Feedback loop between kinship and dominance: the macaque model

J. theor. Biol. (1990) 145, 511-521

Feedback Loop between Kinship and Dominance: the Macaque Model BERNARD THIERRY

Laboratoire de Psychophysiologie, C N R S ( U R A 1295) and Universitd Louis Pasteur, 7 rue de l' Universitd, 67000 Strasbourg, France (Received on 11 December 1989, Accepted in revised form on 9 March 1990) There is growing evidence that macaque social systems represent sets of coadapted traits in which strength of hierarchies and degree of nepotism covary. A framework is developed to explain the link between dominance and kinship phenomena, assuming that power brought by alliances among non-kin is allometrically related to those involving relatives. This can account for the type of social relationships observed in "despotic" systems vs. "egalitarian" ones. When social bonds are mostly founded on kinship, lineages are closed and social power generated by coalitions among relatives may reach high levels; social power frequently outweighs the fighting abilities of single individuals, and asymmetry of dominance between group members may be marked. When lineages are more open, social bonds and alliances are less kin-biased, social relationships are more equal, and as the influence of coalitions is less important, the individual retains a certain degree of freedom in relation to the power of kin-networks. Acknowledging that the balance between individual and social power is not set at the same level across different species can explain a number of variations in rules of rank inheritance and relative dominance of males and females among macaques. The framework illustrates how epigenetic processes may shape complex features of primate social systems, and offers opportunities for testing.

Introduction It is recognized that an integrated view o f evolution needs to account for the occurrence o f cpigenetic processes during development (Gould & Lewontin, 1979; Webster & Goodwin, 1982; Hughes & Lambert, 1984; Maynard Smith et al., 1985). However, in the study of social systems, explanations have until now been framed almost exclusively in functionalist terms, and it is most recent that a handful of authors have begun to investigate how social behaviour is constrained by structures (Thierry, 1985a, 1989; Jamieson & Craig, 1987; Irwin, 1988). In particular, a debate concerning the dualism between ultimate and proximate causes of behaviour has arisen (cf. Chadwick-Jones, 1987; Jamieson, 1989); many authors deny that some important behavioural features might not have direct adaptive function but that they represent the result of pleitropic effects, secondary consequences or structural constraints (Alcock, 1987; Sherman, 1988). They suggest that any attempt to explain a behaviour by proximate causes only is founded on a misunderstanding of the independence o f levels o f analysis. In this context, it is crucial to offer examples in which the analysis of proximate causes may account for phenomena which so far defeated ultimate explanations. The study of the social systems of macaques provides such a case. In what follows, the results of research showing that the various 5||

0022-5193/90/016511 + 11 $03.00/0

© 1990 Academic Press Limited

512

a. THIERRY

components of macaque social systems are coadapted is first summarized. Then, how dominance and kinship may be linked in a feedback loop is developed and the argument quantified in a simple model. Lastly, the model is applied in explaining observed variations in the patterns of rank inheritance among macaques.

