Infanticide by Male Lions Hypothesis: A Fallacy Influencing Research into Human Behavior

by Anne Innis Dagg
Infanticide by Male Lions Hypothesis: A Fallacy Influencing Research into Human Behavior
Anne Innis Dagg
American Anthropologist
Start Page: 
End Page: 
Select license: 
Select License


Independent Studies, University of Waterloo

Waterloo, Ontario, Canada N2L 3G 1

Infanticide by Male Lions Hypothesis: A Fallacy Influencing Research into Human Behavior

Recently, familial abuse and killing of children have been correlated with infanticidal behavior by nonhuman male animals which is postulated to be genetic. However, infanticide by males as an evolutionary mechanism is unlikely to occur in pri- mates, and is equally unlikely in lions. To investigate the sexual selection hypothesis of infanticide for lions, four questions should be considered: (1) because they all mate freely with lionesses, are male pride lions closely related? (often they are not), (2) do males kill the young cubs when they join a pride? (few such cases have been observed), (3) do the females then mate with the new males to produce offspring'? (yes), and (4) do the males remain with the pride for over two years to pro- tect this new generation? (often not). Field data indicate that in NO instance do we know that male lions killed cubs in a new pride, mated with their mother, and then remained with the pride to protect their own young. Rather, female lions may kill more cubs than males and certainly cause many more cub deaths by allowing them to starve. [injanticide by male lions,

injanticide hypothesis, faulty scierzce]

The great tragedy of Science-the slaying of a beautiful hy- pothesis by an ugly fact.

-T. H. Huxley (Huxley 196858)

In the mid-1970s, the infanticide by adult males hy- pothesis was formulated to explain the evolutionary sig- nificance of the behavior of some langurs and subsequently the behavior of other primates (Hausfater and Hrdy 1984; Hrdy 1974). It postulates that a mature male newly arrived in a group of females will kill the females' infants so that the females will quickly come again into estrus and mate with him to produce his offspring.

The issue of infanticide as an evolutionary strategy of males, however, has recently been reassessed for primates. After an extensive review of the primate behavioral litera- ture, Bartlett et al. (1993) show that in fact very few males are known for certain to have killed a female's young and then mated with her to produce new offspring, so that their future efforts on behalf of the troop would benefit their own young. These authors found that "most cases of infan- ticide may simply be a genetically inconsequential epiphe- nomenon of tense, aggressive episodes" caused by the "generalized, overt aggression exhibited by adult males"

984). Ln only 6 out of 48 cases of infanticide was there evidence consistent with the sexual-selection hypothesis
976). Although it seems clear from these data that in- fanticide by primate males is not a significant genetic strat- egy (i.e., that those that practice it will have an evolution- ary reproductive advantage over those that do not), the

concept has captured public imagination and gained wide- spread attention; it is repeated in many books, articles, and films on primates. The hypothesis resonates with Western culture in which many people accept male dominance and aggression and condone in part the control of female sexu- ality by men. Reports on infanticide by male lions as an evolutionary strategy that immediately followed those on langurs have also gained widespread public attention, in part because they help to bolster the evolutionary explana- tion for infanticide by primate males.

The intellectual satisfaction with the hypothesis was and is not only great, but apparently cumulative: Hausfater and Hrdy (1984:xiv) write that "Once infanticide began to be explained in evolutionary terms [Hrdy 19741, published re- ports of infanticide in mammals increased dramatically." However, although many researchers apparently thought otherwise, the growing number of reports on infanticide in a wide variety of species neither strengthened nor weak- ened an evolutionary explanation.

At the same time, research into infanticide in some so- cial animals also has had great significance for studies in human behavior, and thus for anthropologists, sociologists, and psychologists. Yet the possible evolutionary mecha- nism is dissimilar because while nonhuman animals almost universally kill young that are not their own, when young children are lulled, it is done usually by their parents (Hrdy 1987:97).

Since 1975, the year that E. 0.Wilson's landmark book Sociobiology appeared, infanticide by male lions (which had one of its first mentions in this work) has been a primary

Anzericun Anthropologi.~t 100(4):940&950. Copyright 01999, American Anthropological Association

example of such research. A review of the literature shows that since that time and continuing to the present, researchers on infanticide in human beings have cited data from lions and other animals to show that such behavior has an evolutionary and not just a cultural base. As early as 1979, David Barash implied directly that the hypothesis could have human implications: "Even humans are not above similar behavior. There is abundant historical evidence that infant killing as a form of 'evil stepparenting' has occurred for thousands of years" (1979:104).(Yet little or no attempt has been made to explain why ALL human cultures do not have this same base if it is biological rather than cultural.) Many academic articles have cited animal data to back up findings of the evolutionary significance of human infanticide (e.g., Daly and Wilson 1987, 1988a, 1988b, 1988c, 1994, 1995; Daly et al. 1982; Dickeman 1975).

If infanticide can be shown to be a product of evolution in human as well as nonhuman animals, then what about child abuse? Malkin and Lamb (1994:121) write: "Extrapolating from animal studies (e.g., Hrdy, 1974, 1976; Nadler, 1980 [their work on primates]; Schaller, 1972 [his book on lions]), researchers have begun to examine the association between biological relatedness and child abuse in humans. . . ."Authors who have looked for support to explain child abuse and neglect from animal studies include Daly and Wilson (1981), Hrdy (1987), Malkin and Lamb (1994), Nadler (1980), Reite and Capitanio (1987), and Wilson and Daly (1987).

Some authors have written popular books that intermix human and animal evidence so thoroughly that human behavior is seen as primarily biological rather than cultural:

When their book Sex, Evolution, and Behavior was published in 1978, Martin Daly and Margo Wilson (pp. 8687) had already accepted the infanticide by male lions hypothesis for lions as well as langurs, even though of the two authors they cite, Schaller had published before the hypothesis was formulated and provided data that undermined it, and Bertram introduced the hypothesis with some hesitation, noting that it was based on "somewhat scanty data" (1975b:65) and writing that the killing of cubs by males "is not a large scale slaughter,and it is difficult to determinehow commonly it occurs" (197%: 478); he himself observed it four times in five years.

Five years later Richard Monis (1983:146)in Evolution and Human Nature expounded on human prehistory as follows: "Suppose that, at one time, hominid males did kill the offspring of strange females. Suppose also that, like lions, they lulled the offspring of females with whom they had never copulated when the latter gave birth. There is nothing unreasonable about either assumption." He goes on to connect this scenario with present-day human beings.

