CAREL P. VAN SCHAIK, MARIA A. VAN NOORDWIJK AND CHARLES L. NUNN
Infanticide by males unlikely or unable to have fathered a female's current dependent offspring is an adaptive male reproductive strategy if the mother can soon be fertilised again and the infanticidal male is in a position to be the likely sire of her next infant (Hrdy, 1979; Hrdy, Janson and van Schaik, 1995). It is remarkably common among primates (Hausfater and Hrdy, 1984; Struhsaker and Leland, 1987; Hiraiwa-Hasegawa, 1988). Its importance as a source of infant mortality probably varies widely, but it is estimated to be responsible for 31-64% of all infant mortality in some well-studied species (hanuman langurs: Sommer, 1994; Borries, 1997; mountain gorillas: Watts, 1989; red howlers: Crockett and Sekulic, 1984). Several detailed studies have demonstrated that it is an adaptive behaviour for males (op. cit.; Hrdy, 1979).
Male infanticide is potentially common in primates because many species have prolonged lactational amenorrhoea. In species without lacta-tional amenorrhoea, early resumption of mating activity ('postpartum oestrus') means that killing infants will not advance the female's next birth. The incidences of postpartum mating and lactational amenorrhoea, in turn, are determined by the relative length of gestation and lactation. Where lactation is longer than gestation, we find lactational amenorrhoea, in both primates and other mammals (van Schaik, in press).
Among primates, there is a strong relationship between the mode of infant care, the incidence of postpartum mating and the incidence of male infanticide. Primates have two radically different modes of infant care: absentee care, in which the offspring are left in a nest or parked somewhere (galagos and lorises, several lemurs, some tarsiers), and permanent care, in which the offspring are carried around by the mother (all monkeys and apes, one tarsier and several lemurs: see van Schaik and Kappeler, 1997). Among species in which females carry their infants, there is an important difference between those in which mothers basically rear their offspring
alone and those in which there is communal rearing, i.e. the mother consistently receives help in rearing the offspring, especially by carrying and sometimes by provisioning. In these communal breeders, postnatal development can be speeded up due to the input by others. Among primates, they are represented by callitrichids and the pair-living cebids (cf. Lee, 1996).
Figure 8.1 shows that species with these three types of care vary in the relative length of lactation: the carrying primates with mother-only rearing have relatively very long lactation periods because their infants develop slowly (cf. Charnov, 1993). As predicted, none of these species shows postpartum mating and conception, whereas they are very common among species with absentee infant care and among species with communal infant rearing (82% of 11 species and 63% of 8, respectively; data on postpartum mating from Appendix 8.1). The frequent incidence of postpartum mating among the communally rearing infant carriers is clearly a derived condition.
The pattern in postpartum mating implies that infanticide is only expected to be common among species that carry their young but do not rear them communally, i.e. some lemuroids, some ceboids, and all cer-copithecoids and hominoids. The data in Figure 8.1 are consistent with this expectation. Male infanticide is only reported for the non-communal infant carriers, for which it is reported in 49% of 61 well-studied species (Appendix 8.1). Although under-reporting may explain its absence in the parkers (0% of 14), which are nocturnal and solitary, it cannot account for the absence among the communal infant carriers (0% of 8).
In non-communally breeding infant-carrying primates, the risk of male infanticide should have selected for female counterstrategies. Female primates in these taxa have evolved three major ways of reducing the risk of infanticide: (i) to gain protection for the infant by associating with the likely sire; (ii) to cut losses through abortion or premature weaning; and (iii) to prevent infanticidal attacks by employing sexual strategies.
There is strong evidence that likely fathers play a particularly important role in protecting primate infants (reviewed in van Schaik, 1996; in press). Male infanticide is therefore expected when this protection is removed. In nature, infanticide is almost always associated with take-overs by outsiders or, less commonly, with dominance upheavals inside groups (Hrdy et al., 1995; Borries, 1997; Steenbeek, 1996). In captivity, infanticide can be provoked reliably by experimentally replacing the adult (or dominant) male (Angst and Thommen, 1977; Kyes et al., 1995). This pattern suggests that females have evolved association with males as a counterstrategy against male infanticide. As expected, permanent male-female association in primates is absent among species with absentee infant care, whereas it is virtually universal among the infant-carrying species (van Schaik and Kappeler, 1997). It is retained among the communally rearing ceboids. However, while effective, it is not a perfect strategy: males can die or be ousted.