Coadaptation between Traits SPECIES

DIFFERENCES

Agonistic behaviour patterns were first studied and compared in three species of macaque (rhesus macaque: Macaca mulatta; longtailed macaque: M. fascicularis; Tonkean macaque: M. tonkeana) (Thierry, 1985a, 1986a). This showed that the more intense the aggression in a species, the more unidirectional are the contests: high frequencies of biting are related to fleeing and submitting, while retaliation is more frequent when risk of injury is low. Correspondingly, the development of conciliatory patterns appears negatively correlated with levels of intensity and asymmetry of agonistic interactions. With regard to dominance and kinship relations, it emerged that asymmetry of relationships was related to the asymmetry of interactions: in Tonkean macaques, there was a high degree of tolerance towards demands by conspecifics, the gradient of dominance being weak. In other words, rank differences between individuals were small, which corresponded to a high frequency of conciliatory patterns, small interindividual distances, a considerable incidence of bidirectional agonistic interactions, and many affiliative interactions, regardless of the relative ranks and relatedness of the interactants (Thierry, 1984, 1986b, 1989; Thierry et al., in press). This picture contrasts with results from the other two species, especially the rhesus macaques, in which intolerance was marked, aggression highly unidirectional and often severe, individuals being separated by strong differences in rank and restricting most of their affiliative interactions to kin (Thierry, 1985a, 1989). Conditions of socialization also reflected the milieu in which young groupmembers develop. The degree of permissiveness of the mother and the development of alioparental care by other females were positively associated with symmetry of the interactions and relations between individuals, level of intensity in motheroffspring conflict at weaning being directly related to these features (Thierry, 1985b, 1989). In summary, all behavioural traits appeared coadapted. The three social systems studied might be considered as representing different states of equilibrium composed of homologous components, and varying along a continuum ranging from "despotism" to "egalitarism" (Thierry, 1986a; cf. Vehrencamp, 1983, for operational use of these terms). It may be of interest to note that the above conclusions about conditions of socialization are consistent with those of other authors studying other species (Rosenblum, 1971; McKenna, 1979; Mason et al., 1989). Also, confirmation of the relations observed among aggression, conciliation, dominance and kinship was recently provided by de Waal & LuttrelI (1989), who compared rhesus with stumptailed macaques ( M. arctoides ).

FEEDBACK

LOOP

BETWEEN

KINSHIP

AND

DOMINANCE

513

ROLE OF EPIGENESIS

While some differences between social systems were found in comparing different species of macaque, similar variations may be found within a single population. For instance, the comparison of agonistic behaviour patterns between kin and non-kin in Japanese macaques (M.fuscata) led to similar conclusions about relations between behavioural traits: bites were less frequent, interactions less often asymmetrical and conciliation more frequent between relatives (Thierry, 1990), reflecting their privileged relationships. It is known in rhesus and Japanese macaques that dominant mothers are more permissive than subordinate ones (White & Hinde, 1975; French, 1981), behaviour of females towards their infant being related to differences in the aggression received. Although coadaptation may have a genetic basis (Ricker & Hirsch, 1988), these instances show that this phenomenon may equally represent an emergent property of the system. The multiple forces in action through epigenetic processes generate structural constraints which promote certain pathways, exclude others, and more generally pattern social systems. The range of possible equilibria among traits is limited and the existence of allometric relations linking them should be expected. The Interplay between Kinship and Dominance ASSOCIATION

BETWEEN

NEPOTISM

AND

HIERARCHY

Most of the differences in the social systems of macaques may be reduced to terms of dominance and kinship relationships. At one end of the continuum, despotic species are characterized by strong hierarchies, with marked asymmetries in power between individuals, intolerance between unrelated individuals, and a tremendous importance of kinship relations, which determine how and with whom to interact, and have an overall influence upon most individual behaviours. Known examples of such species are rhesus (Sade, 1972; Missakian, 1972; Ehardt & Bernstein, 1987) and Japanese macaques (Kawai, 1958; Kawamura, 1958; Yamada, 1963; Koyama, 1967; Kurland, 1977). At the other end of the continuum, more egalitarian species display weak differences in dominance ranks between individuals and high levels of tolerance, and kinship relations do not represent barriers to affiliative interactions between unrelated individuals. Tonkean (Thierry, 1984, 1986a; Thierry et al., in press) and bonnet macaques (M. radiata) (Rosenblum, 1971; Sugiyama, 1971; Defter, 1978; Caine & Mitchell, 1979; Silk et al., 1981) are representative of this category, as are Barbary macaques (M. sylvanus) (Paul & Kuester, 1987; see also patterns of interactions among females in Deag, 1977, and among males in Deag, 1980; Taub, 1980). Hierarchies and nepotism rule the social life in the former two species, whereas the role of these phenomena is real but less important in the latter three species. Such correlated variation is not explained by current general theories about primate social evolution. These theories make reference to ecological (Wrangham, 1980; van Schaik, 1989) or genetic differences between species (Wade, 1979; Caldecott, 1986) which, in the macaques, remain unsupported by evidence. Rather than imagine a combination of ultimate factors having produced the observed association between

