S08.5: Sex allocation in co-operatively breeding birds

Department of Biology, Queen’s University, Kingston K7L 3N6, Ontario, Canada, fax 613-545-6617, e-mail yezerins@biology.queensu.ca

In theory, the sex allocation that maximises parental fitness is affected by any factor that predictably changes the relative cost of the sexes or their reproductive potential. In co-operative breeders these factors include the sex of helpers and their effects on reproductive success of the group, the relatedness of group members, and the dispersal of offspring. Most of these and other key factors are known to vary among species, among groups, and even among individuals within co-operative groups. Interspecific comparative tests for overall differences in sex allocation are necessary for factors that show little intraspecific variation (e.g. sexual size dimorphism), but they are problematic because of the difficulty in measuring all relevant factors and because the joint effect of all the factors has not been integrated theoretically. Intraspecific tests for facultative changes in sex allocation that examine individuals in contrasting circumstances are more promising for testing single factors. Although few empirical studies of avian sex allocation have been published, this will change with recent technical innovations for sexing offspring. I review results from empirical studies of co-operatively breeding birds. These illustrate that adaptive changes in sex allocation can occur. Nonetheless, these studies also highlight challenges for sex allocation theory.

Sex allocation is the distribution of resources by parents into the production of male versus female offspring. Parents are naturally selected to make this division so as to maximize their genetic contribution to future generations. Although the theory linking parental cost-benefit tradeoffs to sex allocation is one of the best developed in evolutionary ecology (Trivers & Willard 1973; Charnov 1982; Frank 1990), empirical evidence of adaptive sex allocation in birds is scant (Gowaty 1991). Ironically, some of the most convincing theory and evidence for adaptive sex allocation in birds comes from a relatively rare reproductive system, co-operative breeding, in which some members of a social group provide care to young that are not their offspring.

Co-operative breeding occurs across a phylogenetically diverse array of species, and among these species the social organization and forms of co-operation are varied (Brown 1987; Stacey & Koenig 1990). Moreover, many aspects of co-operative breeding seem likely to alter the relative worth of sons and daughters to parents, which should affect sex allocation. In this review, I begin with an outline of basic sex allocation theory as it applies to birds. I then discuss a variety of specific social and ecological circumstances that could create variation in adaptive sex allocation in birds with a co-operative social system. Lastly, I review empirical studies of sex allocation in co-operatively breeding birds. Throughout I draw attention to key theoretical questions that need to be addressed and empirical issues that either facilitate or prevent strong tests of theory.

Darwin noted in 1871 that ‘no one...has paid attention to the relative numbers of the two sexes throughout the animal kingdom.’ He gathered information on the numerical proportions of males and females at birth and maturity because he reasoned that the abundance of one sex and rarity of the other affected the opportunity for sexual selection. Darwin reported a bias towards more male offspring in four of the five species of birds in his review, though the reliability of these anecdotal reports is questionable, especially as nearly equal proportions of male and female offspring were recorded for the fifth species, the only one with quantitative data. Nonetheless, Darwin puzzled over how a tendency to bias the sex of offspring could be selected for and concluded that ‘the whole problem is so intricate that it is safer to leave its solution for the future.’ Modern theory has indeed shown that the problem is intricate, and that Darwin’s notions of how selection affected sex allocation were incorrect. Yet Darwin (1871) provided the first hypothesis for how co-operation could bias sex allocation; he suggested that for animals living in herds or troops in which males defend the group, better defended groups might leave more descendants, and this pattern could result in the practice of female infanticide because males provided an advantage for subsequent reproduction. The present review shows that both theory and empirical results support Darwin’s hypothesis.

Fisher (1930) outlined how total parental expenditure on each sex was regulated by frequency dependent natural selection. Specifically, he noted (1) that the total reproductive value of males and females in a population is equal because every offspring in a sexually reproducing species has one mother and one father, and (2) that parents should invest their reproductive resources into the production of sons and daughters so as to maximize their genetic contributions to future generations. Working from these two assumptions he argued that (1) with any departure from equal investment there arises a greater return per unit of investment in the sex with lower total investment; in which case, (2) parents that make a larger investment in the sex currently receiving less investment are favoured, and this condition is maintained until total investment in the two sexes is once again equal. Fisher’s conclusion was that the population is continually drawn to an equilibrium at which total allocation to each sex is equal.

Four important points from Fisher’s equilibrium theory need emphasising for their theoretical or empirical importance. First, the theory shows that frequency-dependent selection regulates sex allocation. The current pattern of parental investment between males and females affects the profitability of additional parental investment in either sex. All theories of sex allocation are built on this premise (see Charnov 1982).

Second, the theory concerns sex allocation, which is different from sex ratio. Both the sex ratio and the patterns of parental allocation to individual sons and daughters determine sex allocation. Biased sex allocation can occur even with an equal sex ratio if individuals of one sex each receive greater investment. Likewise, a biased sex ratio can occur even when sex allocation is equal if individuals of the rarer sex receive proportionately greater resources per individual. The terms ‘sex ratio’ and ‘sex allocation’ are sometimes used interchangeably in the literature, without the caveat that this is valid only when sons and daughters have the same cost for parents. Empirical studies that measure sex ratio without measuring sex allocation of parental resources need to be cautious in using sex allocation theory to interpret the results.

Third, the theory concerns total allocation of parental resources into sons and daughters. Parental resources in this context are synonymous with parental investment as defined by Trivers (1972). Parental investment in birds generally includes more than one resource (e.g. time and energy) and each single resource is often invested in different ways (e.g. energy into food provisioning and brooding). Thus, measuring total investment is impracticable, and empirical studies are compelled to focus on just one or more constituents of investment as an index of the total. There is a need for empirical studies that integrate allocation across different parental resources and investigate whether the sex allocations of different resources match one another. For example, if nestlings of one sex are provisioned more as nestlings, do they attain independence earlier and thus receive less post-fledging care than the other sex, or do they also receive a greater allocation of post-fledging care? There is also a need for theoretical investigation of the consequences of multifaceted parental investment for total sex allocation (see Rosenheim et al. 1996). For example, are there particular conditions that increase or decrease the correlation between the sex allocations of different resources, and how are allocations of different resources related to total allocation? The issue of how different resources are allocated is especially acute in birds that breed co-operatively because parental care is protracted and the kinds of care are varied.