Where protection fails, females may respond to acute infanticide risk by terminating their investment in vulnerable offspring. Thus, a pregnant female could resorb or abort a fetus. This has been reported for patas (Rowell and Hartwell, 1978), hanuman langurs (Sommer, 1994), gelada baboons (Mori and Dunbar, 1985), captive hamadryas baboons (Colmenares and Gomendio, 1988), and for a yellow baboon group with an unusually aggressive immigrant male (Pereira, 1983).1 When the female already has an infant, and infanticide risk suddenly arises, she could wean her infant sooner than she would otherwise have done, even if it would reduce the infant's survival, as noted for wild gelada (Dunbar, 1980), hamadryas baboons (Sigg et al., 1982), and captive vervets (Fairbanks and McGuire, 1987). Resorption, abortion and infant abandonment are probably rare among primates.
A less costly alternative counterstrategy would be to prevent infanticide by manipulating the male's assessment of paternity or his actual chances of it through sexual behaviour.2 This chapter explores the possibility that the variation in primate sexual behaviour and physiology is an adaptation to reduce the risk of male infanticide (see also Hrdy, 1981; Hrdy and Whitten, 1987). The hypothesis examined in this chapter is that female sexuality in species vulnerable to male infanticide has been molded by the dual need for paternity concentration and confusion: concentration in order to elicit infant protection from the likely father, confusion in order to prevent infanticide from non-likely fathers (see also Nunn, in press). The right balance of confusion and concentration is achieved by (i) prolonged mating periods, often accompanied by actively pursued polyandrous mating, (ii) mating during non-fertile periods, such as pregnancy, (iii) making ovulation unpredictable, thus blurring the correlation with attractivity, and (iv) attracting males of all ranks by producing gradually changing, exaggerated sexual swellings.
Sexual counterstrategies against male infanticide
Whenever the threat of male infanticide occurs, a female would benefit from having a protector male, most likely the infant's sire. This protection is best obtained by concentrating paternity into a single male. Such concentration is normally achieved by default because, regardless of morphological advertising, dominant males mate preferentially with females around the time of likely ovulation and often guard them so as to monopolise mating (Kaufman, 1965; Hausfater, 1975; Glander, 1980; Harcourt et al, 1980; van Noordwijk, 1985).
On the other hand, female primates can also reduce the risk of infanticide by mating with at least some of the other males present before birth and thus confusing paternity. Paternity confusion is a feasible option because mammalian males cannot recognise infants as kin (Elwood and Kennedy, 1994). Thus, male decisions about whether to defend an infant, to ignore it, or, alternatively, to attempt to kill it, should be based on assessments of the likelihood of paternity, weighted against likely paternity of the future infant. These assessments are necessarily based entirely on mating history. If a male has exclusive mating access to a female and if matings only take place during the cycle in which fertilisation occurs, the male rule can be simple: if he mated with the female, his paternity is ensured. Field studies on primates with unimale groups suggest that when a newly immigrated male mates with a pregnant female, the probability that he will subsequently commit infanticide is reduced (e.g. blue monkey: Fairgrieve, 1995; hanu-man langur: Sommer, 1994), although it is not clear how long before birth this mating must take place. Experimental work on rodents confirms these rules (Perrigo and vom Saal, 1994). Where multiple males mate with the female before she gives birth, more complex rules are needed; these have not been studied in detail.
Females have two fundamentally different tactics at their disposal to confuse paternity. First, they mate polyandrously during regular ovarian cycles. This tactic should reduce the risk of infanticide by males that were already in the group but did not mate (enough) with the female and that rose dramatically in rank after conception (only high-ranking males would be likely to benefit from committing infanticide). Second, they pursue matings with one or more males when these males appear when fertilisation is impossible (e.g. during pregnancy).3 This tactic should reduce the risk from males who immigrate into the group after the female's conception cycle and become dominant, thus gaining good prospects of siring the female's next infant. In addition, post-conception matings could provide additional paternity confusion for males already present during the conception cycle.