514

B. THIERRY

hierarchies and nepotism, it seems more reasonable to look for a possible connection between them. In this context, I have pointed out at the special status of kin-based alliances (Thierry, 1989): support of relatives by aggressive means is a common behaviour among macaque species (e.g. Kawai, 1958; Kurland, 1977; Massey, 1977; de Waal, 1977; Bemstein & Ehardt, 1985; Chapais, 1988), it might well represent the required link between dominance and kinship relations. The conditions under which this may operate are developed in the following section. KIN-BIAS

IN ALLIANCES

Dominance status in a group of macaques results from the combination of two kinds of power (cf. Kawai, 1958; Datta, 1983). Individualpower (I) depends on the fighting abilities of the individual, it is directly related to body size (Datta, 1983, 1988), and probably to behavioural features such as experience and personality (el. Smuts, 1985). Social power (S) is attained in contests by the support of allies. In what follows, it is assumed that the two kinds of power do not follow the same laws of variation, and that differences in social power depend on the form of the social system.

Assumptions Although an ally may bring all its own individual power in supporting a relative in a given contest, it is not always in proximity when help is needed and, therefore, the overall power it brings (Sk, social power given to kin) is to be considered to be only a fraction of its individual power:

Sk =aI,

(1)

where a is a constant inferior to 1. (Note that the degree of kin-bias in alliances representing the variable under study, the present demonstration focuses on social power brought by relatives; however, this does not detract from the existence of coalitions between non-relatives.) An ally's support depends on its relationship with the opponent. As there exist social bonds among unrelated individuals, the support given should be weaker when the bond between ally and opponent is stronger. This may be seen as a counter-power to nepotistic power. Social power may consequently be depicted as non-kin power (S.) subtracted from kin power (Sk):

s=s~-s..

(2)

Because individuals are immersed in a complex network of relationships, Sk and S, are dependent variables. Each participant in a conflict may call for help from several allies, relatives or not, and any intervention by a potential ally takes into account all the ally's social bonds with potential partners and opponents. The same individuals may be involved as allies or opponents in various combinations of alliances, or be neutralized through their bonds with several opposed parties. This produces a system involving multiple simultaneous relations between components, which may be described using the equations developed in systems theories (cf. von

FEEDBACK

LOOP

BETWEEN

KINSHIP

AND

DOMINANCE

515

Bertalanffy, 1968). Given that alliance networks are stable (a reasonable assumption for periods measured in weeks or months), the relation between kin and non-kin social powers may be approximated by the following allometrical equation: S, = bSL (3) where b and a are constants, a represents the proportion of non-kin vs. kin support within the social system considered, it may take values comprised between 0 and 1. From eqns (2) and (3), it follows that social power S = SR -- bS~, (4) and, from eqns (1) and (4),

S=I

I 1-a

ab

which reduces to S = ~ S I l-~

(with/3=l/ab).

(5)

Graphic m o d e l

Assuming that individual power ranges on a scale of 10°, it is possible to draw the curves corresponding to formula 5 (Fig. 1). Coefficient a varies according to the social system considered. /3 influences the increment of social power. Examination of the graphic model leads to the following statements: When a is elevated (i.e. set at ½ or ~), social power remains at low values (Fig. 1). This means that a superiority in individual power is of great advantage in contests. Social power makes a difference only when there is little discrepancy in individual power between antagonists. It is generally impossible to outrank a definitely stronger opponent, unless one is able to enlist a much larger number of allies than the opponent. On the whole, differences in social power between group members remain weak, and relationships are consequently quite egalitarian. When social bonds between unrelated individuals are weak, a is low (e.g. equal to ~ and ~0) and social power may reach very high levels (Fig. 1). A superiority in individual power is advantageous only when relative social power is weak. Results of contests are mainly ruled by coalitions; even intrinsically strong individuals may be outranked by challengers of weak individual power if the latter are supported by influential allies. As discrepancies in social power between individuals may be very large, dominance asymmetries are often strong, resulting in the overall social system being of the despotic kind. Implications

