S47.5: Reproductive Skew in cooperatively-breeding-birds: An overview of the issues

Section of Neurobiology and Behavior, Cornell University, W323 Mudd Hall, Ithaca, NY , 14853-2702, USA, fax 607 254 4308, e-mail ste1@cornell.edu

Emlen, S.T. 1999. Reproductive Skew in cooperatively-breeding-birds: An overview of the issues. In: Adams, N.J. & Slotow, R.H. (eds) Proc. 22 Int. Ornithol. Congr., Durban: 2922-2931. Johannesburg: BirdLife South Africa.

The term reproductive skew refers to the distribution of reproduction among individuals living within social groups. High skew societies are those in which reproduction is monopolised by one or a few dominant individuals; low skew societies are those in which reproduction is distributed more equitably among all group members. Two general theoretical frameworks exist for explaining variation in reproductive allocation: optimal skew (concessions) theory and incomplete control theory. The two frameworks represent very different views of conflict resolution within societies. I discuss several misconceptions surrounding optimal skew models. These include: (1) the semantic error that because dominants are said to 'concede' reproduction to subordinates, no overt conflict over breeding is expected, (2) the view that many cooperatively breeding birds do not fit the conditions for optimal skew models to apply because helpers do not always increase the nesting success or survival of the breeders they assist, and (3) the idea that inbreeding avoidance represents a third, and mutually exclusive, theoretical framework for understanding reproductive skew. I then describe three areas in which the predictions of optimal skew and incomplete control models differ, areas that I believe should become foci for future empirical testing of the theories. These involve: (1) the role of ecological constraints, (2) the importance of genetic relatedness, and (3) the expected patterns and levels of aggression. I argue that nuclear family (singular breeding) species are poor candidates, but that extended family (plural breeding) species are promising candidates, for differentiating between the different skew theories. Finally, I list five areas where optimal skew models need to be expanded by incorporating additional factors that will make them more applicable to vertebrate organisms.

When Morne du Plessis and I first proposed a symposium on reproductive skew in cooperatively breeding birds, our goal was to draw attention to an understudied question. That question, simply stated, is: What determines the allocation of reproduction among sexually mature individuals living within breeding groups? More specifically, why is it that in some species a single, dominant pair monopolises breeding, while in others all individuals share equitably in the group's reproduction? Among cooperatively breeding birds, we find both of these extremes and everything in between.

Reproductive skew is the term given to the apportionment of reproduction among same-sex individuals living within a social group. High skew societies are those in which reproduction is concentrated on one or a few dominant breeders; low skew societies refer to those in which reproduction is distributed more evenly among group members. The term 'reproductive skew' carries no theoretical connotations; it is merely the descriptor of the amount of reproductive sharing that occurs within the social group(s) of interest.

Studies of cooperatively breeding birds have played a major role in the development of social evolution theory for many decades. From the first published observations of helping-at-the-nest by Alexander Skutch in 1935 (Skutch 1935), through the pioneering quantitative studies of fairy-wrens and choughs by Ian Rowley in the 1960's (Rowley 1965a, b), to the recent surge of studies combining detailed behavioural with molecular parentage data today, cooperatively breeding birds have been at centre stage in the theatre of ideas in behavioural ecology.

The focus of research efforts on cooperatively breeding birds, however, has changed dramatically over the years. Early studies were primarily descriptive, verifying the existence of 'helping' behaviours and cataloguing the diversity of types of helping associations found in nature. After W. D. Hamilton published the theory of kin selection in 1964 (Hamilton 1964), there followed a surge of interest in examining the importance of genetic relatedness in explaining the seemingly altruistic behaviour of helpers. It was soon discovered that most (but not all) avian co-operative groups form when offspring delay their dispersal beyond the age of sexual maturity and remain with their parents on the natal territory. More recent work has emphasised the multi-generational family nature of such groups and has sought to find general rules that govern avian 'family dynamics' (e.g. Emlen 1995, 1996, 1997).

Not all families, however, are structured alike. In simple, nuclear families, only a single male and female of the group breeds; in extended families, two or more group members of one or both sexes reproduce. In the terminology of Jerram Brown (Brown, 1987), simple families represent singular breeding groups; extended families, plural breeding groups. Singular breeders, by definition, are characterised by high skew; plural breeders, by moderate to low skew.