More recently, in Robert Wright's The Moral Animal (1994:68), under the subheading "What Else Do Women Want?," we are reminded that in animals that are not promiscuous (such as the gorilla), the dominant male is indulgent to the group young that are likely his own; it would apparently follow that if women are promiscuous rather than monogamous, the chances are that their young will be less safe. Similarly, if an infant does not belong to the troop male, it may be at great risk, as are langur young that may be killed by another male "as a kind of sexual icebreaker, a prelude to pairing up with the (former) mother."

Because of the widespread and important implications of the hypothesis,which extend even to human beings, it is critical that the existence of infanticide by male lions be examined, since this is the other group besides primates in which the hypothesis has been widely promulgated. The position against the overall hypothesis will be strengthened if the data show that for male lions the infanticide hypothesis is as tenuous as it is for primates. In this paper I analyze behavioral data for lions and show that for this species, too, the infanticide by male lions hypothesis is certainly unproven (indeedimpossibleto prove given the wide-ranging habits of lions), and the infanticidal behavior probably has little or no evolutionary significance. The cub killings that occur seem rather to have been caused by generalized male and female aggression against adults and young at times of intergroupturmoil.

Infanticideby Male Lions Hypothesis

For the evolutionary infanticide by male lions hypothesis to be true, we would expect the following conditions to be important.

  1. The coalition of male lions newly joining a pride would be closely related, because estrus females mate with several or all of the new males and the males do not fight over access to the females; if the males were not related, one would expect that they would compete for mating privileges to try to make sure that their genes (rather than those of an unrelated male) were passed on to the next litter.
  2. A coalition of males, if it succeeds in "talung over" a pride of females, will kill cubs and because of this, the mothers will soon come into estrus again.
  3. The new males will mate with the estrus pride females to produce new cubs that are their own offspring.
  4. The males will stay with the pride and chase away other adult males for over two years until these cubs become more or less independent; otherwise the lives of the cubs might in turn be forfeited or shortened by the next male lions tojoin the pride. If this were the case, the genetic inheritance of the earlier pride males would be lost.

Behavioral Reports on Serengeti Lions from 1966 to Present

Although authors have written about lions in the wild for many years, their long-term behavior has been studied in detail only recently, after biologists began field research on the Serengeti Plains of Tanzania in 1966. The first ob- server was George Schaller who concentrated on two lion prides, the Seronera and the Masai, although he observed many other lions as well. He recognized individuals by natural characteristics or by the ear tags with which he had marked 156 lions. Schaller (1972a) reported that prides were composed of about 3 to 11 female relatives such as grandmothers, mothers, daughters, half-sisters, nieces, and cousins. (Some nomad females tried to raise young on their own, but found this almost impossible because they had to be away from their vulnerable cubs while hunting for food and because of widespread lion aggression.) Pride females often had cubs at the same time, suckled these cubs com- munally, and spent many of their days together. They hunted together to bring down large game to feed them- selves, their young, and the pride males. The females did not form a close bond with the pride males unless they were in estrus and mating, but they needed them not only for reproduction but to help them protect from other lions the pride territory and the prey it contained.

A pride typically produces far more young than the pride area can sustain. As Schaller (1972b:86) noted, "Having few enemies, the lion must regulate its population in part by limiting the number of cubs that reach adult- hood." He found that in lions estrus was often delayed, 80% of mating sessions did not produce young, and 17% of females did not come into estrus (p. 89); the Seronera pride produced only about one quarter the theoretically ex- pected number of cubs (p. 91). Even so, far more cubs were produced than the prey could sustain should they grow to adulthood. It was essential, therefore, that many cubs die, as many did. This could be accomplished by having fe- males ambivalent about their young-newborn cubs weigh only about 1.7 kg so that a lioness has little initial invest- ment in them (Schaller 1972a: 143). Schaller wrote:

The response of a lioness to her cubs is so finely balanced between care and neglect, between her own desires and the needs of her offspring, that the survival of the young ones is threatened whenever conditions are not at the optimum. As an intensely social creature, a lioness prefers to be with other pride members rather than separated from them with her small cubs, and as a voracious eater she bolts any meat herself rather than share it. [I 972a: 142-1431

When environmental or social conditions are poor the fe- male may lose her cubs, but she survives to produce more. Although Schaller's account appears to endorse group se- lection, the fact remains that cub mortality is extremely high from a variety of causes including, to a very large ex- tent, female neglect.

Schaller found that most prides included two or more pride males (rarely one or sometimes up to nine) which re- mained with the pride females as pride males for a few years. Pride males were important in keeping other lions away from the pride area by scent marlung, patrolling, and roaring. Sometimes these males left a pride voluntarily, but usually stronger or more males challenged them and, after considerable turmoil, replaced them as pride males. In the field, a pride of lions comprised a far-flung aggregation of individuals considered a pride because they did not fight or flee when they met each other (although they still squab- bled over a kill). It took Schaller several months to be sure which lions belonged to a pride because members were often many kilometers apart; estrus females and those with cubs tended especially to remain separate from the other pride members. Indeed, he never saw the 13-member Se- ronera pride all together during his entire study. Because of the looseness of each pride, it was impossible to see what all the members were doing at any one time, with the result that most lion behavior went unobserved.

As we have seen, many cubs in a pride did not survive. Schaller (1972x428) gave causes of cub deaths during his

observations as follows:
lions 11 cubs
leopard and hyena 2 cubs
starvation 15 cubs
unknown but not starvation 25 cubs

The 11cubs killed by lions included the following:

3 killed by non-pride males (determined by their spoor)

who did not then "take over" the pride (p. 50) 3 killed "by a lioness, judging from the tracks" (p. 50) 1 killed and eaten by female no. 60 (p. 77) 1killed by a non-pride female (p. 174) 1 inadvertently rolled on by a male (p. 146) 1 killed by a nomadic male (Plate 16) 1 perpetrator not given by Schaller

Schaller found that other causes of death were more common. Starvation killed 15 cubs, not because the moth- ers themselves were starving (p. 190), although their milk might have dried up, but often because they would not share food with their young. Schaller (1972b:86) wrote, "it was depressing to see a starving cub totter to its mother, each rib sharply outlined beneath its unkempt hide, and re- ceive a vicious cuff instead of a bite to eat." (By contrast, sometimes a male lion allowed a cub but not a female to feed near him [p. 861.) At a time when their cubs were starving, the females did not hunt more to provide food for them as long as they had enough meat for themselves (p. 100). Thus the starvation of cubs was often caused by ma- ternal neglect. Two mothers did not protect their cubs killed by a leopard and a hyena (Schaller 1972a: 145,188).