The model illustrates the fact that relative weights of individual and social power vary according to the society considered. It implies that degrees of kin-bias in alliances and asymmetry of power are linked in a positive feedback loop. When relatives form cohesive and closed subgroups, their relations with non-kin are correspondingly quite tenuous, and there is a marked imbalance between support given to members of one's lineage and that accorded to other individuals. Each individual in a high-ranking subgroup benefits from the ability of the subgroup to dominate others, which should strengthen bonds between relatives and further favour their mutual preferential treatment. Heightened nepotism in turn increases

B. THIERRY

516

I0

15

8

h

a =

II~ / a=l/3

6 10 4

2 0

I

I

23

I

I

45

I

I

67

I

I

8910

25 0 ~ o o.

I

I

I

2 3 4

I

I

5

6 7

I

8 910

I

I

I

I

20

20

o

p, O3

15

15

I0

I0

0

I

I

2 3

I

I

4 5

I

I

6 7

I

I

a =1/5

I

I

I

8 910 2 :5 4 Individual power

I

5 6

I

7 8

I

9 I0

FIG. 1. Curves of equality between individual and social power. Social power was calculated using the formula S = f l i t - ' . ot represents the proportion of kin versus non-kin support. For every given value of a, the curves were drawn for three possible values of the coefficient/3 (1, 2 or 3).

separation of lineages and stratification of the society (cf. de Waal, 1986), which further augments rank distances between unrelated individuals. In egalitarian systems, the existence of such a strong feedback loop should not be expected. Individuals are not separated by strong dominance asymmetries; the occurrence of frequent affiliative interactions between non-relatives favours the formation of social bonds between them and lineages remain open. This makes room for alliances not founded on kinship, which inhibits any increase of dominance asymmetry between unrelated individuals. To summarize, in despotic systems, individual dominance status is mainly dependent on the power of the kin subgroup, and any challenge by unrelated inferiors encounters a strong coalition. In more egalitarian systems, power is more balanced among group members and the individual retains a certain degree of freedom with

FEEDBACK

LOOP

BETWEEN

KINSHIP

AND

DOMINANCE

517

regard to power networks. It may be noted that, in groups composed of unrelated individuals, privileged relationships emerge among some individuals and, therefore, the existence of a similar relation between degree of dominance asymmetry and degree of exclusivity in social bonds should also be expected in such artificial groups. Understanding Variations in the Transmission of Dominance FEMALES

That kinship and dominance phenomena may be interconnected at an epigenetic level, like other features of macaque social systems, has important consequences for the understanding of power inheritance among female macaques. It is a well acknowledged fact (Kawamura, 1958; Koyama, 1967; Sade, 1967, 1972; Missakian, 1972; Chapais, 1988; Datta, 1988) in rhesus and Japanese macaques that (1) females inherit their mother's rank, (2) mothers dominate daughters, (3) younger sisters dominate older ones. That is, females rank just below their mother; within families, sisters' ranks are strictly ordered inversely with age, and, as relatives are close in rank, families themselves are ordered along hierarchies. Notwithstanding a few exceptions (Chikazawa et al., 1979; Chapais, 1985; Datta, 1988; Mori et al., 1989), the above rules were found to apply with considerable regularity, so it is possible to predict quite accurately the dominance rank of a female at adulthood from her birth order and family rank. These hierarchical patterns have been shown to result from maternal support to the youngest daughter, and existing alliances of the mother (Berman, 1980; Chapais, 1988; Datta, 1988). These were once believed to be universal among macaques, and two different interpretations were advanced to account for them. One suggests that the mother favours her youngest daughter because she has the highest reproductive value (Chapais & Schulman, 1980). The other proposes that she does so in order to prevent an eventual coalition among her adult daughters, which could outrank her (Horrocks & Hunte, 1983). Both kinds of explanation aim to demonstrate the adaptiveness of the rules observed, and rely exclusively on ultimate causes. As they stand to be very general, they fail to explain cases where the rules do not apply. In Barbary (Paul & Kuester, 1987), bonnet (Silk et al., 1981) and Tonkean macaques (Thierry, unpublished data), old matriarchs are frequently outranked by their adult daughters. In addition, while there is a tendency for sisters to rank in inverse order of age, the rule is too often violated to hold: a number of maturing younger sisters do not outrank their older sisters (Silk et al., 1981; Paul & Kuester, 1987). Such patterns may be understood by considering them within the scope of the framework previously developed. In the more egalitarian systems of these species, social power is relatively weak compared to individual power. Therefore, a female of strong individual power may outrank a younger sister in spite of the mother's support. Also, any decline in fighting abilities due to the mother's age enables successful contest by daughters: the support of the mother's allies is not sufficient for her to maintain her rank. Such effects cannot easily occur in despotic species: any inferiority in individual power is generally outweighed by the social power brought by related allies. It is noteworthy that the few exceptions in which daughters