Both vertebrate and invertebrate co-operative breeders can be arrayed along a common axis on the basis of skew values (Sherman et al. 1995). The heuristic value of such a 'eusociality continuum' lies in encouraging cross-taxonomic thinking about possible common mechanisms underlying variation in skew. How, then, can we explain the variation in the partitioning of reproduction found among cooperatively breeding birds?

Currently there are two theoretical frameworks that attempt to answer this question. One is optimal skew theory (hereafter OSM, to refer to optimal skew models); the second is incomplete control theory (hereafter ICM, for incomplete control models).

Optimal skew theory, also called concessions theory (Clutton-Brock 1998), refers to a growing set of models based on the idea of a social contract. These models assume that 1) the dominant breeder benefits from the activities of a same-sex subordinate, and 2) the dominant is able to control reproduction by the subordinate.Under these assumptions, a dominant may sometimes benefit by conceding some reproduction to the subordinate as an incentive for the latter to remain peacefully in the group. OSM attempt to explain the degree of skew by predicting the conditions under which dominant breeders should yield just enough reproduction to make it favourable for a subordinate to stay in the group and co- operate peacefully, rather than to leave the group and attempt to reproduce on its own or fight to usurp the dominant position within the group for itself. OSM incorporate ecological, social (dominance), and genetic relatedness factors to predict when such a concession is expected, and what its magnitude will be (Vehrencamp 1979, 1983a, b; Emlen 1982, 1995, 1997; Emlen & Vehrencamp 1983; Reeve 1991; Reeve & Ratnieks 1993; Keller & Reeve 1994; Reeve & Keller 1997).

Incomplete control models, in contrast, assume that (1) the dominant breeder has only limited ability to control a subordinate's reproduction, and (2) the dominant and subordinate directly compete with one another to increase their respective shares of the group's total reproductive output. The effort expended in this competition, however, has a cost and that cost is measured in reduced group productivity. This is logical since effort expended in competition is effort not put into reproduction. ICM solve for the optimal amount of effort that each participant will invest in this competition and calculate the resulting degree of skew (Reeve et al. 1998).

Differentiating between these two theoretical frameworks is important because the incomplete control and concessions models, respectively, 'are examples of two different general views of intra-societal evolution: the tug-of-war view, in which group members engage in a struggle over resources, and the transactional view, in which group members exchange parcels of reproduction to induce beneficial behaviour from each other' (Reeve et al. 1998, p 267).

In my opinion, there are several widespread misunderstandings surrounding optimal skew theory which are hampering efforts to empirically tests its predictions. I will attempt to clarify three of these.

The first misunderstanding is semantic. In optimal skew models, dominant breeders are varyingly described as 'offering incentives', 'forfeiting fitness' , or 'making concessions'to subordinates (Reeve & Ratnieks 1993; Emlen 1982; Clutton-Brock 1998, respectively). Participants are said to form 'social contracts' with one another.

If one interprets these semantics literally, one might mistakenly conclude that the partitioning of reproduction is arrived at peacefully and harmoniously when optimal skew models apply, i.e. that aggression and conflict are not expected under OSM. Clutton-Brock, in a recent critique of OSM (1998, p. 289), stated 'The presence of overt conflict suggests that subordinate reproduction might occur where dominants are unable to prevent subordinates breeding or where the costs of suppressing subordinates outweigh the benefits'. Accurate assessment of relative status and resource holding power, however, are essential elements of both OSM and ICM (a point also acknowledged by Clutton-Brock, ibid, p. 289). Within the context of OSM, aggression is simply part of the mechanistic process by which reproductive control is established and the enforcement of the stable partitioning of reproduction is maintained. In fact, one predicted outcome of optimal skew models is that levels of both 'testing' and 'assertion' of status (i.e. of aggression) will increase as the degree of monopolization of reproduction increases (Reeve & Ratnieks 1993; Reeve & Keller 1997). The greater the fitness differential between status levels, the greater the benefits of challenging the dominant (for a subordinate) and of suppressing such challenges (for a dominant). Simple observations of intense aggression and conflict over reproduction are thus consistent with both OSM and ICM and do not constitute evidence differentially supporting either model.

OSM are based on the premise that a dominant breeder benefits from the presence and activities of a same-sex subordinate. Adopting the now-standard notation of OSM, if we let k equal the expected total output of the group if a subordinate stays, relative to a value of 1.0 if the subordinate leaves, then OSM only apply when k > 1. When this inequality does not hold and k < 1, there is no benefit to the dominant in retaining the subordinate and no reproductive incentive or concession is expected.