Schaller wrote (1972b:84) that "Many cubs simply dis- appeared, often whole litters of vigorous youngsters, and I think that in some instances they were simply abandoned.

Perhaps leaving them to die in the obscurity of a thicket, their mother returned to the pride's conviviality." He con- cluded (1972a: 190) "Probably more cubs died as a result of having been abandoned by their mothers than through any other cause, but I have no data to support this assumption." He did document some examples of certain and probable abandonment (p. 190-1 91). Rudnai (1973:249), who watched 30 lions in Nairobi National Park over a period of five years (but reported no infanticide by males), hypothesized that it might not be worth a female's while to spend two years rais- ing a single young when by abandoning the cub she could be- come pregnant again and raise a larger litter.

In summary, Schaller did not write in terms of an evolu- tionary infanticide by male lions hypothesis that had not yet been formulated. His data (1972a: 192) showed that about half of all cubs died from a variety of causes and that female lions were far more to blame than males for ob- served or presumed cub deaths.

Brian Bertram, who studied these same two prides from late 1969 to 1974, found that up to 80% of cubs died before reaching adulthood (1975b:59). In carrying on Schaller's re- search project, Bertram (1975a:465) noted that compared to Schaller's his own observations were "less intensive, visits to the prides' areas being made once or twice per week." In a pa- per published in 1975, he analyzed the reproductive history of these lions over seven years, from 1966-1972, as reported both by Schaller and himself. From these data @s figures 1 and 2) he concluded that males did in fact have an evolution- ary advantage if they killed the cubs of pride females when they joined a pride (1975a:479), citing the new hypothesis as applied to langurs: "Male langurs newly in command of a troop frequently attack and kill infants in that troop; it appears that in such cases the same advantages accrue to them as for lions indulging in infanticide" (1975a:481). These comments seem more conclusive than the facts would warrant: for the Seronera and Masai prides over seven years, when Bertram's figures 1 and 2 indicate that there were seven "takeovers," he noted that only "Four cases were found of newly arrived males lulling one or more cubs in prides which they had just taken over" (1975a:473). Of course, cubs are small and could readily be lulled and eaten without this being observed by re- searchers. However, Bertram's charts show that cub deaths were not as closely correlated with male "takeovers" as Packer and Pusey (1 984) would later aver. If "takeovers" occurred when food was scarce (such as in the summer of 1969 in the Seronera pride [Schaller 1972a: 18,190; Bertram's fig- ure I]), it was much more likely that cubs died from starva- tion or abandonment rather than from being killed by males. Bertram (1978:99) writes:

Certainly the new males do not indulge in a wholesale slaugh- ter of cubs, but on the other hand the takeover of a pride does result in an increase in the mortality of cubs of any age in that pride. Some of these deaths may be only indirectly due to the new males. Possibly, for example, their presence causes stress for the lionesses who therefore hunt less efficiently, or pro-

duce less milk, or waste time and effort trying in vain to guard their cubs. . . .

Bertram (1976) reports that because pride males were closely related, there was no competition to mate with pride females; whichever male sired a cub would provide it with some genetic material common to all the males.

The Serengeti research on lions was next carried on by Jeannette Hanby and David Bygott from 1974 to 1978. They broadened the research to cover 15 to 20 prides of about 300 lions in the Serengeti and in the Ngorongoro Crater (Packer et al. 1988:363) while focusing especially not on the Seronera and Masai prides, but on the Sametu Pride that was formed in 1974 from nine young females born in 1971 into the Masai Pride. The data they collected, published in part in Lions Share (Hanby 1982) (a popular rather than an academic/scientific book such as Schaller's), agree with the information gathered by Schaller and Ber- tram (although not with the infanticide hypothesis). Many cubs died, usually in response to extreme conditions when large game had migrated elsewhere; this caused starvation of cubs, their abandonment by their mother herself desper- ate for food, and the harassment of pride members by other lions traveling long distances in search of prey (chapter 10). They surmised that the Loliondo pride males may have killed some cubs, and one certainly mated with a Sametu female (p. 159), but these males did not sub- sequently "take over" the Sametu pride, nor did the two groups remain connected in any way (p. 215). For this pride over a five-year period there was no report of males killing cubs when they "took over" a pride.

Craig Packer and Anne Pusey continued in-depth lion research in the Serengeti and in the Ngorongoro Crater from 1978 to 1981, during which time 1,237 cubs were born and one case of infanticide by a male was observed (Packer et al. 1988:365). Unllke Schaller, they concen- trated not on the activities within two prides, but each day "recorded the incidence of wounds and the reproductive condition and mating activity of the females of as many prides as possible" (Packer and Pusey 1983a:717). They have not written up their continuing observations in a book as Schaller did, but rather produced a number of scientific articles. Although they reported extensively on supposed killing of cubs by males, they did not include data on fe- males killing either cubs or adult lions, which would have weakened their hypothesis. (We know that females they observed did kill other lions, because this is mentioned in Packer's popular book Into Africa (1994: 100) and in their 1997 article in Scientijc American [Packer and Pusey


Packer and Pusey (1984) documented a total of 7 cases, involving a total of at least 11 cubs, in which cubs were seen to have been killed by male lions since 1966 (but they did not include Schaller's data of 5 cubs being killed by fe- male lions). However, noting a case of infanticide by a male does not mean that it has evolutionary significance, as the four conditions postulated above indicate.

The first of 7 cases of cub killings by male lions put for- ward by Packer and Pusey to illustrate their hypothesis was observed by Schaller, who reported that the Kamarishe males killed three Seronera cubs. However, Packer and Pusey (1984:32) write that "the Kamarishe males were not seen mating with females of the Seronera pride until 6 months after lulling the cubs"; no males associated with the Seronera pride for several months, and then for a short period both Masai and Kamarishe males mated with them until the Masai males, not the Kamarishe males, joined the Seronera pride permanently (Schaller 1972a:50-5 1). This example therefore is not one of males killing cubs at the time of a "takeover." Packer and Pusey stated that publish- ed details were unavailable for the three cases of infanti- cide noted by Bertram.