518

B. THIERRY

outrank mothers in rhesus and Japanese macaques involved cases where alliance with a non-kin occurred (cf. Chapais, 1985). Rules of rank inheritance should not be considered as the direct result of natural selection but as emergent properties of the social system and, as such, dependent on it. MALES

As male macaques may transfer from one group to another several times in their life, the influence of lineage on their social status is not as important as for females. However, in rhesus and Japanese macaques, the dominance rank of maturing males is highly dependent on the mother's hierarchical position. While most of them migrate between 3 and 5 years of age, those remaining in their natal group usually outrank their mother and sisters at around 5-7 years of age (Koyama, 1967; Missakian, 1972; Sade, 1967, 1972; Datta, 1983; Colvin, 1983). However, a number of adult males are usually subordinate to certain adult females (Yamada, 1963; Packer & Pusey, 1979; Johnson et al., 1982; Chapais, 1983). In Barbary macaques (Kuester & Paul, 1988), in contrast, the rank of maturing males depends on their age (i.e. mainly on their physical strength), the correlation between mother and offspring rank not being perfect, in contrast to rhesus and Japanese macaques. Young males dominate all females of their age and, by 5-6 years, they outrank all females, including their mother, adult sisters, and unrelated high-ranking females (Deag, 1977; Kuester & Paul, 1988). Migration occurs late, after 5 years of age (cf. Colvin, 1983; Mehlman, 1986). Although male relationships are less documented in Tonkean and bonnet macaques, it appears that subadult males of these species remain fully integrated in groups, as in Barbary macaques; in particular, males dominate females by 5-6 years in Tonkean macaques (Thierry, unpublished data), and migration is also delayed in bonnets (Simonds, 1973; Colvin, 1983). The above interspecific variations may be explained by considering the balance between individual and social powers. In despotic species, coalitions between relatives determine the rank of maturing males. In addition, some adult males may be unable to outrank females of strong social power. No such effects should be expected in more egalitarian species: fighting abilities of males may outweigh differences in social power, explaining the weaker effect of lineage and the fact that they rapidly outrank all females. Their high social status, together with the affiliation with other males may account for the late age at migration (Simonds, 1965, 1973; Taub, 1980; Bartecki, 1986). A further difference may be explained by the model. In rhesus and Japanese macaques, contests and reversals mainly occur between maturing males of different age. In Barbary macaques, in contrast, most rank reversals occur among members of the same age-cohorts, and there are few overlaps in hierarchy between these cohorts (Paul & Kuester, 1988). From the model, conflicts are ruled by alliances in despotic species and in effect, at a given moment, high-ranking mothers' sons are in the process of outranking older males belonging to lower lineages (Sade, 1967; Koyama, 1967; Datta, 1983). In Barbary macaques, strong differences in size cannot

FEEDBACK LOOP BETWEEN KINSHIP AND DOMINANCE

519

be easily countered by enlisting allies and, therefore, most challenges occur among same-aged peers. Conclusion