There are a number of studies that report no significant effect of helpers (i.e. subordinates) on nesting success or survival of breeders (i.e. dominants) (e.g. Ligon & Stacey 1989; Cockburn 1998). Such studies are sometimes taken as evidence that the range of social associations for which OSM apply is limited (e.g. Clutton-Brock 1998; Heinsohn et al. 1999; Magrath 1999). If these cases truly represent species or populations in which k is (and historically has been) consistently less than 1, then I agree that OSM will not apply to them.

Typically, however, k is calculated on the basis of annual data. But virtually all cooperatively breeding birds live for many years. If k is calculated on the basis of lifetime gains from the social association, then I suspect that cases where k < 1 will prove to be rare. This is because parents often retain offspring when ecological constraints are severe. Brown & Brown (1984) were the first to propose that, under such conditions, parents (i.e. dominants) may gain future fitness benefits by facilitating the process by which their young (i.e. subordinates) achieve breeding status, even if such facilitation involves a current fitness cost. One frequent 'route' by which young achieve breeding status in cooperatively breeding species is through inheritance, by taking over the parental territory and breeding position at the time of the death or emigration of the parent (Brown 1987; Stacey & Koenig 1990; Emlen 1991). By increasing the likelihood that their offspring will become established as breeders in the future, parents increase their number of grand-offspring. When such 'parental facilitation' effects (Brown & Brown 1984) are incorporated into calculations of k, cases where dominants actually lose lifetime fitness from associating with subordinates (where k is truly less than one) may largely disappear. If this reasoning is correct, then the scope of social associations for which OSM will potentially apply will be broadened to encompass most cooperatively breeding birds.

Some workers, including myself, have considered inbreeding avoidance to be a third theoretical framework for understanding reproductive skew (Emlen 1996; see also Clutton-Brock 1998; Heinsohn et al. 1999; Magrath 1999). The active avoidance of incestuous matings appears to be a common feature of many vertebrates, including most cooperatively breeding birds (e.g. Emlen 1995; Koenig et al. 1998; but see Heinsohn et al. in this symposium for an exception). Cooperatively breeding species typically live in familial associations, and within families most group members are close genetic relatives. Thus, for mature offspring, the only potential sexual partners residing within the social group are often its opposite-sex siblings and parent.

Inbreeding avoidance, when it occurs, can exert a strong influence on resulting skew (Emlen 1996, Magrath 1999). But, in my current thinking, there is little justification for treating it as a true alternative to either OSM or ICM. The avoidance of inbreeding is not an absolute. One cannot dichotomise species or populations in an all-or-none fashion on the basis of whether they do or do not exhibit incest avoidance. Rather there are costs to incest that reduce the value of inbred offspring (e.g. inbreedng depression). But there are also costs to not breeding at all if unrelated mates are unavailable within the group (Bengtsson, 1978; Waser et al. 1986) as well as to mating outside, but attempting to rear offspring within, the family association (Reeve & Keller 1996).

The costs and benefits of inbreeding can be quantitatively incorporated into both OSM and ICM (although I acknowledge that this has not yet been done). In one possible rendition (Reeve & Emlen, unpublished manuscript), inbreeding would increase genetic relatedness between parent and offspring but decrease the reproductive value of such offspring (because of their reduced survival and lower genetic quality). Incorporation of such costs and benefits into existing models will allow determination of the importance of inbreeding depression relative to other factors in determining the resulting skew.

OSM and ICM are mutually exclusive explanations for reproductive partitioning in social groups. They are not additive; they are not complimentary. They represent very different views on how conflict is resolved within social groups. Incest avoidance, by contrast, is not incompatible with either OSM or ICM. It makes no unique predictions of its own. For this reason, I believe that the importance of inbreeding avoidance will be better understood by incorporating its costs and benefits into existing theoretical frameworks rather than treating it as a separate alternative hypothesis.

How then can we empirically discriminate between the two primary theoretical frameworks for explaining variation in reproductive skew? As we have seen from the papers in this symposium, the task is often not an easy one.

The recent formalisation of two incomplete control models (Reeve et al. 1998) has generated three predictions that should ease the task of differentiating between the alternatives.