Packer and Pusey (1984:32-33) themselves saw the fifth case in 1980 when a male killed two cubs:

The infanticidal male was one of a group of four males that were simultaneously resident in two adjacent prides. His three companions started to associate with this third pride 6 months before the infanticide, but the infanticidal male was not seen with the pride until 6 weeks after the mother had conceived and was not observed with the female again until the day he killed the cubs. During the period in which the mother had conceived, a second male group was also observed mating with these females. Several weeks after killing her cubs, the infanticidal male was observed consorting with the female.

This example is perplexing. In effect, a male member of a pride killed cubs in that pride six months after his male group joined the pride. Thus he could have killed the cubs sired by his own male cohort (or, unobserved by Packer and Pusey, even by himself). In any case, the killing was not associated with a "takeover" that presumably had oc- curred six months earlier.

In a sixth case in mid-1 982, a nomadic male was seen to kill a cub: "The male had not associated with the pride be- fore, and there is evidence that a male takeover was in pro- gress" (1984:33). However, Packer and Pusey give no fur- ther information on what then happened. In a seventh case observed farther north, three new males hlled two cubs, became resident with the pride, and "the infanticidal male might have fathered the female's next litter." Again there is no further information on what happened next, or even any confirmation that the infanticidal male was, in fact, the fa- ther of the new litter.

In summary, in over 17 years of observations of many thousands of lions, there were only seven incidents, involv- ing at least 11cubs, of known infanticide by males (and ap- parently at least 5 of infanticide by females even though for most of this period infanticide by females may have oc- curred and been observed but not reported); in NO instance is there evidence that the males that lulled the cubs then mated with their mothers and remained with them for two

years to protect the new cubs until they were more or less independent.

A few more cubs were killed by males after the 17-year period. Packer (1994:73) wrote that in 1987 a camera crew following successively two mothers with young cubs day after day and into the night for about six weeks saw the cubs hlled by males. One female was a member of the Kibumbu pride who then came into estrus two days later and mated with the male; the other was a lone female whose pride mates had died earlier. The first incident could be an example of the infanticide by male lions hypothesis, but both may also reflect harassment by the camera crew. By 1991, Packer (1994:73) reported that infanticide by male lions had been observed about a dozen times, but he did not further expand on these events.

Packer and Pusey (1984:33) write that in addition to their observations, "there is circumstantial evidence that infanticide occurs almost every time a new coalition of males takes over a pride." The circumstantial evidence was the disappearance of cubs. As we have seen, the seven ob- servations do not corroborate the hypothesis (although in some cases they could support it if we had further informa- tion), so that the hypothesis rests entirely on circumstantial evidenceldisappearance of cubs. By 1987, Packer and Pusey were asserting that not almost all but all cubs died, stating: "Incoming males kill or evict all of the dependent young when they first enter the pride" (1987:638), citing Bertram (1975a) who did not state this, themselves, and Hanby and Bygott (1987) who, in turn, had cited Packer and Pusey's work (p. 164). Packer et al. (1988:373) noted that "Since cubs' deaths are rarely witnessed, it is usually difficult to determine the exact cause. Nevertheless, there are two contexts in which the cause of death can be reliably inferred: male takeovers and season of low prey availabil- ity." They claimed, on circumstantial evidence based on their unsubstantiated hypothesis, that "All cub mortality within two months of a male takeover is therefore consid- ered here to be due to infanticide" (Packer et al. 1988: 374)-presumably in two months the males would have time to come into contact with all the pride mothers and their young.

In summary, there are two trends present during these periods of surveillance of Serengeti and Ngorongoro lions. In one, increasing numbers of lions were observed by suc- cessive watchers which necessitated individual lions being given less attention. In the other, observers increasingly re- ported evidence allegedly reinforcing the infanticide by male lions hypothesis, but less often gave information that would put this behavior in context (such as the extent of starving cubs and of infanticide by females). The first observer (Schaller) noted infanticide of cubs by both male and female adults. The second (Bertram) reported only in- fanticide by males and that the number of cub deaths in- creased at the time of male "takeovers." The third set of ob- servers (Hanby and Bygott) did not focus on this behavior, and the fourth (Packer and Pusey) assumed with almost no evidence that all cub mortality within two months of a male "takeover" was caused by infanticide by males.

Is the Infanticide by Male Lions Hypothesis True?

Because the unwarranted assumption that infanticidal behavior by males is important in lion evolution skews the information provided by later lion workers, and because later workers did not provide many raw data, the facts pre- sented here are taken largely from the work of Schaller and other early researchers who published before the hypothe- sis of sexual selection by infanticide had been formulated. Here I shall consider the four relevant conditions as stated earlier that should be considered in assessing the infanti- cide by male lions hypothesis:

1. Pride males should be closely related as are brothers1 half-brothers1 cousins

Pride males of a coalition mate freely with pride fe- males, with both males and females having multiple part- ners (Schaller 1972a). Copulations for a couple have to- taled 157 in 55 hours, or an average of one every 21 minutes (Schaller 1972b:89). When the male tires, the fe- male may continue mating with a second male. Bertram (1975a:77) several times saw an estrus female "get up quickly and trot over to a waiting male, apparently before her attendant consort realized what was happening. At once the tables were turned, and the new male was in pos- session and dominant over the previous boss."'

The initial belief that male pride members were closely related, either brothers or half-brothers or cousins, fitted with the infanticide by male lions hypothesis; it did not matter greatly which males mated with the pride females because the result would be offspring with a similar ge- netic inheritance (see Bertram 1976). However, Packer and Pusey (1982) more recently report that in fact "our data show that 42% of breeding coalitions of known origins contained non-relatives" (p. 740). This information would seem to weaken the close-relationship hypothesis; al- though the authors provide an explanation for the mixed- breeding coalition (it prevents male conflict and keeps the group together), they do not acknowledge that it under- mines their fundamental hypothesis. In reality, a male might stay with a pride for years and protect it even though none of the young were his or his relatives.