From the simple assumption that kin-bias in the formation of social bonds and alliances varies among species, it is possible to explain a number of variations in the form of macaque social systems. When there is a marked imbalance between individual and social power, strong asymmetries in dominance relationships are made possible; it is difficult for an individual to counter the effect of kin-based coalitions, and an offspring's rank is mainly decided by its place within the kinship network. When networks of relatives have less weight in the social system, relationships remain more symmetrical and there is more room for individual assertiveness. This framework allows predictions to be made. For instance, it was shown in Japanese macaques that removing one individual from a dominant coalition of relatives may lead to an overthrow by a subordinate coalition gathering more partners (Chapais, 1988). In the same experimental conditions, no such marked effect should be expected in more egalitarian species; while the number of related allies is of primary importance in despotic systems, this is less so in egalitarian systems. Also, it may be predicted that spectacular rises in rank by individuals belonging to subordinate lineages should be more frequent in weakly despotic species since help from high-ranking individuals may be available (cf. Small, 1989, a report of such an event in Barbary macaques). More generally, as there exist 16 species of macaque, of which only a few are well-known, increased study of the genus offers the opportunity to verify whether strength of hierarchies, degree of nepotism, modes of dominance transmission, and relative rank of males and females are indeed covarying patterns. The terms "despotism" and "egalitarism" do not represent a black-and-white dichotomy but, rather, the so-called ends of a continuum along which a number of species should be found to occupy intermediate positions. Available data tentatively suggest that longtailed (Angst, 1975; de Waal, 1977; Thierry, 1986a) and pigtailed macaques (M. nemestrina) (Bernstein, 1969; Rosenblum, 1971; Defter, 1978) are located towards the despotic side, while stumptailed macaques (Bertrand, 1969; Nieuwenhuijsen et al., 1988; de Waal & Luttrell, 1989) are closer to the egalitarian side. The framework presented here should be viewed as open to further refinement. As social systems develop under socio-ecological constraints, integrating the influence of demographic variations (cf. Hausfater et aL, 1987; Datta, 1989) should increase the explanatory and predictive power of the model, and allow it to be extended to primates other than macaques. Life-history variables such as birth rate or life-expectancy, in particular, may affect size of lineages and, consequently, the balance between kin and non-kin power. Further, the form of any society containing coalitions is liable to be influenced by their degree of kin-bias. The existence of a link between reciprocity and relaxed dominance, observed particularly in humans (Trivers, 1971), could be related to this phenomenon. However, caution is required in generalizing the model since additional interactions are to be expected

520

n. T H I E R R V

as a function of differences in quantity and quality of the components in the systems being compared. Finally, it should be emphasized that some of the more complex features of macaque social systems are easily explicable by structural considerations, without resorting to any ultimate causes. That dominance and kinship, usually interpreted in sociobiology and behavioural ecology exclusively in terms of adaptation to the environment, are the phenomena concerned, is revealing of the overwhelming albeit underestimated importance of epigenetic processes in the development of social behaviour. While the present model stresses the importance of interconnections among traits, it does not indicate which one would be possibly most sensitive to environmental selective pressures and prone to be pacemaker of the evolution of the whole system. As an example, conciliatory patterns displayed by bonnet, Barbary and Tonkean macaques appear much more frequent and sophisticated than those observed in rhesus and Japanese macaques (cf. Simonds, 1965; Sugiyama, 1971; Silk & Samuels, 1984; Deag, 1977, 1980; Thierry, 1984, 1986a). It remains to assess whether such behaviour patterns would merely represent secondary features of social systems in which degree of nepotism or power asymmetry are ecologically selected adaptations (cf. Vehrencamp, 1983; van Schaik, 1989) or, on the contrary, represent key innovations leading to improved tradeoff abilities among individuals and the consequent emergence of weakly despotic systems. The

author is grateful to Ph. Ropartz and J. Anderson for their comments on the manuscript. REFERENCES