In the ICM, the subordinate's share of reproduction is largely unaffected by the magnitude of ecological constraints, whereas it varies inversely with the severity of such constraints in OSM. This is because, in OSM, the dominant concedes just enough reproduction for a subordinate to reach its 'break-even point', the point at which the subordinate's benefits of staying just exceed those of leaving. This break-even point is strongly affected by the expected success associated with dispersal and independent reproduction (i.e., by the magnitude of ecological constraints). In ICM, in contrast, a subordinate always reproduces above this point (i.e. it obtains more than its staying incentive); hence the subordinate is favoured to remain in the group irrespective of its opportunities for independent breeding. If a subordinate fails to achieve this 'break-even' amount of personal reproduction through competitive effort in the ICM, the dominant is predicted to yield this same amount in the OSM. Otherwise the subordinate will leave and the dominant will lose the benefits of grouping. At this point the ICM become identical to the optimal skew model (Reeve et al. 1998).

In the ICM, the subordinate's share of reproduction is largely insensitive to, or increases with, the subordinate's genetic relatedness to the dominant. This is because the optimal levels of competitive effort of both dominant and subordinate decrease with increasing relatedness, and the effects of these changes roughly cancel out their effects on skew (Reeve et al. 1998, Fig. 1).

This insensitivity contrasts sharply with the corresponding prediction from OSM, in which a subordinate's share of reproduction decreases as relatedness increases. This follows because related subordinates will gain larger indirect fitness benefits from the enhanced productivity of the group than will unrelated individuals; they thus require smaller incentives to be induced to remain and co-operate peacefully.

In the ICM, both dominant and subordinate are predicted to expend less effort in competition over their share of reproduction (i.e. to be less aggressive) as relatedness between them increases. Further, for all values of relatedness, the models predict higher levels of attempted selfishness by subordinates than by dominants (Reeve et al. 1998).

In OSM, in contrast, aggression levels are predicted to increase as relatedness increases. High relatedness fosters high skew, and as skew increases, so too does the fraction of 'disputable' reproduction. Disputable reproduction is the gap between the minimal reproductive share required for a subordinate to stay and co-operate and the maximal share above which a dominant will gain by evicting the subordinate (Reeve & Keller, 1997). As this fraction increases, the degree to which dominant and subordinate can try to selfishly augment their reproductive shares without destabilising the group also increases. Thus, the predicted frequency of conflict should increase as relatedness (and skew) increases (Reeve & Ratnieks 1993; Reeve & Keller, 1997).

Nuclear family groups are the most common form of helper-at-the-nest associations among birds. Unfortunately, they are poor candidates for testing skew theory (Emlen 1996; Heinsohn et al. 1999; Magrath 1999). Why is this so?

First, parent and offspring are typically asymmetrically (i.e. unequally) related to one another's offspring. Assuming social parents are genetic parents and no parental mate change occurs, an offspring is related to future offspring of its parents by 0.50, whereas a parent is related to future offspring of its offspring (i.e. its grand-offspring) by 0.25. When a family consists only of two dominant parents and a single, mature subordinate offspring, neither OSM nor ICM predict subordinate reproduction (Reeve et al. 1998). This is because, to a subordinate, rearing full siblings is genetically equivalent to rearing its own offspring. Simply stated, the subordinate does not increase its inclusive fitness through personal reproduction in either model. This is true even before incorporating incest avoidance, which would simply reinforce total skew (Emlen 1996).

Even when relatedness between a parent and offspring is symmetrical, which will be true if the parent changes mates between reproductive events or if a high level of extra-pair fertilisations or intra-specific brood parasitism occurs, the built-in generational asymmetry in age and dominance between parent and offspring will continue to promote high skew under both models (Emlen 1996; Magrath, 1999).

The difficulties inherent in using small nuclear family groupings to differentiate between skew models were confirmed by Magrath who, despite five years' data on scrubwrens, was still unable to clearly refute either hypothesis (Magrath 1999).

Tests of the different skew models are somewhat easier when nuclear families are of larger size. This is because when families contain two or more full-sibling offspring, OSM but not ICM predict that subordinate offspring will sometimes reproduce, especially when ecological constraints are weak (Reeve et al. 1998; equation 9).

