Species of Mammals from Which Both Adoption

—Considerable attention has been paid to the phenomenon of infanticide in recent years. Five functional categories of infanticide have been defined. Here I concentrate on those that either have been described as the outcome of possible competition for limited resources or have by default been classified as the result of social pathology. Many of the species that show infanticide of this nature also show adoption of unrelated young at fairly high frequencies. I suggest that the possibility of caring for nonfilial offspring creates an intergenerational conflict, or arms race, whereby offspring separated from their parents or receiving parental care of substandard quality (insufficient for their survival) should be selected to solicit care from adults other than their parents and the potential adoptive parents are selected to avoid giving such care. Evidence suggests that most examples of supposedly pathological infanticide, or resource-based infanticide, are the result of potential foster parents killing unrelated offspring when these offspring can clearly be identified as nonrelatives. Support for this idea comes from observations that (1) such infanticide is most common in group-living or colonial species, where chances of encountering wandering offspring are high; (2) infanticidal individuals come almost exclusively from the sex that bears the primary costs of adoption; (3) such infanticide occurs only under conditions where victims can clearly be identified as nonfilial; and (4) in species with little or no cost to adoption, adoption is common, but infanticide is nonexistent. In recent years, a great deal of attention has been paid to the phenomenon of infanticide (reviews in Hausfater and Hrdy 1984). Five major functional categories of infanticide have been defined: (1) exploitation of the infant as a food resource, that is, cannibalism; (2) sexual selection: individuals (typically males) improve their opportunities for breeding by eliminating dependent offspring of a prospective mate; (3) parental manipulation: parents increase their own reproductive success (on average) by eliminating particular offspring; (4) competition for resources: death of the infant potentially increases resource availability either for the killer or its descendants; and (5) social pathology: killing of unrelated offspring with no adaptive explanation (from Hrdy 1979; Hrdy and Hausfater 1984). Numerous cases of infanticide involve breeding adults that attack and kill offspring of other breeding adults (e.g., Davis and Dunn 1976; Pierotti 1980, 1982a, 1988; Hausfater and Hrdy 1984 and reviews therein; Hoogland 1985; Pierotti and Murphy 1987). In many of these cases, offspring are not eaten, which eliminates cannibalism (category 1) as an explanation. Category 2 (sexual selection) can be eliminated in many other cases since the killer does not kill offspring of prospec* Present address: Department of Zoology, University of Arkansas, Fayetteville, Arkansas 72701. Am. Nat. 1991. Vol. 138, pp. 1140-1158. © 1991 by The University of Chicago. 0003-0147/91/3805-0005$02.00. All rights reserved. INFANTICIDE VERSUS ADOPTION 1141 tive mates. Category 3 (parental manipulation) can be eliminated since the killer is not killing its own offspring. Elimination of categories 1-3 in these cases leaves only categories 4 and 5 as possible explanations. Category 4, competition for resources (Pierotti 1980, 1982 ;̂ Sherman 1981), is easily invoked, but this is the most difficult category to demonstrate convincingly (Hrdy and Hausfater 1984). As a consequence, competition for resources and social pathology have become default explanations for many cases of infanticide. This reliance on alleged resource competition and social pathology is unsatisfying. Resource competition is a nebulous concept and could be invoked in any situation since it can be argued that killing a conspecific will always reduce potential competition for resources. To satisfactorily demonstrate that infanticide occurred to reduce competition, it must be shown that (a) the resource in question is actually limiting and (b) killers, or their relatives, thereby gain increased access to that resource. Similarly, pathology can always be invoked when one conspecific kills another. For pathology to be established, it must be convincingly argued that the infanticidal individual is behaving in a manner that indicates malfunction. In this article, I examine the evidence in relation to an alternative explanation for cases of infanticide that have been attributed to either competition for resources or social pathology, that is, avoidance of adoption and provision of parental care to unrelated offspring (Pierotti and Murphy 1987; Pierotti 1988; see also Sherman 1981; Elwood and Ostermeyer 1984; Mock 1984). I do not discuss cases of infanticide that have solid functional explanations. These include instances where the infanticidal adult kills its own offspring to increase chances of survival of itself or of other offspring (parental manipulation) or kills offspring of prospective mates in order to bring the mates into sexually receptive condition, for example, male langurs or lions. In cases where the infanticidal adult consumes or partially eats the offspring it kills, the exact cause of the infanticide may be problematical. As a result, I discuss below several cases where cannibalism occurs in conjunction with infanticide. THE INTERGENERATIONAL CONFLICT HYPOTHESIS All of the proposed explanations of infanticidal behavior ignore the possible role of the offspring in the interaction and treat offspring as if they were passive recipients of their fate. This may be an important oversight, for there will always be strong selection on traits that maximize chances of offspring survival to independence regardless of parental interests (Trivers 1974; Pierotti and Murphy 1987; Pierotti 1988). In most natural populations, some offspring receive inadequate parental care because they have become separated from their parent(s), a parent has died, or their parents are inexperienced or inept. Under these circumstances, selection would exist for traits that enable young to solicit parental care from adult individuals other than their biological parents (Mock 1984; Pierotti and Murphy 1987; Pierotti 1988). Adoption or care of unrelated offspring has been reported from a number of species in which infanticide that has been attributed to either competition for resources or social pathology has also been reported (tables 1 and 2). The re-