2. New pride males kill the cubs of pride females who quickly come into estrus again

Males do sometimes kill cubs when they join a new pride, and their mothers then do come into estrus and may mate with them a few days later, but only a few such cases have been seen. One would presume that a male lion that carried the genetic components that made him kill unfamil- iar cubs would pass this inheritance to his own sons if this were an evolutionary strategy. However, observed cases of males killing cubs are rare, suggesting that infanticide by males is uncommon, and certainly not universal, although it could easily be if new males were persistent and deter- mined to kill cubs. If males that committed infanticide were more successful reproductively than those that did not, one would expect that all cubs would be killed when new males join a pride, but this does not happen (Bertram 1975a: figures l,2). It may be that cub-hlling only occurs with a few bad-tempered irascible males. In other words, for the hypothesis to be validated, we would have to know that the "evolutionarily adaptive" behavior is uniformly distributed across all or most male adults. If the attributed behavior is adaptive as the hypothesis suggests, then all or most males should lull all cubs new to them. Packer (1994:75) indeed states that all male lions are infanticidal and have probably murdered at least one cub each, but there is no evidence that this is so.

One might infer from the infanticide by male lions hy- pothesis that general aggression among lions is rare; pro- ponents of the hypothesis seldom mention that fighting and killing among all lions is not uncommon. For example, ag- gression was widespread even within a pride during com- munal feeding (Schaller 1972x132-136): 6% of lions were blinded by such encounters and others were wounded. One female was bitten on the hip by her son while both were at a kill (Scott 1992239).

Aggression was far greater outside the pride, with non- related lions sometimes killing each other if they came in contact. In the Nairobi National Park, for example, two males killed two non-pride females and a sub-adult male in 196 1 and 1962, and two more females were killed in 1963, although Schaller (1972a:47) notes that such mayhem was unusual. Females on occasion kill male lions (Pusey and Packer 1994:297), other females (Packer and Pusey 199754) and cubs (Schaller 1972a), just as do males. Schaller (1972a:428) in his research reported that females apparently killed as many young (5 cubs) as did males. It is unclear why the aggression of lions against each other can be taken for granted and largely ignored, while a small sub- category of that behavior-the killing of cubs by males only-merits an evolutionary explanation. There seems to be no reason to give a complicated evolutionary explana- tion for the occasional case of infanticide by male lions; it is both simpler and more plausible to explain such cases by general aggression when new males and females come into contact, conflicts which have caused the death of adults as well as cubs.

3. New pride males then mate with pride females to produce cubs

DNA tests by Gilbert et al. (1991) indicate that in a sam- ple of 78 cubs, all had pride males as fathers. However, cubs could also be fathered by nomads. Schaller (1972a:SS) writes that "any lioness in estrus accepts and is accepted by both pride males and nomads." Scott (1992:53) noted that the females "willingly mate with any new males wandering through their area." It is almost im- possible, therefore, to know with what males a female has mated. The situation is complicated further in that 80% of sexual sessions that Schaller (1 972a: 178) observed did not result in pregnancy. We do not know with certainty in the field, therefore, which males fathered which cubs.

Linked with the infanticide hypothesis is the observation that pride females usually have cubs about the same time; the inference is that the "takeover" males will have killed all the cubs, and therefore this synchrony is to be expected (Packer and Pusey 1983b). However, starvation of cubs and their abandonment when prey is scarce produce the same dearth of cubs at one time. Synchrony is physiologi- cal, with estrus in one female apparently triggering that condition in her colleagues (Schaller 1972a: 180); thus pride females tend to mate and produce cubs within a short time period. This is beneficial because all the females suckle all the pride cubs and act as inadvertent guards when they stay nearby (Schaller 1972a:149, 146). If cubs of varying ages are born into a pride, it may prove lethal for the younger ones; Scott (1992:53) described the Bila Shaka pride in which three females at about the same time gave birth to 10 cubs, with a fourth female producing four more cubs some months later. He wrote that "the benefits of syn- chronized births soon became apparent. Despite her efforts to keep the older cubs from toying endlessly with her young, the lioness found it impossible to keep them at bay." Before long her younger cubs were all dead.

4. Pride males will stay with the pride for over two years

If males newly joining a pride kill all the young cubs and chase away the larger offspring (Packer et al. [1988:374] state that young "continue to be at risk [from infanticide] until about twenty months"), then it seems important that the male lions that sired the cubs stay with the pride to pro- tect them until they are old enough to survive on their own.

After a gestation period of three and a half months, cubs remain dependent on their mothers for food for at least two years (Bertram 1975b:59). Some potential pride males mate with pride females who are not currently near a male, but do not then stay with the pride, so that their progeny (if any) remain unprotected (Schaller 1972x42). In some cases, males have been reported to leave a pride of females voluntaily, and so withdraw their protection from the pride and their young (Schaller 1972b:71). The impetus for males to mate and then remain with the female pride is not, therefore, imperative. In NO case is it known that after "taking over" a pride and lulling the cubs, the new males then remained in the pride until their own progeny were grown. Moreover, as already noted, some cubs might have been sired by nomad males.

In summary, the foregoing makes clear that pride males are often not related; that up to 80% of cubs die for a vari- ety of reasons far more often related to female rather than male behavior; that pride females may mate with non-pride males or with unrelated pride males with the result that an infanticidal male could kill his own young or could help raise another's young believing they were his own; and that a male may mate with a pride female yet make no ef- fort to become a pride male to ensure the future safety of the pregnant female and her cubs.

As with primates, it seems certain that the death of large numbers of young is caused not by infanticide by males as an evolutionary mechanism, but by shortage of food, by the lack of pride males that leaves young vulnerable to predators, and by general hostility that occurs with a change in male attachment to a pride. Scott (1992:88) wrote that after a male "takeover" "two or three months of disruption and hostility" occurred before life returned to normal. Hanby (1 982) describes in graphic detail such dis- ruption when there was little food or no pride males--cubs died of starvation because their mothers could not feed them, and cubs were abandoned when a mother decided to stay with other pride females rather than return a long dis- tance to her cubs.

Spurious Science and Its Consequences

The issue of infanticide by male lions as an evolutionary mechanism is a classic case of a satisfying hypothesis be- coming accepted in the academic world and then in popu- lar culture far beyond proper validation, even though the scientific procedure followed seemed to be adequate, with scientific papers reviewed by referees. How could this hap- pen and what are the repercussions?