The best opportunities for differentiating between alternative skew theories in cooperatively breeding birds lie, I believe, in studies of species that (1) live in extended family groups (Emlen 1997), and (2) exhibit clear group productivity benefits (i.e. k>>1). In such species, more than one breeder of one or both sexes is the norm, so the potential for variation in skew is large. Further, because there often are multiple breeders, the complexity of group genealogies increases so the potential for variation in relatedness values is also large. (This should be true for relatedness of different dyads within groups, as well as for average relatedness of members across different groups.) Finally, assuming variability exists in such factors as territory quality, levels of critical resources controlled, and opportunities for dispersal and independent breeding, variability can be expected in both the magnitude of group productivity benefits and the severity of ecological constraints.

It then becomes possible to examine variation in the partitioning of reproduction at several levels---among different populations, among different social groups within the same population, and within the same social groups across different years. Provided that differences exist among the units of comparison (populations, groups, or years) in the parameters of interest (genetic relatedness, magnitude of group productivity benefits, and severity of ecological constraints on independent breeding), one can ask whether such differences are significant predictors of skew variation.

Since OSM and ICM predict different resulting patterns of skew with differences in both genetic relatedness and the severity of ecological constraints, the likelihood of cleanly differentiating between the theoretical alternatives is heightened.

The final step in any testing of skew theories must be experimental. The causality underlying any significant correlations that are found in analyses of naturally occurring patterns of variation in skew must be examined experimentally. This could be done by manipulating the critical parameters of interest to determine whether resulting skew can be changed and, if so, whether the changes are in the directions and of the amounts predicted by OSM or ICM.

Studies examining reproductive skew are only beginning to be undertaken on cooperatively breeding birds. Promising subjects among plural breeding, separate nesting species include miners of the genus Manorina in Australia, bee-eaters of the genus Merops and starlings in the genus Spreo in Africa, and numerous jays of the genera Aphelocoma, Cyanocorax and Calocitta in the Americas. Among plural breeding, joint nesting species, promising subjects include various rails and gallinules of the genera Gallinula, Porphyrio and Tribonyx, the anis and guiras (genera Crotophaga and Guira) and some woodpeckers in the genus Melanerpes.

What, then, are the next steps for research on the partitioning of reproduction within social associations? There is a need for further theoretical development of both optimal skew (concessions) and incomplete control models. Fortunately there is an emerging consensus on what extensions to current OSM are most needed (Clutton-Brock 1998; Magrath 1999; Johnstone et al. 1999; Reeve & Emlen unpublished manuscript.). These include the following:

1. The incorporation of the costs and benefits of inbreeding

By allowing for increased relatedness of inbred offspring, but decreased reproductive value of such offspring, such additions may render 'incest avoidance' null and void as an independent, competing hypothesis to either OSM or ICM (Reeve & Emlen, unpublished manuscript.)

By including future (lifetime) costs and benefits of current social associations in calculations of expected skew, OSM will become better tailored to fit the life history parameters of long-lived, family-dwelling, bird species. Under some circumstances, retention of offspring may be costly to parents in the short-term yet beneficial in the long-term. This is most likely to occur when ecological constraints are severe and continued tolerance increases an offspring’s likelihood of inheriting or otherwise successfully obtaining a breeding position within the population (thus producing more grand-offspring for the original parent). I anticipate that incorporation of such 'parental facilitation' effects ' (sensu Brown & Brown 1984) will broaden the range of conditions for which OSM are potentially relevant to encompass virtually all avian co-operative breeders (Reeve & Emlen unpublished manuscript).

Current models deal with dyads of same-sex individuals. Yet many social associations consist of more than mere dyads. Complex alliances and social coalitions occur regularly in higher vertebrates. This is especially true of plural breeders among cooperatively breeding birds, and of social carnivores and primates among mammals. Work is just beginning on expanding skew models to multi-member groups (e.g. Johnstone et al. 1999).

Current skew models deal only with same-sex resolutions, although they can be applied to either males or females. The genetic interests of the two sexes, however, are seldom congruent. What happens when the fitness of an individual of one sex is strongly influenced by the pattern of reproductive sharing of the other? The work of Davies and colleagues on polyandrously-mated dunnocks Prunella modularis provides such an example (Davies 1992). The nesting success of a female paired with two mates is significantly enhanced only if both males contribute to the care of the young in her nest. But the contribution of the subordinate male is dependent upon his sexual access to the female during her prior fertile period. The female seemingly encourages, while the dominant male actively discourages, such access. This problem is similar to the 'third party' problem mentioned above, but with the added complexity of potential inter-sexual influences on the resulting magnitude of concessions offered.