Fourth, sex allocation theory makes predictions about the behaviour of individual parents towards offspring. In co-operatively breeding species parents (sometimes more than two) and non-parents (sometimes kin) all contribute to care of offspring in a brood. Different individuals may have different optimal sex allocations and thus there can be conflicts of interest over the allocation of care to offspring. A practical consequence is that the sex allocation of each ‘parental’ individual needs to be measured directly; it cannot be estimated indirectly by measuring the total allocation received by male and female offspring from all caregivers. There is a also a need for theoretical investigation of how conflicts of interest in sex allocation arise and are resolved when multiple individuals invest in the same offspring. The topic has been considered in some detail for taxa other than birds (e.g. Trivers & Hare 1976; Pamilo 1991), but rather little for species with biparental care such as most birds (see Stamps 1990; Gowaty & Droge 1991; Lessells 1998). The issue of whether different caregivers have different allocations and how conflicts of interest are resolved is especially acute in co-operative breeders because there are generally more than two caregivers and often more than two genetic parents contributing to a single brood of offspring.

Fisher’s (1930) theorem has remained the foundation for studies of sex allocation, but many of its key elements or latent assumptions have become the focus for substantial modifications. Most fundamentally, Charnov (1982 and references therein) and his colleagues showed that Fisher’s implicit assumption that a change in sex allocation results in equal proportional changes in genetic returns for both sexes is critical to the conclusion that sex allocation should be equal. Fisher (1930) assumed that when a parent doubles (or halves) its investment in females (or males), it also doubles (or halves) it genetic returns from daughters (or sons). Thus, the conclusion of equal allocation depends upon males and females being equally costly to produce and on incremental changes of investment being equally valuable at all levels of initial investment. Because these two conditions are probably rarely met in nature (see below), it is important to see how relaxing them affects sex allocation.

The sexes are unequal in cost when raising one causes a greater reduction in a parent’s residual reproductive value than the other does. This condition is likely to be fulfilled whenever males and females have different life histories. For example, consider a co-operative society such as Splendid Fairy Wrens (Malarus splendens) where many individuals of both sexes disperse and must settle into an existing group to breed, but males face stiffer competition than females from current group members trying to exclude them from the group (Rowley & Russell 1990). If competitive ability has a greater effect on the likelihood of males breeding, and parental investment affects competitive ability, then a small additional parental investment in males will increase the probability of sons joining a group more than it diminishes the chances for daughters. In this case, increased investment in sons would yield a greater genetic return for parents than the loss suffered by daughters that do not receive this investment. The population, as well as individual parents, will be drawn to an equilibrium at which allocation to males is greater than that to females, because only then will the return for parents per unit investment be the same for sons and daughters (Charnov 1982). In general terms, any time the relationship between parental investment and offspring fitness differs between sons and daughters, unequal rather than equal sex allocation is expected.

The second condition necessary for equal sex allocation is that changes in investment have the same proportional effects on offspring fitness at all levels of investment. This condition seems unlikely to be met in most birds because it is violated whenever there is a nonlinear relationship between offspring fitness and parental investment. For example, a non-linear relationship between offspring fitness and investment exists when the benefit of post-fledging care in terms of survival declines with offspring age. Thus if the rate of change in survival differs consistently between sons and daughters for any reason, then unequal allocation is expected. Recall that according to Fisher’s theorem the equilibrium sex allocation is reached when any increase in allocation to one sex yields a benefit equal to the cost incurred by the other sex from the reduction in allocation. When the relationship between offspring fitness (parental return) and parental investment is nonlinear this condition cannot be satisfied unless sex allocation is unequal. There is, however, one apparent exception to the rule that nonlinear functions create unequal sex allocation. This exception would occur if the relationship between investment and return were exactly symmetrical around the point of equal allocation and the same for both sexes. Thus, despite the function being non-linear, changes in sex allocation would always give the same benefit for one sex as the cost to the other as required by Fisher’s equal allocation theorem. However, it is doubtful that such exact conditions occur in nature.

In summary, basic sex allocation theory shows that either equal or unequal sex allocation can maximise parents’ total genetic return on investment. Equal sex allocation is expected when sons and daughters cost the same, that is, when they give equal fitness returns for a given amount of parental investment. Unequal sex allocation is expected when sons and daughters achieve the same fitness expectation only with different amounts of parental investment, or when the relationship between offspring fitness and parental investment is non-linear. When the sexes differ in cost, or when investment-return relationships are non-linear, sex allocation is expected to be biased such that the rate of genetic return per unit of parental investment is the same for sons and daughters.

The key to testing sex allocation theory in nature is identifying the specific circumstances that affect either the relative cost of producing each sex or the reproductive potential of the sexes. From theory we know that anything that predictably changes the relative cost or reproductive potential of offspring will alter the investment-return relationships for parents, and should therefore affect sex allocation. Not surprisingly imaginative theorists and pragmatic field biologists have developed a variety of distinct models that incorporate specific social and ecological circumstances that can affect sex allocation. There are a number of reviews dealing with these ideas, and I refer the reader to them for a comprehensive treatment (Williams 1979; Charnov 1982; Clutton-Brock & Iason 1986; Frank 1990; Hardy 1997). Here I focus on sex allocation models in which the key ecological and social circumstances are encountered only in co-operative breeders or whose effects are more likely to be manifest in these species because of the natural history of co-operative breeding. The hypotheses are easily separated into three broad categories: local mate competition, local resource competition, and local resource enhancement. Nevertheless, these three hypotheses also share a conceptual unity. Each can be considered a special case of either negative (local mate competition, local resource competition) or positive (local resource enhancement) frequency dependent selection within a brood, similar to Fisherian frequency dependence within the whole population. For example, negative frequency dependence within a brood occurs when the function relating offspring fitness to sex ratio in the brood declines more rapidly for one sex due to competitive effects between members of the brood. Whereas, positive frequency dependence within a brood occurs when the function relating offspring fitness to sex ratio in the brood is more rapidly increasing for one sex due to mutual enhancement between members of the brood. Below I give for each of local mate competition, local resource competition, and local resource enhancement an explanation of the mechanism by which sex allocation is supposedly affected, provide a simple example of how allocation is affected, and then point to the natural history conditions that make certain species particularly suitable subjects for investigation.