Researchers who supported the hypothesis did not supply data that made it possible for readers to evaluate for themselves if the hypothesis was true; they had to accept the authors' opinion. (This is not surprising, as scientific journals are typically short of space and reject raw data.)
Researchers ignored or downplayed data that contra- dicted the hypothesis. For example, although Packer et al. (1988:374) stated that "All cub mortality within two months of a male takeover is therefore considered here [in their paper] to be due to infanticide," Bertram (1975a) and Schaller (1 972a) showed this not to be so, since cubs died of many causes at all times of the year. Concerning starva- tion of cubs, Packer et al. (1988:374) write "Schaller (1 972) [1972a] suggested that 50% of cub mortality in the Serengeti was the result of starvation, but though many malnourished cubs have been observed, the death of a starving cub has been seen only once." (Packer [1994:70] later wrote, however, that during their first few years in the Serengeti prey was scarce on occasion and "many small

cubs starved.") For one of many examples, the nine fe- males that made up the Sarnetu pride were all born into the Masai pride in 1971 and left the pride together in 1974 (Bertram 1975a: figure 2; Hanby 1982:213); during that time there were two male "takeovers" in the Masai pride, yet these nine young females not only survived, but thrived.

(3) Researchers published papers based on invalid as- sumptions, as the following examples show. Packer and Pusey's (1983a) article entitled "Adaptations of Female Lions to Infanticide by Incoming Males" has two funda- mental problems: (I) it is based on the unfounded assump- tion that the infanticide by male lions hypothesis is valid, and (2) it explains female lion behavior as a response to this unfounded hypothesis. The five adaptations they cite (defense of cubs, avoidance of new males, pseudo-estrus, spontaneous abortion, and abandonment) cannot be adap- tations to infanticidal evolutionary behavior by males if that behavior rarely if ever happens. The female lions' be- havior seems simply to be a response to stress caused by generalized male and female aggression.

As already noted, Packer et al. (1988:374) state that all cub mortality within two months of a male takeover was due to infanticide. They continue (p. 381): "The results in figure 23.5 and table 23.4 show that infanticide is a major cause of cub mortality . . .", although these data are based on circumstantial evidence, since as mentioned earlier the killing of cubs by males has been observed only about a dozen times in 25 years (Packer 1994:73), and by far the most cubs die of other causes. (How can one justifiably graph dozens of incidents of infanticide when it is un- known if any at all occurred?)

Using the two-months' assumption, Packer et al. (1988) calculated the proportion of cub deaths attributed to infan- ticide by noting how many cubs died during the two months after a takeover as a percentage of how many died over the year, concluding that "27% of all mortality before twelve months of age occurs in this context," that is, the context of a recent takeover (p. 374). They ignore the fact that cubs die from many other causes even at the time of a "takeover"; in the Seronera pride, for example, Schaller (1972a: 190) reported that lack of food was causing cubs to die of starvation in the summer of 1969 when a "takeover" occurred (see his figure 1). In the methodology of Packer et al.'s 1988 paper, such cubs that died of starvation would be counted as having been killed by males. Although this value of 27% is highly questionable, it is referred to by Packer and his colleagues again in 1990, 1994, and 1997 (Packer et al. 1990; Packer and Pusey 1997; Pusey and Packer 1994).

(4) Other animal researchers accepted the published data and cited them in their own research papers, giving the hy- pothesis legitimacy that it should not have had. For exam- ple, Hrdy et al. (1995: 153) bolster the argument for the in- fanticide hypothesis in primates with supporting evidence

from lions, stating that for this species "27% of all cub deaths in first year [are] due to infanticide," citing the work of Packer and his colleagues. The value 27% sounds de- finitive, yet as we have seen has no firm basis in reality.

Researchers have used the infanticidal data on lions to bolster sociobiological research on human beings, an in- sidious approach because it makes cultural practices seem inborn rather than culturally transmitted; there is little hope of changing society if social behavior is considered as a ge- netic trait.
The infanticidal hypothesis was made to seem so certain (despite being based entirely on circumstantial evi- dence), that criticisms of it were disregarded. In 1982 I could find no scientific journal that would publish my cri- tique of it. Eventually I was able to publish short appraisals in what became a little-read book and an article in a femi- nist journal (Dagg 1983,1984).

Sarah Blaffer Hrdy, who studied at Harvard University under E. 0.Wilson and postulated the infanticide by males hypothesis for langurs (1974), argues along with her col- leagues (1995) (in a response to and refutation of the posi- tion of Sussman et al. [1995]) that there are two types of re- searchers of primate social behavior. The first, who are made to sound like unimaginative drones, focus on specific groups and do not generalize "beyond the specifics of the case in hand (Hrdy et al. 1995:154). The second, with whom she presumably identifies, "derive their greatest pleasure from noting that so many findings could have been correctly predicted on the basis of pitifully incom- plete data sets merely by relying on logic, comparisons, and extrapolations guided by evolutionary theory." How- ever, there is also a third category of researchers who are so enamored of theory that when the whole of the data do not fit their hypotheses, they ignore basic facts (often produced by the unimaginative drones) in order to salvage at all costs their theory-driven hypothesis. This can lead to fallacious conclusions and is not good ~cience.~


It is lamentable that the infanticide by male lions hy- pothesis disproven here has been so widely accepted. It has been routinely cited since 1975 as fact and has promoted for other species the notion that infanticide must have an evolutionary purpose. Pusey and Packer (1994:29&297) note that sometimes male lions lull a female lion, which "is disadvantageous both to the female and the males in- volved, so why should it ever occur?" I think the answer is that not all animal behavior is evolutionarily adaptive. In- fanticidal behavior has no apparent adaptive significance in pinnipeds where males often kill pups, although such deaths do not cause the mother to resume ovulating sooner, nor is the male more apt to mate with her than other females (Le Boeuf and Campagna 1994). For another example, this time for chimpanzees at Gombe in Tanzania during two decades of observations, seven cases of infanticide have been recorded; four have been by males, while three have been by a mother-daughter pair killing and eating young from their own communities (Fossey 1984:232). Fossey writes (p. 234) that the data raise "the possibility that some of the cases reviewed may indeed reflect idiosyncratic or pathological behavior of litte evolutionary significance."

Behavior of human beings is far more likely than that of other animals to be culturally influenced rather than being primarily the result of organic evolution. It is regrettable, therefore, that research into human infanticide and child abuse has drawn in part on animal research, specifically on the sexual selection hypothesis of infanticide, for its vali- dation. Human studies that emphasize sociobiology and the predominance of biology over culture make the possi- bility of change for society seem more difficult. If sociobi- ological studies are not based on reliable data, they cause two-fold problems: the published errors appear as true facts and may influence other research, and these false data bolster the belief that change in human society is difficult if not impossible.