ALCOCK, J. (1987). Nat. Hist. 96, 4. ANGST, W. (1975). In: Primate Behavior (Rosenblum, L. A., ed.) pp. 325-388. New York: Academic Press. BARTECKI, U. (1986). Primates 27, 173-184. BERMAN, C. M. (1980). Int. J. Primatol. 1, 153-170. BERNSTEIN, I. S. (1969). Anita. Behav. 17, 452-458. BERTRAND, M. (1969). The Behavioral Repertoire of the Stumptail Macaque. Basel: Karger. BERNSTEIN, I. S. & EHARDT, C. L. (1985). Am. J. Primatol. 8, 37-52. VON BERTALANFFY, L. (1968). General Systems Theory. New York: Bra2iller. CA1NE, N. G. & MITCHELL, G. (1979). Primates 20, 583-590. CALDECOTT,J. O. (1986). Anita. Behav. 34, 208-220. CHADWICK-JONES, J. K. (1987). Ethology 74, 164-169. CHAPAIS,B. (1983). In: Primate Social Relationships (Hinde, R. A., ed.) pp. 171-175. Oxford: Blackwell. CHAPAIS, B. (1985). Primates 26, 407-423. CHAPAIS,B. (1988). Anita. Behav. 36, 1025-1037. CHAPAIS, B. Rr SCHULMAN, S. R. (1980). J. theor. Biol. 82, 47-89. CHIKAZAWA, O., GORDON, Z., BEAN, T. & BERNSTEIN, I. S. (1979). Primates 20, 301-305. COLViN, J. (1983). In: Primate Social Relationships (Hinde, R. A., ed.) pp. 160-171. Oxford: Blackwell. DATrA, S. B. (1983). In: Primate Social Relationships (Hinde, R. A., ed.) pp. 93-102. Oxford: Blackwell. DA3-rA, S. B. (1988). Anita. Behav. 36, 754-772. DA'I-rA, S. B. (1989). In: Comparative Socioeeology (Standen, V. & Foley, R. A., eds) pp. 265-284. Oxford: Blackwell. DEAG, J. M. (1977). Anita. Behav. 25, 465-474. DEAG, J. M. (1980). Behaviour 75, 54-81. DEFLER, T. R. (1978). Primates 19, 153-167. EHARDT, C. ,t, BERNSTEIN, I. S. (1987). Am. J. Primatol. 13, 255-270. FRENCH, J. A. (1981). Int. J. Primatol. 2, 237-246. GOULD, S. J. & LEWONT1N, R. (1979). Proe. R. Soc. Lond., B Biol. Sci. 205, 581-598.