The term local mate competition was coined to describe competition between siblings of one sex for mating access to either their siblings or unrelated individuals of the opposite sex (Hamilton 1967). Competition of this sort over mating opportunities, by definition, reduces the reproductive success of some or all of the competing individuals. Hamilton (1967) described the first evidence of how local mate competition can affect sex allocation. He noted that many species of parasitic wasps had female-biased sex ratios and that offspring of these species usually mated close to their birthplace. Hamilton suggested that because females have a lower reproductive capacity than males, local mating made females a limiting resource for male reproductive success. As a result, the fitness returns to parents from investment in males became diminishing before they did so for females. Thus, parents benefited from a greater allocation to females because this increased the reproductive returns from both sexes.

The example of wasps is a specific case that illustrates the general conclusion that mating competition between relatives can lead to biased sex allocation. The key is that one sex competes for matings within a more restricted arena containing a higher proportion of same sex siblings than in the population as a whole. The result is that the reproductive output of the competing sex is limited by access to members of the opposite sex. Consequently, a greater allocation to the limiting sex allows offspring of the limited sex to increase their reproductive output at the same time as the reproductive output of the limiting sex is increased, which, of course, increases the total reproductive output of the offspring and the genetic benefits to parents. More generally, local mate competition need not occur only among siblings in order to affect sex allocation; mating competition among more distantly related kin can have the same effect. The only critical factors are that related males (or females) compete locally for mates and access to mates limits the reproductive capacity of the competing sex.

There is a variety of theoretical treatments that consider the degree to which allocation should be biased as a result of local mate competition (Charnov 1982; Frank 1986; Antolin 1993; Hardy 1994). All predict that the degree of bias in sex allocation that is required to maximise parental fitness depends on the number of females whose offspring mate with one another. The smaller the number of mothers that contribute offspring that interbreed, the greater the expected bias in allocation toward the limiting sex. In the extreme case, where one mother contributes offspring to a mating group, her best allocation provides just enough investment in males to ensure that all her daughters are fully mated. At the other extreme, when all offspring compete for mates to the same extent with all other males in the population, there is no local mate competition and no expected effect of mate competition on sex allocation. Overall, the degree of bias in allocation will depend upon the relative difference between sons and daughters in the effects of local mating competition on reproductive output.

Are co-operative breeders subject to local mate competition in a way that affects sex allocation? The first condition necessary for local mate competition is that the reproductive capacity of one sex sets a limit to that of the other sex. This seems to be a nearly universal property of animal mating systems (see Kvarnemo & Ahnesjö 1996). The second requirement is that mating competition between siblings or other kin is stronger in one sex. If both sexes disperse only short distances or when both sexes remain in their natal territory until maturity, then local mate competition is likely to be equally strong in both sexes and no change in sex allocation is expected. If just one sex is philopatric and the other disperses before sexual maturity, then local mate competition is likely to be stronger in the philopatric sex and a shift in sex allocation toward the dispersing sex is expected. Finally, the conditions for local mate competition are probably all but eliminated when both sexes disperse long distances before sexual maturity. Among co-operatively breeding birds, very few species apparently have natural histories whereby individuals of both sexes routinely remain in their natal territory until maturity (Stacey & Koenig 1990; Koenig et al. 1992). In the vast majority of co-operative breeders females disperse and males are philopatric (e.g., Pukeko, Porphyrio porphyrio Craig & Jamieson 1990), so there is the potential for local mate competition among brothers and between sons and fathers. In some species where females disperse only one to two territories away and males are philopatric (e.g., Splendid Fairy-wren, Malarus splendens, Rowley & Russell 1990), there remains the possibility of local mate competition in both sexes, though it is difficult to judge whether it is stronger in one sex. Despite these possibilities there is rather limited evidence from co-operative species of whether the strength of mating competition among relatives differs between the sexes (Brown 1987, Stacey & Koenig 1990), yet these data are necessary if we are to make predictions of local mate competition effects on sex allocation.

The Local Resource Competition theory of sex allocation was formulated to account for the effects of competition between siblings (Clark 1978). Thus, in its original form Local Resource Competition is superficially rather similar to the theory of Local Mate Competition, in that it involves siblings. The key distinction between the two hypotheses is that competition is over resources rather than mates. Local resource competition amongst siblings or between offspring of one sex and unrelated individuals has the potential to affect sex allocation, if the competition involves food, territory, or other limited resources that are required for reproduction. The effect of competition is a reduction in the reproductive output of some or all of the competing individuals. The key point is that competition occurs among siblings of only one sex, or competition is more intense and has a stronger negative effect on one sex. In such cases the genetic benefit to parents per unit investment in that sex is lower, and parents can increase their genetic returns by making a larger allocation towards the non-competing or less-affected sex.