Ever since the mid-1970s when the infanticide by males hypothesis was formulated for langurs, it has also been ap- plied to lions, even though the behavioral data to support it are weak, illusory, or circumstantial. In reality, aggression is common among all lions; adult male and female lions as well as cubs are lulled by both male and female adults at times of stress (such as lack of food or inter-group turmoil). Primary data show that the females much more than the males are responsible for cub deaths, both by killing non- pride cubs directly and by abandoning their own cubs to starvation. Moreover, even when males do kill cubs, this behavior appears to be indiscriminate, with little evidence suggesting that it is directly linked to bringing females into estrus or to the "takeover" of prides.

Daly and Wilson's (1988b:108) pronouncement that "The reluctance of many writers to consider infanticidal behavior adaptive seems to derive more from the distaste- fulness of the phenomenon than from evidence or theory" is misleading. What needs to be explained is not the reluc- tance to accept distasteful theory, but the willingness to promulgate unfounded theory. The reluctance to accept in- fanticide in lions as an evolutionary adaptation stems from the faulty evidence on which the hypothesis is based. The refutation of this hypothesis for lions indicates that related hypotheses for other species, including human beings, should be reassessed.

No one would deny the existence of many genetic sources of human behavior. However, sociobiology clearly treads on dangerous ground when it incorrectly gives un- warranted prominence to the biological basis of human be- havior. When it does so, it locates us in a biological fatal- ism that a correct cultural interpretation would avoid.


Acknowledgments. I am grateful to Brent Wootton for his feedback on this paper, to Alan Cairns who suggested im- provements and discussed the paper's various implications with me, and to anonymous reviewers who offered much con- structive criticism.

  1. Bertram's choice of words such as in possession, domi- nant, and previous boss seem curiously male-biased given the situation he is describing.
  2. For further case studies about the perpetuation of error in academic analyses, see Richard Hamilton's The Social Mis- construction of Reali~ (1996); his trenchant criticisms apply to such eminent figures as Max Weber and Michel Foucault and their subsequent interpreters.

References Cited

Barash, David 1979 The Whisperings Within. New York: Harper and Row. Bartlett, Thad Q., Robert W. Sussman, and James M. Cheverud

1993 Infant Killing in Primates: A Review of Observed Cases with Specific Reference to the Sexual Selection Hy- pothesis. American Anthropologist 95:958-990.

Bertram, Brian C. R. 1975 Social Factors Influencing Reproduction in Wild Li- ons. Journal of Zoology, London 177:463482. 1975b The Social System of Lions. Scientific American 232(5):54-65.

1976 Kin Selection in Lions and in Evolution. In Growing Points in Ethology. P. P. G. Bateson and R. A. Hinde, eds. Pp. 281-301. Cambridge: Cambridge University Press.

1978 Pride of Lions. New York: Scribners. Dagg, Anne Innis 1983 Harems and Other Horrors: Sexual Bias in Behavioral Biology. Waterloo, ON: Otter Press.

1984 Sexual Bias in the Literature of Social Behaviour in Mammals and Birds. International Journal of Women's Studies 7(2): 1 18-1 35.

Daly, Martin, and Margo Wilson 1978 Sex, Evolution, and Behavior: Adaptations for Repro- duction. North Scituate, MA: Duxbury Press.

1981 Abuse and Neglect of Children in Evolutionary Per- spective. In Natural Selection and Social Behavior: Recent Research and New Theory. R. D. Alexander and D. W. Tin- kle, eds. Pp. 405416. New York: Chiron.

1987 Evolutionary Psychology and Family Violence. In Sociobiology and Psychology: Ideas, Issues and Applications. Charles Crawford, Martin Smith and Dennis Krebs, eds. Pp. 293-309. Hillsdale, NJ: Lawrence Erlbaum Assocs.

1988a Evolutionary Social Psychology and Family Homi- cide. Science 2425 19-524.

1988b The Darwinian Psychology of Discriminative Paren- tal Solicitude. In Comparative Perspectives in Modern Psy- chology: Nebraska Symposium on Motivation, 1987. Daniel W. Leger, ed.Pp. 91-144. Lincoln, NE: University of Nebraska Press.

1988~ Homicide. New York: Aldine de Gruyter Press.

1994 Some Differential Attributes of Lethal Assaults on Small Children by Stepfathers versus Genetic Fathers. Ethology and Sociobiology 15:207-2 17.

1995 Discriminative Parental Solicitude and the Relevance of Evolutionary Models to the Analysis of Motivational Systems. In The Cognitive Neurosciences. Michael S. Gaz- zaniga, ed. Pp. 1269-1286. Cambridge, MA: MIT Press.

Daly, Martin, Margo Wilson, and Suzanne J. Weghorst 1982 Male Sexual Jealousy. Ethology and Sociobiology

3:1 1-27. Dickeman, Mildred

1975 Demographic Consequences of Infanticide in Man. In Annual Review of Ecology and Systematics, vol. 6. Rich- ard F. Johnston, Peter W. Frank, and Charles D. Michener, eds. Pp. 107-137. Palo Alto, CA: Annual Reviews Inc.

Fossey, Dian 1984 Infanticide in Mountain Gorillas (Gorilla gorilla ber- ingei) with Comparative Notes on Chimpanzees. In Infanti- cide: Comparative and Evolutionary Perspectives. Glenn Hausfater and Sarah Blaffer Hrdy, eds. Pp. 2 17-235. New York: Aldine Publishing.

Gilbert, D. A., C. Packer, A. E. Pusey, J. C. Stephens, and

S. J. O'Brien

1991 Analytical DNA Fingerprinting in Lions: Parentage, Genetic Diversity, and Kinship. Journal of Heredity 82: 378-386.

Hamilton, Richard F. 1996 The Social Misconstruction of Reality: Validity and Verification in the Scholarly Community. New Haven, CT: Yale University Press. Hanby, Jeannette 1982 Lions Share: The Story of a Serengeti Pride. Boston: Houghton Mifflin. Hanby, J. P., and J. D. Bygott 1987 Emigration of Subadult Lions. Animal Behaviour 35:161-169. Hausfater, Glenn, and Sarah Blaffer Hrdy 1984 Infanticide: Comparative and Evolutionary Perspec- tives. New York: Aldine. Hrdy, Sarah Blaffer 1974 Male-Male Competition and Infanticide among the Langurs (Presbytis entellus) of Abu, Rajasthan. Folia Pri- matologica 22: 19-58. 1976 Care and Exploitation of Nonhuman Primate Infants by Conspecifics other than the Mother. In Advances in the Study of Behavior. Jay S. Rosenblatt, Robert A. Hinde, Evelyn Shaw, and Colin Beer, eds. Pp. 101-158. New York: Academic Press. 1987 Sex-Biased Parental Investment among Primates and Other Mammals: A Critical Evaluation of the Trivers-Wil- lard Hypothesis. In Child Abuse and Neglect: Biosocial Di- mensions. Richard J. Gelles and Jane B. Lancaster, ed~. Pp. 97-147. New York: Aldine de Gruyter.

