ratio variation in mammals. – Q. Rev. Biol. 61: 339– 374. Clutton-Brock, T. H., Albon, S. D. and Guinness, F. H. 1988. Reproductive success in male and female red deer. – In: Clutton-Brock, T. H. (ed.), Reproductive success. Univ. of Chicago Press, pp. 325– 343. Degayner, E. J. and Jordan, P. A. 1987. Skewed fetal sex ratios in white-tailed deer: evidence and evolutionary speculations. – In: Wemmer, C. M. (ed.), Biology and management of Cervidae. Smithsonian Inst. Press, pp. 178– 188. Ellard, M. E. and Crowell-Davis, S. L. 1989. Evaluation equine dominance in draft mares. – Appl. Anim. Behav. Sci. 24: 55 – 76. Festa-Bianchet, M. 1991. The social system of bighorn sheep: grouping patterns, kinship and female dominance rank. – Anim. Behav. 42: 71– 82. Festa-Bianchet, M. 1996. Offspring sex ratio studies of mammals: does publication depend upon the quality of the ´ coscience 3: research or the direction of the results? – E 42 – 44. Fowler, C. W. 1988. Population dynamics as related to rate of increase per generation. – Evol. Ecol. 2: 197– 204. Gaillard, J., Festa-Bianchet, M. and Yoccoz, N. G. 1998. Population dynamics of large herbivores: variable recruitment with constant adult survival. – Trends Ecol. Evol. 13: 58 – 63. Gilpin, M. E., Case, T. J. and Ayala, F. J. 1976. U-selection. – Math. Biosci. 32: 131– 139. Green, W. C. H. and Rothstein, A. 1991. Sex bias or equal opportunity? Patterns of maternal investment in bison. – Behav. Ecol. Sociobiol. 29: 373– 384. Hardy, C. W. 1997. Possible factors influencing vertebrate sex ratios: an introductory overview. – Appl. Anim. Behav. Sci. 51: 217 – 241. Hass, C. C. 1991. Social status in female bighorn sheep (O6is canadensis): expression, development and reproductive correlates. – J. Zool. 225: 509– 524. Hayward, L. 1987. World studbook of Grevy’s zebra (Equus gre6yi ). – Marwell Zoological Park, Hampshire, UK. OIKOS 94:2 (2001)