The first evidence of how local resource competition affects sex allocation was described in bush babies by Clark (1978), who observed a male-biased sex ratio in this species. Moreover, she noted that male offspring disperse when they are young but females tend to stay near their place of birth throughout life. The explanation Clark offered for the apparently male-biased sex allocation was that because sisters do not disperse they compete with one another for limited resources necessary for breeding, whereas brothers do not. As a result, the return on investment for a parent is lower for daughters than for sons, which favours parents that invest more heavily in sons than daughters. The example of bush babies illustrates a specific case of a more general conclusion. If offspring of one sex suffer more from competition with same-sex siblings, competition with other kin, or competition with unrelated individuals, then the sex suffering greater competition provides a smaller marginal, genetic return on parental investment. As a result, sex allocation is expected to be biased towards the sex experiencing less local competition. Theoretical treatments of the problem of local resource competition differ somewhat in their formulations (Charnov 1982; Silk 1983, 1984; Bulmer 1986), but all agree that the magnitude of the bias in sex allocation should be proportional to the frequency with which same sex siblings compete, or more precisely, the relative difference between sons and daughters in the effects of same-sex competition on reproductive output.

Just as with local mate competition, which was originally formulated in terms of sibling competition but can be extended to competition with other kin, local resource competition between offspring of one sex and other kin must also be considered. Whenever local resource competition by one sex results in a larger reduction in the reproductive output of the kin group then a parent is expected to bias sex allocation towards the sex exhibiting less competition. According to this view, competition between offspring and parents can also be interpreted as a form of local resource competition, and it is expected to affect sex allocation. For example, if among mature offspring that are independent of parental care, one sex has greater overlap in resource use with parents and this overlap reduces the subsequent reproductive success of those offspring, then parents should bias their sex allocation toward the other sex. Likewise, if competition between parents and mature offspring of one sex reduces the subsequent reproductive success of the parents, then parents should bias their sex allocation toward the other sex. In the latter case the effect can alternatively be regarded as an additional form of parental investment in the offspring with different timing of investment into either sex.

Are co-operatively breeding birds subject to local resource competition in a way that could affect sex allocation? Probably yes, and probably often. For example, one way in which more intense competition among siblings of one sex may originate is a sex difference in dispersal coupled with competition for limited breeding territories. In the majority of co-operatively breeding birds one sex disperses but the other breeds close to the place of birth (Stacey & Koenig 1990). Generally females disperse and males stay. Moreover, there is evidence from some species that males are philopatric in part because of a limitation on breeding vacancies and females (e.g., Superb Fairy-wrens, Malurus cyaneus, Pruett-Jones & Lewis 1990). What is now needed are more studies that seek to demonstrate explicitly whether competition amongst siblings or between parents and offspring reduces the reproductive success of the offspring or parents and whether these effects in turn affect parental allocations (see Lessells 1998).

Just as competition between kin can reduce reproductive success and favour a sex allocation that is biased away from the competing sex, co-operation between kin can increase reproductive success and favour a sex allocation biased towards the co-operating sex. The term local resource enhancement was coined to describe situations where the behaviour of adult offspring of one sex increases the ability of parents to control or benefit from local resources that are necessary for reproduction. For example, in some species territorial ownership of resources is essential for breeding and competition for territory maintenance is intense. If daughters disperse but sons are philopatric and help defend the territory, then the presence of sons could lengthen the parents’ tenure of a territory and increase their lifetime reproductive success. Under such conditions sons would be providing a greater return to parents than daughters would and parents would be expected to bias their sex allocation towards sons. This form of the local resource enhancement hypothesis is often referred to as the repayment model, because offspring of one sex essentially repay part of the parental investment they received by helping their parents to produce more offspring. Theoretical models of this problem have been developed that incorporate relevant parameters for co-operatively breeding birds (Emlen et al. 1986, Lessells & Avery 1987), and they do indeed predict that sex allocation should be biased towards the sex that repays parents by helping to raise additional offspring. Specifically, the expected equilibrium sex allocation can be predicted using data from natural populations. The degree of bias in sex allocation is expected to be proportional to the probability that an offspring helps and the effectiveness of the helping behaviour relative to the average productivity of an unassisted parent.

Local resource enhancement may also affect sex allocation independently of direct effects on parents. This is possible if co-operation between siblings enhances the reproductive success of these siblings. For example, if offspring of one sex form alliances for the defence or takeover of territories. Lessells & Avery (1987) concluded from their model that when siblings co-operate only amongst themselves, no effect on sex allocation is expected. This conclusion, however, is based upon the assumption that the benefits of co-operation are linearly related to the number of individuals co-operating. If, however, offspring of one sex form alliances and the benefits of these alliances increase exponentially with group size, then parents may be expected to bias their allocation towards this sex. Such a bias may in turn create a Fisherian frequency-dependent mating advantage for the opposite sex. The overall effect is difficult to predict without a formal model. It is unclear whether local resource enhancement among siblings would select for overall changes in the population sex allocation or simply for facultative changes, whereby equal numbers of parents bias their allocation toward one sex for local resource enhancement and toward the opposite sex for the resulting Fisherian frequency-dependent mating advantage.

Local resource enhancement has received the most attention as a cause of biased sex allocation in co-operatively breeding birds. This is because the most common form of co-operative breeding involves a breeding pair being assisted by offspring from previous broods (Stacey & Koenig 1990; Emlen 1991) and often these offspring appear to increase the reproductive success or survival of the breeding parents (Brown 1987; Stacey & Koenig 1990; Emlen 1991). The key point being that helping enhances parental fitness and that it is biased with respect to offspring sex. Under such circumstances, the repayment model would seem to apply. Thus, there is great opportunity for testing this hypothesis amongst co-operatively breeding birds.

An important distinction among models of sex allocation is between population equilibrium models and non-equilibrium models. Population equilibrium models assume that all parents are essentially identical and therefore make the same sex allocation. Non-equilibrium models do not make this assumption. Rather in these models individual differences between parents are meaningful because individual parents are assumed to make different facultative changes to their sex allocation depending upon some parental or environmental variable. Trivers and Willard (1973) first described a facultative-change model. In their model the relative fitness of sons and daughters changed with maternal condition. Below a threshold level of maternal condition one sex gives a better return on investment, and above this level of maternal condition the other sex gives a better fitness return. Hence the parent is selected to produce one sex below the threshold and the other sex above the threshold. By contrast, equilibrium models make one prediction for all individuals in the population, they cannot be applied separately to different individuals or territories if the offspring of these groupings are in reproductive competition.