Hrdy, Sarah Blaffer, Charles Janson, and Carel van Schaik 1995 Infanticide: Let's Not Throw Out the Baby with the Bath Water. Evolutionary Anthropology 3:15 1-154. Huxley, Laura Archera 1968 This Timeless Moment: A Personal View of Aldous Huxley. New York: Ballantine Books.

Le Boeuf, Burney J., and Claudio Campagna 1994 Protection and Abuse of Young in Pinnipeds. In Infan- ticide and Parental Care. S. Pannigiani and F.S. von Saal, eds. Pp. 257-276. Chur, Switzerland: Harwood Academic Publishers.

Malkin, Catherine M., and Michael E. Lamb 1994 Child Maltreatment: A Test of Sociobiological The- ory. Journal of Comparative Family Studies 25: 121-133. Morris, Richard 1983 Evolution and Human Nature. New York: Seaview1 Putnam. Nadler, Ronald D. 1980 Child Abuse: Evidence from Nonhuman Primates. Developmental Psychobiology 13507-5 12. Packer, Craig 1994 Into Africa. Chicago: University of Chicago Press.

Packer, Craig, Lawrence Herbst, Anne E. Pusey, J. David

Bygott, Jeannette P. Hanby, Sara J. Cairns, and Monique

Borgerhoff Mulder 1988 Reproductive Success in Lions. In Reproductive Suc- cess. T. H. Clutton-Brock, ed. Pp. 363-383. Chicago: Uni- versity of Chicago Press.

Packer, Craig, and Anne E. Pusey 1982 Cooperation and Competition within Coalitions of Male Lions: Kin Selection or Game Theory? Nature 296: 74G742. 1983a Adaptations of Female Lions to Infanticide by In- coming Males. American Naturalist 121 :7 16728. 1983b Male Takeovers and Female Reproductive Parame- ters: A Simulation of Oestrous Synchrony in Lions (Pan- thera leo). Animal Behaviour 3 1334-340. 1984 Infanticide in Carnivores. In Infanticide: Comparative and Evolutionary Perspectives. Glenn Hausfater and Sarah Blaffer Hrdy, eds. Pp. 31-42. New York: Aldine.

1987 Intrasexual Cooperation and the Sex Ratio in African Lions. American Naturalist 130:636442. 1997 Divided We Fall: Cooperation among Lions. Scien- tific American 276(5):52-59. Packer, Craig, D. Scheel, and Anne E. Pusey 1990 Why Lions Form Groups: Food Is Not Enough. American Naturalist 136: 1-19. Pusey, Anne E., and Craig Packer

1994 Infanticide in Lions: Consequences and Counter-strategies. In Infanticide and Parental Care. S. Parrnigiani and F. von Sad, eds. Pp. 277-300. London: Harwood Aca- demic Publishers.

Reite, Martin, and John P. Capitanio 1987 Child Abuse: A Comparative and Psychobiological Perspective. In Child Abuse and Neglect: Biosocial Dimen- sions. Richard J. Gelles and Jane B. Lancaster, eds. Pp. 149-174. New York: Aldine de Gruyter.

Rudnai, Judith 1973 Reproductive Biology of Lions (Panthera leo mas- saica Neumann) in Nairobi National Park. East African Wildlife Journal 1 1 :24 1-253.

Schaller, George B. 1972a The Serengeti Lion. Chicago: University of Chicago Press. 1972b Serengeti: A Kingdom of Predators. New York: Knopf.

Scott, Jonathan Wilson, Margo, and Martin Daly

1992 Kingdom of Lions. London: Kyle Cathie. 1987 Risk of Maltreatment of Children Living with Step- Sussman, Robert W., James M. Cheverud, and Thad Q. Bartlett parents. In Child Abuse and Neglect: Biosocial Dimen- 1995 Infant Killing as an Evolutionary Strategy: Reality or sions. Richard J. Gelles and Jane B. Lancaster, eds. Pp.

Myth? Evolutionary Anthropology 3:149-15 1. 215-232. New York: Aldine de Gruyter. Wilson, E. 0. Wright, Robert 1975 Sociobiology: The New Synthesis. Cambridge, MA: 1994 The Moral Animal: Evolutionary Psychology and Belknap Press. Everyday Life. New York: Vintage Books.

Representdtions of the Exotic
C'ercr M.lson explorcs rhc tcuts, yantlngs,
dr.rnlngL l~hocoSr.~Phb, and museum
d~\~>l,~~sIn \r.h~ih the exot~c has bccn Race, Cldss, Ethnicity, 2nd Ndtionalism
  Case Scudies in
  edlted by N~naGlirk Srh~lle~ Linda Basrh, African Gender
    gro~~~)of norcd author~t~c\ to offrr .rn In-
    depth look at women who bclo~ig to royal
    dnd cl~tcgroLIps In s~~b-S.lli.lra~i Afr~ca.
    R~scdon or~g~li~llrese.lrch ionducrcd
  high level oS~nvol\ernent ~n both thc~r 1~1ostIy111 N1gcr1.1, tlic C.11ncroo11. Gh.111a. .rnd the Rcp~lblli of Renln, clicsc c~ghteen
  home .lnii host soilcclcs. The .1~1thors si~ggesc thc cr.~ns~i.~t~o~~.ll IIVES of colltcnl- CA~Csti~d~csoffer .I rn~~lr~d~si~~>l~naryportr.lyal of ASr~c.ln
    women who excrilsc power In borl~ the
    trlnpor.~l .Inti sp~r~runl
    ilnn.iI5 of-cllc, hi.r~>brX ilr;1iir1n, of~.'
  • Recommend Us