FEEDBACK

LOOP BETWEEN KINSHIP

AND DOMINANCE

521

HAUSFATER, G., CAIRNS, S. J. & LEVIN, R. N. (1987). Am. J. Primatol. 12, 55-70. HORROCKS, J. A. & HUNTE, W. (1983). Am. Nat. 122, 417-421. HUGHES, A. J. & LAMBERT, D. M. (1984). J. theor. Biol. 111, 787-800. IRWIN, R. E. (1988). Anint Behav. 36, 814-824. JAMIESON, I. G. (1989). Anint Behav. 37, 696-697. JAMIESON, I. G. & CRAIG, J. L. (1987). Perspect. Ethol. 7, 79-98. JOHNSON, O. F., MODAHL, K. B. & EATON, G. (1982). Anita. Behav. 30, 383-392. KAWAI, M. (1958). Primates 1, 111-148. [English translation In: Japanese Monkeys (Altmann, S. A., ed.) pp. 66-86. Atlanta: published by the editor]. KAWAMURA,S. (1958). Primates 1, 149-156. [English translation In: Japanese Monkeys (Altmann, S. A., ed.) pp. 105-112. Atlanta: published by the editor]. KOYAMA, N. (1967). Primates 8, 189-216. KURLAND, J. A. (1977). Kin Selection in the Japanese Monkey. Basel: Karger. KUESTER, J. & PAUL, A. (1988). Folia Primatol. 5, 33-44. MASON, W. A., LONG, D. D. & MENDDZA, S. P. (1989). Am. J. Primatol. 18, 155. MASSEY, A. (1977). Behav. Ecol. Sociobiol. 2, 31-40. MAYNARD SMITH, J., BURIAN, R., KAUFFMAN, S., ALBERCH, P., CAMPBELL, J., GOODWlN, B., LANDE, R., RAUP, D. & WOLPERT,L. (1985). Q. Rev. Biol. 60, 265-287. MCKENNA, J. J. (1979). Am. Anthropol. 81, 818-840. MEHLMAN, P. (1986). Am. J. Primatol. 10, 67-81. MISSAKIAN, E. A. (1972). Primates 13, 169-180. NIEUWENHUIJSEN, K., SLOB, A. K. & VAN DER WERFF TEN BOSCH, J. J. (1988). Psychobiology 16, 357-371. MOR1, A., WATANABE, K. '¢" YAMAGUCHI, N. (1989). Primates 30, 147-173. PACKER, C. & PUSEY, A. E. (1979). Folia Primatol. 31, 212-218. PAUL, A. & KUESTER, J. (1987). Behav. Ecol. Sociobiol. 21, 323-331. RICKER, J. P. & HIRSCH, J. (1988). 3". cutup. Psychol. 102, 203-214. ROSENBLUM, L. A. (1971). In: Proceedings of the 3rd International Congress of Primatology, Ziirich 1970, Vol. 3 (Kummer, H., ed.), pp. 79-84. Basel: Karger. SADE, D. S. (1967). In: Social Communication among Primates (Altmann, S. A., ed.) pp. 99-114. Chicago: The University of Chicago Press. SADE, D. S. (1972). In: The Functional and Evolutionary Biology of Primates (R. H., ed.) pp. 378-398. Chicago: AIdine. VAN SCHAIK, C. P. (1989). In: Comparative Socioecology (Standen, V. & Foley, R. A., eds) pp. 195-218. Oxford: Blackwell. SHERMAN, P. W. (1988). Anita. Behav. 36, 616-619. SILK, J. B. & SAMUELS,A. (1984). Am. J. Primatol. 6, 373-376. SILK, J. B., SAMUELS, A. & RODMAN, P. S. (1981). Primates 22, 84-95, S1MONDS, P. E. (1965). In: Primate Behavior: FieM Studies of Monkeys and Apes (DeVote, I., ed.) pp. 175-196. New York: Holt, Rinehart & Winston. SIMONOS, P. E. (1973). Am. J. phys. Anthropol. 38, 599-604. SMALL,M. F. (1989). Nat. Hist. !, 10-12. SMUTS, B. B. (1985). Sex and Friendship in Baboons. New York: Aldine. SUGIYAMA,Y. (1971). Primates 12, 247-266. TAUB, D. M. (1980). In: The Macaques (Lindburg, D. G., ed.) pp. 287-344. New York: van Nostrand Rheinhold. THIERRY, B. (1984). Behaviour 89, 1-28. THIERRY, B. (1985a). Aggress. Behav. 11, 223-233. THIERRY, B. (1985b). Behav. Proc. 11, 89-95. THIERRY, B. (1986a). In: Primate Ontogeny, Cognition and Social Behaviour (Else, J. G. & Lee, P. C., eds) pp. 307-313. Cambridge: Cambridge University Press. THIERRY, B. (1986b). Am. J. Primatol. 11, 89-97. THIERRY, B. (1989). 21st International Ethological Conference, Utrecht August 9-17. TmERRY, B. (1990). C. R. Acad. Sci. Paris 310, III, 35-40. THIERRY, B., GAUTHIER, C. & PEIGNOT, P. (1990). Int. J. Primatol. (in press). TRIVERS, R. L. (1971). Q. Rev. Biol. 46, 35-57. DE WAAL, F. B. M. (1977). Z. Tierpsychol. 44, 225-282. DE WAAL, F. B. M. (1986). Anint Behav. 34, 1033-1040. DE WAAL, F. B. M. & LUTTRELL, L. M. (1989). Ant J. Primatol. 19, 83-109.

522

B. T H I E R R Y

VEHRENCAMP, S. (1983). Anita. Behav. 31, 667-682. WAD~, T. D. (1979). Primates 20, 355-370. WEBSTER, G. & GOODWIN, B. C. (1982). J. Social Biol. Struct. 5, 15-47. WHITE, L. E. & HINDE, R. A. (1975). A n i n t Behav. 23, 527-542. WRANGHAM, R. W. (1980). Behaviour 75, 262-300. YAMADA, M. (1963). Primates 4, 43-65.