Even were such models available it is difficult to imagine a comprehensive empirical test. For it would require measuring for both sons and daughters, at least all of the following: the costs of rearing, the distances and outcomes of dispersal, intensity of competition with relatives, and subsequent reproductive success of both offspring and parents. In short, in order to test a comprehensive theory that predicts optimal patterns of sex allocation in nature it is critical to have detailed life history data and measures of fitness for male and female offspring produced under the full range of environmental variation, as well as to experimentally allocate parents to these different conditions so as to demonstrate causation. Clearly an impossible task at present. A promising alternative approach is to test single hypothesised effects by contrasting the allocation of individuals in different circumstances. For example, comparing the sex allocation of individuals in closely related species that differ in the dispersal of offspring, may be one way to test whether local resource competition and/or local mate competition affect sex allocation. A difficulty with this approach is evolutionary differences between the species other than dispersal may have affected sex allocation. Another approach is to use inter-population comparisons because differences between populations are less likely due to other evolutionary changes than are differences between species. Similarly, comparisons between individuals in different social or ecological circumstances within the same population are useful for testing whether particular hypotheses cause individuals to make facultative changes in their sex allocation. A difficulty with these latter approaches is that the factor(s) causing the change in social or ecological circumstances between populations or individuals are themselves potential explanations for changes in sex allocation.

The authors considered three potential explanations for biased sex allocation in this species. First, that differences in the costs of daughters and sons arise due to differences in growth or survival during the period of parental care. This explanation, does not seem to apply because nestlings are sexually monomorphic in size and they survive equally well as fledglings. A second explanation, local resource competition, also seems to be an unlikely explanation for these patterns because it is sons that are philopatric and all daughters disperse by one year after birth. Under these circumstances, local competition is more likely to arise between sons and mothers, which is expected to bias sex allocation toward daughters rather than sons as was observed. Finally, local resource enhancement seems to be consistent with the observed bias; males are more philopatric than females, helpers are usually males, and reproductive success appears to increase with helpers. These results from Red-cockaded Woodpeckers prompted Emlen et al. (1986) to develop a formal model for the repayment version of local resource enhancement. The model confirmed that when sons repay part of their costs by helping parents, the greater return on initial investment to parents can lead to a male biased allocation. The model was corrected and elaborated by Lessells & Avery (1987).

Although both repayment models (Emlen et al. 1986; Lessells & Avery 1987) predict a male biased sex allocation for Red-cockaded Woodpeckers, and therefore provide qualitative support for this hypothesis, the quantitative predictions are not met exactly. This reinforces the point that testing sex allocation theory requires precise measurement of many natural history parameters and the interpretation of this study suffers from insufficient data to evaluate alternative influences on sex allocation. Nevertheless, this study is noteworthy because it provided the first evidence of biased sex allocation in a co-operatively breeding bird and it sparked interest in the role of local resource enhancement in sex allocation in co-operative breeders.

In a study of sex ratio of fledglings over seven breeding seasons, Ligon & Ligon (1990b) found that females with few helpers (0 to 2) produced a female biased sex ratio in their first nests of a season, but an even sex ratio in all other nests. Females with many helpers produced even sex ratios in all nests. Because starvation was extremely rare, the bias was not the product of differential starvation of the sexes as nestlings. Rather, it must have arisen during the egg stage, either as a result of sex biased ovulation or hatching success.

Using data on the natural history of Green Woodhoopoes, the authors presented three possible explanations for the sex bias in Woodhoopoes. First, because females are about 20% smaller than males, the authors surmised that males may be more expensive to rear and therefore unassisted females could be biasing their allocation away from males because they require a greater investment to achieve the same fitness. Second, because daughters are more helpful for subsequent reproductive attempts, unassisted females may be biasing their allocation to the helping sex in a manner consistent with the repayment model of local resource enhancement. Lastly, the authors suggested that biased sex allocation in Woodhoopoes may result from another form of local resource enhancement; territory turnover is high in Green Woodhoopoes, and adults of both sexes are dependent upon allies of the same sex, usually close relatives, to defend their territory against immigrant teams of related same sex individuals that attempt to usurp territories. Thus, if an unassisted female Woodhoopoe produces a daughter her security relative to groups of females in neighbouring flocks should be increased and her probability of a long territorial tenure increased. In this way, daughters also provide local resource enhancement to their mothers (and perhaps indirectly to their fathers) in the form of territory defence.

These three hypotheses are not mutually exclusive, and the data currently available are weakly supportive of them all. Strong tests of these hypotheses will require more data on the reproductive output of females with different numbers of helpers of each sex. Experiments in which the number of daughters allowed to remain on the territory are altered and then territory tenure and reproductive success are monitored would be most helpful.

The Eclectus Parrot, Eclectus roratus, is the only parrot known to breed co-operatively. The little that is known of this species in the wild suggests that breeding groups of up to ten individuals are composed primarily of males and that groups persist between seasons. Unlike the majority of birds, Eclectus Parrots develop sex-specific coloration in their beaks and plumage soon after hatching and therefore can be sexed easily. This fact, along with the popularity of the species in zoos and with private breeders, has produced evidence of remarkable biases in sex allocation in this species (Heinsohn et al. 1997). Specifically, females appear to be able to bias their sex allocation to such a degree that (in captivity at least) they are capable of producing long unbroken runs of a single sex (the maximum was 20 males in a row). Heinsohn et al. (1997) gathered data from breeders and zoo records and demonstrated statistically (1) that the sex ratio of 209 fledglings produced in captivity did not differ significantly from 50:50 and (2) that the biased sex allocation apparent in runs of a single sex probably arises at ovulation but could be further adjusted by maternal infanticide. This species provides convincing evidence that females can and do bias their sex allocation, but whether such control is adaptive in nature remains to be addressed with data from field studies.