Most current skew models assume simultaneous decisions by dominant and subordinate (see Cant 1998, for an exception). This assumption may be unrealistic for some cooperatively breeding birds, particularly when the provisioning contribution of subordinates is a major factor increasing the group’s total productivity. In such situations, a subordinate might alter the magnitude of its later provisioning contribution depending upon the amount of personal reproduction obtained earlier (at the time of fertilisation and egg laying). One effect of such temporal separation of decisions is that the magnitude of the benefit gained by a dominant would not be fixed at the time of yielding the reproductive concession to a subordinate (see also Magrath 1999). Again, the dunnock provides a real-life example. The effect of such sequential decisions on predicted levels of skew has yet to be modelled.

In addition to extending existing theory, there is also an urgent need for careful empirical tests of the predictions that differentiate between the major skew models. To date, few such studies have been undertaken. The difficulties, as well as the promise, of conducting such tests with cooperatively breeding birds have been discussed above.

Why is an understanding of the allocation of reproduction in social groups such an important question? As Reeve, Emlen & Keller recently wrote (1998, p. 276): 'the relative applicability of optimal skew versus incomplete control models has rather profound implications for our understanding of the evolution of animal societies. If the optimal skew models are correct, the possibility of a truly unified theory of social evolution is greatly enhanced because these models provide a fairly straight-forward theoretical apparatus for linking the ecology and genetic structure of societies to their reproductive partitioning and patterns of intra-group conflicts. If the incomplete control models are correct, this linkage is largely severed, and the models describing internal social dynamics will have to be decidedly more 'local' (i.e. tailored quite differently to different social systems).

'Even more fundamentally, the two different kinds of models entail different views about the degree of sophistication in interactions between members of an animal society. In the incomplete control model, interactants are engaged in a tug-of-war struggle over resources, but in the optimal skew models... organisms are engaged in higher-order transactions --- exchanges of parcels of reproduction designed to induce recipients to behave in a more favourable way. It is of fundamental importance to determine if the latter transactional view applies outside of humans, not only to non-human vertebrate societies, but also to invertebrate societies. If so, the implication will be that we have previously underestimated the power of selection to design social behaviours of considerable intricacy and subtlety in the absence of cognitive complexity.

I thank M. du Plessis for co-organising this symposium with me, and R. Heinsohn, I. G. Jamieson, J. Komdeur, and R. D. Mcgrath for participating. Peter Wrege kindly stood in for me when temporary medical complications prevented my attendance. My research on both skew theory and on cooperatively breeding birds has been supported by the National Science Foundation. Thanks to R. Heinsohn, I.G. Jamieson, and R.D. Magrath for sharing their unpublished manuscripts from this symposium, and N.J. Demong, M. duPlessis, R.D. Magrath, H.K. Reeve, and P.H. Wrege for comments on the manuscript.

Bengtsson, B.O. 1978. Avoiding inbreeding: at what cost? Journal of Theoretical Biology 73: 439-444.

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

Brown, J.L. & Brown, E.R. 1984. Parental facilitation: parent-offspring relations in communally breeding birds. Behavorial Ecology and Sociobiology 14: 203-209.

Cant, M.A. 1998. A model for the evolution of reproductive skew without reproductive suppression. Animal Behavior 55: 163-169.

Clutton-Brock, T.H. 1998. Reproductive skew, concessions and limited control. Trends in Ecology and Evolution 13: 288-292.

Cockburn, A. 1998. Evolution of helping behavior in cooperatively-breeding birds. Annual Review Ecology & Systematics 29: in press.

Emlen, S.T. 1982. The evolution of helping. II. The role of behavioral conflict. American Naturalist 119: 40-53.

Emlen, S.T. 1991. Evolution of cooperative breeding in birds and mammals. In: Krebs, J.R. & Davies, N.B. (eds) Behavioural ecology: an evolutionary approach. 3rd ed.: 301-337. Oxford; Blackwell Scientific.

Emlen, S.T. 1995. An evolutionary theory of the family. Proceedings of the National Academy of Science 92 (18): 8092-8099.

Emlen, S.T. 1996. Reproductive sharing in different types of kin associations. American Naturalist 148: 756-763.