Hewison, A. J. M. and Gaillard, J. M. 1996. Birth sex ratios and local resource competition in roe deer Capreolus capreolus. – Behav. Ecol. 4: 461–464. Hewison, A. J. M. and Gaillard, J. M. 1999. Successful sons or advantaged daughters? The Trivers-Willard model and sexbiased maternal investment in ungulates. – Trends Ecol. Evol. 14: 229–234. Hewison et al. 1999, p. 3 in MS, ??? Hirotani, A. 1990. Social organization of reindeer (Rangifer tarandus) with special reference to relationship among females. – Can. J. Zool. 68: 743–749. Jorgenson, J. T., Festa-Bianchet, M., Lucherinin, N. and Wishart, W. D. 1993. Effects of body size, population density and maternal characteristics on age of first reproduction in bighorn ewes. – Can. J. Zool. 71: 2509– 2517. Kent, J. P. 1992. Birth sex ratios in sheep over six lambing seasons. – Behav. Ecol. Sociobiol. 30: 151–155. Kent, J. P. 1995. Birth sex ratios in sheep over nine lambing seasons: years 7–9 and the effects of aging. – Behav. Ecol. Sociobiol. 36: 101–104. Kohlmann, S. G. 1999. Adaptive fetal sex allocation in elk: evidence and implications. – J. Wildl. Manage. 63: 1109– 1117. Kojola, I. 1989. Mother’s dominance status and differential investment in reindeer calves. – Anim. Behav. 38: 177– 185. Kojola, I. 1997. Social status and physical condition of mother and sex ratio of offspring in cervids. – Appl. Anim. Behav. Sci. 51: 267 –274. Kojola, I. 1998. Sex ratio and maternal investment in ungulates. – Oikos 83: 567–573. Kojola, I. and Eloranta, E. 1989. Influences of maternal body weight, age, and parity on sex ratio in semidomesticated reindeer (Rangifer t. tarandus). – Evolution 43: 1331– 1336. Kruuk, L. E. B., Clutton-Brock, T. H., Albon, S. D. et al. 1999. Population density affects sex ratio in red deer. – Nature 399: 459–461. Kucera, T. E. 1991. Adaptive variation in sex ratios of offspring in nutritionally stressed mule deer. – J. Mammal. 72: 745 – 749. Lloyd, P. H. and Rasa, O. A. E. 1989. Status, reproductive success and fitness in Cape mountain zebra (Equus zebra zebra). – Behav. Ecol. Sociobiol. 25: 411–420. Locati, M. and Lovari, S. 1991. Clues for dominance in female chamois, age weight or horn size. – Aggressive Behav. 17: 11 – 16. Lowe, V. P. W. 1969. Population dynamics of the red deer (Cer6us elaphus L.) on Rhum. – J. Anim. Ecol. 38: 425– 457. Mansell, W. D. 1974. Productivity of white-tailed deer in Bruce Peninsula, Ontario. – J. Wildl. Manage. 38: 808– 814. Mason, D. R. 1990. Juvenile survival and population structure of blue wildebeest and warthogs in the Central Region of the Kruger National Park. – Koedoe 33: 29 –46. Meilke, D. B., Drickamer, L. C., Vessey, S. H. et al. 1993. Maternal dominance rank and secondary sex ratio in domestic swine. – Anim. Behav. 46: 79–85. Novellie, P. A., Millar, P. S. and Lloyd, P. H. 1996. The use of VORTEX simulation models in a long term program of re-introduction of an endangered large mammal, the Cape mountain zebra (Equus zebra zebra). – Acta Oecol. 17: 657 – 671. Oftedal, O. T. 1985. Pregnancy and lactation. – In: Hudson, R. J. and White, R. G. (eds), Bioenergetics of wild herbivores. CRC Press, pp. 215–238. Owen-Smith, N. R. 1988. Megaherbivores. – Cambridge Univ. Press. Owen-Smith, N. R. 1990. Demography of a large herbivore, the greater kudu Tragelaphus strepsiceros, in relation to rainfall. – J. Anim. Ecol. 59: 893–914. Pederson, J. C. and Harper, K. T. 1984. Does summer range quality influence sex ratios among mule deer fawns in Utah. – J. Range Manage. 37: 64– 66. OIKOS 94:2 (2001)

Post, E., Forchhammer, M. C., Steneth, N. C. and Langvatn, R. 1999. Extrinsic modification of vertebrate sex ratios by climatic variation. – Am. Nat. 154: 194– 204. Robinette, W. L., Hancock, N. V. and Jones, D. A. S. 1977. The Oak Creek mule deer herd in Utah. – Utah State Div. Wildl. Res. Publ. 77-15. Rutberg, A. T. 1986. Lactation and fetal sex ratios in American bison. – Am. Nat. 127: 89– 94. Rutberg, G. T. and Greenberg, S. A. 1998. Dominance, aggression frequencies and modes of aggression competition in feral pony mares. – Anim. Behav. 40: 322– 331. Sæther, B.-E. 1997. Environmental stochasticity and population dynamics of large herbivores: a search for mechanisms. – Trends Ecol. Evol. 12: 143– 149. Saltz, D. 1996. Minimizing extinction probability due to demographic stochasticity in a reintroduced herd of Persian fallow deer (Dama dama mesopotamica). – Biol. Conserv. 75: 27 – 33. Saltz, D. and Rubenstein, D. I. 1995. Population dynamics of a reintroduced Asiatic wild ass (Equus hemionus) herd. – Ecol. Appl. 5: 327– 335. Saltz, D., Rowen, M. and Rubenstein, D. I. 2000. The impact of space use patterns of reintroduced Asiatic wild ass on effective population size. – Conserv. Biol. 14: 1852– 1862. Shy, E., Frankenberg, E., Kaplan, D. et al. 1998. The effect of management on mountain gazelle populations in Israel. – Gidier Faune Sauvage 15: 617– 634. Silk, J. B. 1983. Local resource competition and facultative adjustment of sex ratios in relation to competitive abilities. – Am. Nat. 121: 56– 66. Skogland, T. 1986. Sex ratio variation in relation to maternal condition and parental investment in wild reindeer (Rangifer t. tarandus). – Oikos 46: 417– 419. Spinage, L. A. 1972. African ungulate life tables. – Ecology 53: 645 – 652. Thomas, D. C., Barry, S. J. and Kilian, H. P. 1989. Fetal sex ratios in caribou: maternal age and condition effects. – J. Wildl. Manage. 53: 885– 890. Trivers, R. L. and Willard, D. E. 1973. Natural selection of parental ability to vary sex ratio of offspring. – Science 179: 90– 92. Verme, L. J. 1983. Sex ratio variation in Odocoileus: a critical review. – J. Wildl. Manage. 47: 573– 582. Verme, L. J. 1985. Progeny sex ratio relationships in deer: theoretical vs. observed. – J. Wildl. Manage. 49: 134– 136. Watson, J. S. 1982. Parental manipulation of the sex ratio of offspring: a historical note. – Am. Nat. 119: 283– 284. Wauters, L. A., de Crombrugghe, S. A., Nour, N. and Matthysen, E. 1995. Do female roe deer in good condition produce more sons than daughters. – Behav. Ecol. Sociobiol. 37: 189 – 193. Williams, G. C. 1979. The question of adaptive sex ratio in outcrossed vertebrates. – Proc. R. Soc. Lond. B 205: 567 – 580. Woolf, A. and Harder, J. D. 1979. Population dynamics of a captive white-tailed deer herd with emphasis on reproduction and mortality. – Wildl. Monogr. 67: 1– 53. Yang, H., Eastham, P. R., Phillips, P. and Whitemore, C. T. 1989. Reproductive performance, body weight, and body condition of breeding sows with differing body fatness at parturition, differing nutrition during lactation, and differing litter size. – Anim. Prod. 48: 181– 201. Zhakikh, T. L. 1997. Dominance hierarchy in harem bands of Przewalski horses Equus przewalkii at the Askania Nova Reserve. – Vestn. Zool. 31: 69– 74.