The Western Bluebird Sialia mexicana is a facultatively co-operative breeder. The majority of reproductive attempts involve only a pair of adults. However, in a 12-year study in Carmel Valley, USA, Dickinson et al. (1996) found that for approximately 7% of nesting attempts parents had helpers, and that these helpers were virtually always sons from the parents’ previous nest that season or from a previous year. Moreover, parents that had sons helping them reduced their feeding rates to nestlings, and fledged more offspring. Thus, according to the repayment model of local resource enhancement, Western Bluebirds are expected to bias their sex allocation toward males.

Koenig & Dickinson (1996) tested this prediction, as well as predictions from four other sex allocation hypotheses using 13 years data on the sex of 2187 nestlings from 549 nests. Unlike the majority of birds, nestling Western Bluebirds display sex differences in plumage at an early age (about one week prior to fledging), so it is possible to estimate sex allocation at the nestling stage without resorting to molecular techniques. The overall sex ratio in this population was 51.9% male, which was not significantly different from 50% and was much lower than the 55.2% predicted by the repayment model. Thus, the Western Bluebird provides no support for adaptive sex allocation in relation to co-operative breeding.

A strength of Koenig & Dickinson’s (1996) study is that they also considered the predictions made by other hypotheses for sex allocation. Most studies of sex allocation focus on just one hypothesis. In the case of Western Bluebirds, in addition to the help sons provide to their parents, sex allocation could also be affected by the size dimorphism of nestlings, differential dispersal of the sexes (causing local resource competition), as well as seasonal changes in maternal condition and the condition of offspring (causing differences in the fitness of sons and daughters). Overall, Koenig and Dickinson (1996) found no support for any of these hypotheses; the sex ratio of Western Bluebirds (and by inference sex allocation) did not vary with brood reduction, mother’s age, presence of male helpers, laying date in the season, condition of the mother, or annual differences in environmental conditions. The authors concluded that studies of sex allocation must be more comprehensive than they generally have been, and that data on various aspects of natural history are required to determine which hypothetical mechanisms (if any) regularly select for biases in sex allocation. This study also underscores the need for a comprehensive predictive model of sex allocation that simultaneously considers multiple factors affecting sex allocation.

The Bell Miner Manorina melanophrys is an obligately co-operative breeder in which breeding pairs of birds are always assisted in rearing young by related and unrelated helpers. Three of the principal conditions necessary for local resource enhancement effects on sex allocation appear to be met in this species. First, the majority of helpers are males and often they are sons of the breeders. Second, daughters disperse at about 8 months of age and only help occasionally at their parents’ nest before dispersing. Third, there is a positive relationship between the number of helpers and the number of offspring fledged (Clarke 1989). Thus, according to the repayment model of local resource enhancement, a male biased sex allocation is expected in this species.

Jones (1998) and collaborators studied sex allocation in Bell Miners using a molecular marker to sex nestlings. They determined the sex of 230 offspring sampled over four different breeding seasons. Consistent with the predictions of the repayment model, 60% of the offspring were male, which is significantly more than half, and the bias towards males occurred in all four breeding seasons in which they gathered samples. Moreover, they assessed the costs of rearing sons and daughters by comparing relative weights of male and female nestlings and by comparing feeding rates at male- and female- biased nests. Thus, they were able to estimate sex allocation in the form of parental feeding and not just sex ratio. They found no significant differences in the apparent costs or resources supplied to sons versus daughters, suggesting that in this species sex ratio might be a reasonable index of sex allocation up to fledging. A limitation, however, of this study with regard to the repayment model was that Jones (1998) could not find empirical evidence that sons increased the reproductive success of their parents. Thus, while the results are suggestive of biased sex allocation, they are equivocal in supporting local resource enhancement as the explanation. To do so, data are needed to confirm that sons repay part of their cost by increasing the reproductive success or survival of their parents or brothers.

A further interesting pattern of sex allocation emerged from this study. There was a seasonal trend in which the proportion of females produced declined over the course of the six-month long breeding season. In Bell Miners, females are more likely to disperse than males and will do so to take up breeding vacancies as early as 8 to 10 months of age. Jones (1998) speculates that the seasonal trend is adaptive because it allows early hatching females to reach the age for dispersal by the start of the following breeding season. These data are consistent with the argument that when there are sex-specific differences in offspring maturation time, parents should bias their sex allocation seasonally so that early in the season they favour the sex whose lifetime reproductive success is most affected by hatching date (Daan et al. 1996). Data are needed, however, to demonstrate that early hatching females are indeed advantaged in gaining a breeding vacancy as speculated.

In the Pied Kingfisher Ceryle rudis, sons occasionally help their parents to raise offspring whereas daughters do not. Thus, there appears to be opportunity for local resource enhancement by sons to lead to a male biased sex allocation. In a study of two different breeding sites in Kenya, Reyer (1990) reported an equal sex ratio of fledglings. Specifically, using clutches in which all offspring were sexed, he found there were 18 males to 20 females at Lake Victoria and 55 males to 49 females at Lake Naivasha. Using the repayment model of Lessells & Avery (1987) along with data on the frequency of helping by presumed kin and its apparent effects on the success of breeding attempts, Reyer (1990) calculated that sex allocation is expected to be male biased in the Lake Victoria population and near equality at Lake Naivasha. As neither population apparently differed from equal sex allocation, these data are equivocal with regard to the repayment model of local resource enhancement. Moreover, without better evidence that helpers are indeed sons of the adults they assist, these data provide only a weak test of sex allocation theory.

The Seychelles Warbler Acrocephalus sechellensis is a rare endemic of the Seychelles Islands. For the past decade this species has been the subject of an extraordinarily intensive study focusing on reproductive biology and co-operative breeding. Komdeur and his colleagues (1996; 1997) have also provided the most convincing evidence yet for adaptive facultative changes in sex allocation in birds.