Paternity is most easily concentrated when a female mates briefly, and exclusively or predominantly with a single male, whereas paternity confusion requires mating with multiple males, and therefore in most cases extended periods of mating. A review of primate mating behaviour (van Noordwijk, unpublished) shows that males can use a variety of anatomical, chemical and behavioural cues to assess the female's oestrogen activity, and thus likelihood of ovulation. However, the feasibility of deceptive receptivity during pregnancy underscores that ovulation itself is truly concealed to males at all times. The cues used by males can also be produced during pregnancy (when ovulation is effectively blocked), although females will be less attractive because of the inevitable endocrine differences (for which they may compensate by being more proceptive, e.g. patas: Loy, 1981). Indeed, males mating with females during pregnancy tend to be lower-ranking or subadult males (e.g. gorilla: Watts, 1991; sooty mangabey: Gust, 1994). We conclude that the mechanisms of male paternity assessment are so crude that there is considerable room for female manipulation of it.
Natural selection may have achieved the balance between paternity confusion and concentration by making ovulation unpredictable. Martin (1992) and Nunn (in press) review quantitative studies of several species and note that ovulation can take place over a range of at least 6 to 13 days. Thus, in species with swellings, ovulation is most likely during peak swelling, but not exclusively so (e.g. Wildt et al., 1977; Whitten and Russell, 1996). In species without morphological advertising, this unpredictability is harder to demonstrate. However, in humans, ovulation can take place during a surprisingly broad time window after last menstruation (Martin, 1992). This implies that the length of the follicular phase is highly variable. Indeed, across studies the standard deviation in the length of the follicular phase is consistently far greater than that of the luteal phase (see Hayssen, van Tienhoven and van Tienhoven, 1993). Thus, although male primates in polyandrous mating situations have endocrine and behavioural indications that ovulation is likely, its exact timing, and thus the best timing for fertile matings, is hard to predict.
Unfortunately, there are no data to test the assumption that the degree of unpredictability depends on the need for confusion. However, the concept of unpredictable ovulation also suggests that there should be a good but not perfect correlation between male rank and paternity, with a small proportion of infants being sired by low-ranking males. A recent review of paternity studies in primates supports this contention (Paul, 1998). We assume that the female can manipulate male paternity assessments and make them non-zero for all males who have mated with her before she gave birth. The males cannot improve much on their rules; they can use the correlation between the strength of the female's signal of likelihood of ovulation and the temporal proximity of mating to birth to obtain approximate clues, but the probabilistic nature of ovulation makes it impossible for natural selection to design a male decision rule that does much better.
We can now develop predictions for species in which infanticide risk is acute and paternity confusion is required. We can use the variation within primates in vulnerability to male infanticide (see Figure 8.1) to explore broad differences in sexual behaviour between the infant parkers and non-communal infant carriers (see below). Specifically, we expect that non-communal carriers should differ from parkers in the following ways. First, females should actively pursue promiscuity. Second, females should show situation-dependent receptivity in situations when new males appear that are potentially infanticidal and when not enough males mated during
Table 8.1. The range of social conditions in which females need to concentrate paternity by mating exclusively or predominantly with a single dominant male around the most likely time of conception, confuse it by mating with all potentially infanticidal males over a minimum frequency and confuse it by mating (deceptively) during times of non-fertility
Case Number of potentially infanticidal males:
Female sexual response:
in group or at periphery
Concentrate Confuse paternity paternity entering post-conception in dominant within situation regular dependent cycles
Was this article helpful?
If Pregnancy Is Something That Frightens You, It's Time To Convert Your Fear Into Joy. Ready To Give Birth To A Child? Is The New Status Hitting Your State Of Mind? Are You Still Scared To Undergo All The Pain That Your Best Friend Underwent Just A Few Days Back? Not Convinced With The Answers Given By The Experts?