Appendix I: The computer model The Leslie matrix model was based on the data and model for Asiatic wild ass (Equus hemionus) described in Saltz and Rubenstein (1995) with density dependence 383

and a randomly varying carrying capacity (K) added arbitrarily. Pulses were one year. Density dependence in the model was derived from the generalized logistic equation (Gilpin et al. 1976) 15

  n

N0(m,f) = % mx(m,f)Nx( f) 1− x=0

Ntot KR

U

where N0(m,f) is the number of male or female progeny produced at each pulse and surviving to age 1, mx(m,f) is the proportion of male or female offspring produced by an average mother of age x at low density and surviving to age 1, Nx(f) is the number of females aged x, Ntot is total population size, KR is realized carrying capacity, and U is the asymmetry of the growth curve. If U is greater than 1, the inflection point occurs above the mid-point of the curve, and vice versa. Adult survival was based on the survival curve of plains zebra (Spinage 1972) and was independent of density. I used the generalized logistic equation for two reasons: (1) Ungulates are expected to have an inflection point greater than the commonly used 0.5K (U \1), reaching as high as 0.8K (Fowler 1988). The high inflection point causes occasional overshoots of K resulting in no reproduction in those pulses, thus inducing the necessary shifts in age structure required by the first scenario. (2) By assigning different values of U to different age groups I could vary the density dependent responses between the age groups as described in the second scenario.

384

Based on Saltz and Rubenstein (1995) expected PSR in a population of Asiatic wild ass with stable age distribution was slightly male skewed. This was probably due to sampling error. In the model I adjusted the overall net reproductive rate (R0) of male offspring so it would be exactly equal to that of females. Thus, in a population with a stable age distribution where density dependence is equal for all age groups PSR is expected to be 1.0. K was set arbitrarily to 10000 and was allowed to fluctuate 920% around the mean as a random uniform variable reflecting annual fluctuations in environmental conditions. Because animal condition and the resulting natality rate are dependent not only on present environmental conditions, but also on past conditions, the realized carrying capacity KR was calculated as (Kt + KR(t − 1))/2, where Kt is carrying capacity at the present pulse and KR(t − 1) is realized carrying capacity at the previous pulse. Thus, 50% of population performance is determined by present conditions, 25% by conditions in the previous pulse, 12.5% by conditions two pulses ago, etc. I started each run with an initial population of 3000 animals (100 males and 100 females in each age group) and ran the model for 1050 pulses. To remove any effects of the initial age structure I used only pulses 1000–1050 to test for shifts in PSR. PSR in each pulse was tested against the expected 1:1 using the normal approximation of the binomial and assuming a sample size of 100.

OIKOS 94:2 (2001)