Amongst Seychelles Warblers, offspring of both sexes will disperse to fill available breeding vacancies but males and, to a greater extent, females are philopatric when vacancies are of lower quality than their natal territories. Philopatric daughters (and infrequently sons) help their parents raise offspring in subsequent nesting attempts. On low-quality territories that have less vegetation, and less insect food, the presence of philopatric offspring actually reduces parents’ reproductive success, presumably due to food limitation, whereas on high quality territories the presence of one or two helpers increases the reproductive success of parents through effects of their helping, and the presence of more than two helpers again reduces parents’ reproductive success apparently due to resource limitation. Thus, the natural history of the Seychelles Warbler appears to satisfy the conditions for both local resource enhancement and local resource competition, depending on the particular territorial and social circumstances in which parents find themselves.

Females of this warbler species generally nest only once per year and 91% of nests contain just one egg. Komdeur and his colleagues (1996; 1997) gathered blood samples from over 200 nestlings from three breeding seasons and used a molecular marker to determine sex. There was so little mortality from egg laying to sampling of nestlings to fledging of offspring that the sex ratio of nestlings provides an accurate index of sex allocation in this species and the results are robust even when potential biases in mortality are assumed. The results provide remarkable evidence for adaptive biases in sex allocation.

As predicted by the local resource competition hypothesis, females on low quality territories biased their sex allocation towards the dispersing sex (producing 77% sons), whereas females on high quality territories biased their sex allocation towards the more philopatric sex (producing 88% females). Moreover, experiments in which the same parents were transferred between islands confirmed the sex ratio differences were due to effects of territory quality. Seven pairs were translocated. The four pairs that moved from a low to a high quality territory switched from producing 90% sons to producing 85% daughters. Three pairs that resided in high quality territories on both islands showed no change in sex allocation, producing 80% daughters on both islands.

As predicted by the repayment model of local resource enhancement, females on high quality territories with just one resident helper biased their allocation towards the philopatric helping sex (producing 85% daughters), whereas females with two helpers already present produced 93% sons. Moreover, an experiment confirmed that this bias in sex allocation was apparently for the purpose of producing helpers; when six females on high quality territories had one of their two helpers experimentally removed they switched from producing all sons to producing 83% females.

The Seychelles Warbler provides the most convincing evidence of biased sex allocation in birds of any mating system, and it is difficult to see how the evidence of facultative biases could be improved upon. Perhaps the most profitable approach would be to use data from this species to construct models that incorporate multiple hypotheses (such as local resource competition and enhancement) to make quantitative predictions about the equilibrium sex allocation expected in the population as a whole. Thus, the strong support for facultative change models would provide a foundation for testing equilibrium models. The large skews in this species, like in the Eclectus Parrot, also provide an unprecedented opportunity to study the proximate mechanisms used for biasing allocation. Lastly, the Seychelles Warbler illustrates that the greatest power in studies of sex allocation comes not from determining the sex of offspring but rather from having comprehensive data on the life history and fitness of individuals with which to interpret the observed allocation.

In summary, the data on sex allocation in co-operative breeders are fragmented. Too few species have been studied for any patterns to emerge about the most common patterns of sex allocation or causes of bias. Moreover, amongst studies reporting a bias in sex allocation, most suffer from having too little supporting data to make strong conclusions about whether the observed biases are adaptive. Nevertheless, studies of co-operatively breeding birds are noteworthy for being among the longest running and most detailed investigations of social behaviour and fitness to be found for vertebrates. Thus, it seems certain over the next few years there will be an explosion of molecularly derived data on sex allocation in co-operatively breeding birds.

I thank Kate Lessells and Jim Quinn for the invitation to participate in the symposium. Kate, Jim and Bob Montgomerie made many helpful comments on the ms. Dave Jones kindly gave permission to cite his MSc. thesis. NSERC (Canada) provided support through a post-doctoral fellowship and the American Ornithologists' Union provided funds for travel.

Antolin, M.F. 1993. Genetics of biased sex ratios in subdivided populations: models, assumptions, and evidence. Oxford Surveys in Evolutionary Biology 9:239-281.

Brown, J.L. 1987. Helping and communal breeding in birds. Princeton, NJ; Princeton University Press: 354.

Bulmer, M.G. 1986. Sex ratio theory in geographically structured populations. Heredity 56:69-73.

Charnov, E.L. 1982. The theory of sex allocation. Princeton, NJ; Princeton University Press:

Clark, A.B. 1978. Sex ratio and local resource competition in a prosimian primate. Science 201:163-165.

Clarke, M.F. 1989. The pattern of helping in the Bell Miner (Manorina melanophrys). Ethology 80:292-306.

Clutton-Brock, T.H. 1986. Sex ratio variation in birds. Ibis 128:317-329.

Clutton-Brock, T.H. &  Iason, G.R. 1986. Sex ratio variation in mammals. Quarterly Review of Biology 61:339-374.

Craig, J.L. & Jamieson, I.G. 1990. Pukeko: different approaches and some different answers. In P. B. Stacey and W. D. Koenig (eds.), Co-operative breeding in birds. Cambridge; Cambridge University Press: 385-412.

Daan, S.,  Dijkstra, C. & Weissing,F.J. 1996. An evolutionary explanation for seasonal trends in avian sex ratios. Behavioral Ecology 7:426-430.

Darwin, C. 1871. The descent of man, and selection in relation to sex. London; John Murray:

Dickinson, J.L.,  Koenig, W.D. & Pitelka, F.A. 1996. Fitness consequences of helping behavior in the Western Bluebird. Behavioral Ecology 7:168-177.

Emlen, S.T. 1991. Evolution of co-operative breeding in birds in mammals. In J. R. Krebs and N. B. Davies (eds.), Behavioural Ecology. Oxford; Blackwell Scientific: 301-337.