Emlen, S.T. 1997. Predicting family dynamics in social vertebrates. In: Krebs, J.R. & Davies, N.B. (eds) Behavioural ecology: an evolutionary approach. 4th ed.: 228-253. Oxford; Blackwell Scientific.

Emlen, S.T. & Vehrencamp, S.L. 1983. Cooperative breeding strategies among birds. In: Brush, A.H. & Clark, J.G.A. (eds) Perspectives in ornithology: 93-133. Cambridge; Cambridge University Press.

Hamilton, H.D. 1964. The gentical evolution of social behaviour. I. and II. Journal of Theoretical Biology 7: 1-52.

Heinsohn, R. Dunn, P. & Legge, S. 1999. Extreme reproductive skew in cooperatively-breeding birds: tests of theory in white-winged choughs. In: Adams, N.J. & Slotow, R.H. (eds) Proc. 22 Int. Ornithol. Congr., Durban, University of Natal.

Jamieson, I.G. 1999. Reproductive skew models & inter-species variation. In: Adams, N.J. & Slotow, R.H. (eds) Proc. 22 Int. Ornithol. Congr., Durban, University of Natal.

Johnstone, R.A., Woodroffe, R., Cant, M.A. & Wright, J. 1999. Reproductive skew in multi-member groups. American Naturalist, in press.

Keller, L. & Reeve, H.K. 1994. Partitioning of reproduction in animal societies. Trends in Ecology and Evolution 9:98-102.

Koenig, W.D., Haydock, J. & Stanback, M.T. 1998. Reproductive roles in the cooperatively-breeding acorn woodpecker: incest avoidance versus reproductive competition. American Naturalist 151: 243-255.

Ligon, J.D. & Stacey, P.B. 1989. The significance of helping behavior in birds. AUK 106: 700-705.

Magrath, R.D. 1999. Variation in reproductive skew within populations of cooperatively-breeding birds. In: Adams, N.J. & Slotow, R.H. (eds) Proc. 22 Int. Ornithol. Congr., Durban, University of Natal.

Reeve, H.K. 1991. The social biology of polistes. In: Ross, K. & Matthews, R. (eds) The social biology of wasps: 99-148. Ithaca, New York; Cornell University Press.

Reeve, H.K. & Ratnieks, F.L.W. 1993. Queen-queen conflicts in polygynous societies: mutual tolerance and reproductive skew. In: Keller, L. (ed.) Queen number and sociality in insects: 45-85. Oxford; Oxford University Press.

Reeve, H.K. & Keller, L. 1996. Relatedness asymmetry and reproductive sharing in animal societies. American Naturalist 148: 764-769.

Reeve, H.K. & Keller, L. 1997. Reproductive bribing and policing as evolutionary mechanisms for the suppression of within-group selfishness. American Naturalist 150:S42-S58.

Reeve, H.K., Emlen, S.T. & Keller, L. 1998. Reproductive sharing in animal societies: reproductive incentives or incomplete control by dominant breeders? Behavioral Ecology 9: 267-278.

Rowley, I. 1965a. The life history of the superb blue wren, Malurus cyaneus. Emu 64: 251-297.

Rowley, I. 1965b. White-winged choughs. Australian Natural History 15: 81-85.

Sherman, P.W., Lacey, E.A., Reeve, H.K. & Keller, L. 1995. The eusociality continuum. Behavioral Ecology 6: 102-108.

Skutch, A.F. 1935. Helpers at the nest. AUK 76:471-477.

Stacey, P.B. & Koenig, W.D. 1990. Cooperative Breeding in Birds: Long-term Studies of Ecology and Behavior. Cambridge; Cambridge University Press.

Vehrencamp, S.L. 1979. The roles of individual kin and group selection in the evolution of sociality. In: Marler, P. & Vandenbergh, J. (eds) Handbook of behavioral neurobiology. Vol 3: 351-394. Plenum Press.

Vehrencamp, S.L. 1983a. A model for the evolution of despotic versus egalitarian societies. Animal Behavior 31: 667-682.

Vehrencamp, S.L. 1983b. Optimal degree of skew in cooperative societies. American Zoologist 23: 327-335.

Waser, P.M., Austad, S.N. & Kena, B. 1986. When should animals tolerate inbreeding? American Naturalist 128: 529-537.