Emlen, S.T.,  Emlen, J.M. & Levin, S.A. 1986. Sex-ratio selection in species with helpers-at-the-nest. American Naturalist 127:1-8.

Fisher, R.A. 1930. The genetical theory of natural selection. Oxford; Clarendon Press:

Frank, S.A. 1986. Hierarchical selection theory and sex ratios I. General solutions for structured populations. Theoretical Population Biology 29:312-342.

Frank, S.A. 1990. Sex allocation theory for birds and mammals. Annual Review of Ecology and Systematics 21:13-55.

Gowaty, P.A. 1991. Facultative manipulation of sex ratios in birds. Rare or rarely observed? Current Ornithology 8:141-171.

Gowaty, P.A. & Droge,D.L. 1991. Sex ratio conflict and the evolution of sex-biased provisioning in birds. Acta XX Congressus Internationalis Ornithologici 932-945.

Gowaty, P.A. & Lennart, M.R. 1985. Sex ratios of nestling and fledgling Red-cockaded Woodpeckers (Picoides borealis) favour males. American Naturalist 126:347-353.

Hamilton, W.D. 1967. Extraordinary sex ratios. Science 156:477-488.

Hardy, I.C.W. 1994. Sex ratio and mating structure in the parasitoid Hymenoptera. Oikos 69:3-20.

Hardy, I.C.W. 1997. Possible factors influencing vertebrate sex ratios: an introductory overview. Applied Animal Behaviour Science 51:217-241.

Heinsohn, R., Legge,S. & Barry, S. 1997. Extreme bias in sex allocation in Eclectus Parrots. Proceedings of the Royal Society of London - Series B: Biological Sciences 264:1325-1329.

Jones, D.A. 1998. Parentage, mate removal experiments and sex allocation in the co-operatively breeding Bell Miner, Manorina melanophrys. MSc. Queen's University, Kingston.

Koenig, W.D. & Dickinson, J.L. 1996. Nestling sex-ratio variation in Western Bluebirds. Auk 113:902-910.

Koenig, W.D., Pitelka, F.A., Carmen, W.J.,  Mumme, R.L. & Stanback, M.T. 1992. The evolution of delayed dispersal in co-operative breeders. Quarterly Review of Biology 67:111-150.

Komdeur, J. 1996. Facultative sex ratio bias in the offspring of Seychelles Warblers. Proceedings of the Royal Society of London - Series B: Biological Sciences 263:661-666.

Komdeur, J., Daan, S., Tinbergen,J. & Mateman, C. 1997. Extreme adaptive modification in sex ratio of the Seychelles Warblers eggs. Nature 385:522-525.

Kvarnemo, C. & Ahnesjö, I. 1996. The dynamics of operational sex ratio and competition for mates. Trends in Ecology and Evolution 11:404-408.

Leimar, O. 1996. Life-history analysis of the Trivers and Willard sex-ratio problem. Behavioral Ecology 7:316-325.

Lessells, C.M. 1998. A theoretical framework for sex-biased parental care. Animal Behaviour in press.

Lessells, C.M. & Avery, M.I. 1987. Sex ratio selection in species with helpers at the nest: some extensions of the repayment model. American Naturalist 129:610-620.

Ligon, J.D. & Ligon, S.H. 1990b. Green Woodhoopoes: life history traits and sociality. In P. B. Stacey and W. D. Koenig (eds.), Co-operative breeding in birds. Cambridge; Cambridge University Press: 31-66.

Ligon, J.D. & Ligon, S.H.. 1990b. Female-biased sex ratio at hatching in the Green Woodhoopoe. Auk 107:765-771.

Pamilo, P. 1991. Evolution of colony characteristics in social insects. I. Sex allocation. American Naturalist 137:83-107.

Pruett-Jones, S.G. & Lewi, M.J. 1990. Sex ratio and habitat limitation promote delayed dispersal in Superb Fairy-wrens. Nature 348:541-542.

Quinn, T.W. 1999. Sex and the single chromosome. In: Adams, N.J. & Slotow, R.H. (eds) Proc. 22 Int. Ornithol. Congr., Durban: 434-449. Johannesburg: BirdLife South Africa.

Reyer, H.U. 1990. Pied Kingsfishers: ecological causes and reproductive consequences of co-operative breeding. In: P. B. Stacey and W. D. Koenig (eds.), Co-operative breeding in birds. Cambridge; Cambridge University Press: 527-558.

Rosenheim, J.A., Nonacs, P. & Mangel, M. 1996. Sex ratios and multifaceted parental investment. American Naturalist 148:501-535.

Rowley, I.C.R. & Russell, E.. 1990. Splendid Fairy-wrens: demonstrating the importance of longevity. In P. B. Stacey and W. D. Koenig (eds.), Co-operative breeding in birds. Cambridge; Cambridge University Press: 1-30.

Silk, J.B. 1983. Local resource competition and facultative adjustment of sex ratios in relation to competitive abilities. American Naturalist 121:56-66.

Silk, J.B. 1984. Local resource competition and the evolution of male-biased sex ratios. Journal of Theoretical Biology 108:203-213.

Stacey P.B. & Koenig, W.D. 1990. Co-operative breeding in birds. Cambridge; Cambridge University Press: 615pp.

Stamps, J.A. 1990. When should parents differentially provision sons and daughters? American Naturalist 135:671-685.

Triver, R.L. & Hare, H. 1976. Haplodiploidy and the evolution of the social insects. Science 191:249-63.

Trivers, R.L. 1972. Parental investment and sexual selection. In B. Campbell (eds), Sexual Selection and the Descent of Man. 1871-1971. Chicago; Aldine: 136-179.

Trivers, R.L. & Willard, D.E. 1973. Natural selection of parental ability to vary the sex ratio of offspring. Science 179:90-92.

Williams, G.C. 1979. The question of adaptive sex ratio in outcrossed vertebrates. Proceedings of the Royal Society of London - Series B: Biological Sciences 205:567-580.