¿ìè¶ÌÊÓƵ

The contractions deepen and the births begin. One hour apart, the two
cubs emerge, their eyes open, incisors and canines fully developed. The
exhausted mother licks her offspring. But within minutes the scene turns
nasty as the first-born cub grips its twin over the shoulders and shakes
it. The pair roll in a bitter embrace, each with the skin of the other locked
in its jaws.

Fortunately, this scene is unfolding in Berkeley California rather than
the African plains, so we can intervene if one cub is in serious jeopardy.
But out in the wild the vicious dual may continue until one cub emerges
victorious – frequently leaving the other dead. Alone among wild mammals
studied to date, the spotted hyena, Crocuta crocuta, produces infants that
appear to be genetically programmed to attack and, in many cases kill their
siblings. Even in captivity, their Cain-like tendencies are obvious. Newborn
hyenas will try to ravage inanimate objects such as a rolled-up towel that
are about the same size and texture as a sibling. Cubs may even attack a
brother or sister that has not yet emerged from its amniotic sac.

The aggression has profound consequences for the hundreds of thousands
of spotted hyenas that roam Africa. I have been studying one large clan
of hyenas in Kenya’s Masai Mara National Reserve over the past 15 years,
and from the start I was surprised by the large number of single cubs. Captive
spotted hyenas produce twins, and observations of fetal litters in the wild
show the same. Yet by the time the litters I observed in the wild were brought
to the communal den, about half had been reduced to a singleton.

Initially, I could see no obvious reason why the cubs should be dying.
Some had scabs and scars on their shoulders which I assumed at first were
caused by some sort of skin disease. But years later the truth dawned. The
cubs were killing each other and, more intriguing still, the killing was
confined to twins of the same sex. The real puzzle became clear: why should
evolution have selected such a seemingly destructive genetic trait? What
advantage does siblicide offer spotted hyenas?

From my own observations of hyenas in the wild, and my and my colleagues’
studies of animals kept in captivity in California, we are beginning to
find out. The clan I study normally numbers over 70 animals, including about
20 cubs. Spotted hyenas have a social system as complex as those of many
primates. Males leave the clan at puberty, while females stay on to form
the group’s social core. All adult females are dominant to all adult males,
and the daughters of high-ranking mothers grow up to become dominant themselves.
Females are not only more aggressive than the males but look just like them:
their enlarged clitorises are the same size as the male penis and are fully
erectile. The vaginal labia are fused to form a ‘pseudoscrotum’; the female
mates and gives birth through the clitoris.

When I started my research, this similarity made it extremely difficult
to determine the sex of younger animals. I soon realised that the testicles
of yearling males formed larger bulges in the scrotum than do the fat and
connective tissues that fill female pseudoscrotums. But I still couldn’t
sex cubs that were less that a year old. In the wild, many youngsters died
before they were old enough to be sexed.

Another obstacle is that the females deposit their newborn cubs in abandoned
aardvark burrows, where they are safe from predators – and out of sight
of the prying eyes of field zoologists. The cubs live in these burrows,
where not even the mother can enter, until they are two to four weeks old,
when their mother brings them to the clan’s creche at the communal den.

As a result of these difficulties, after four years of study I knew
the sexes of only four twin litters. All were male-female pairs. There were
no twins of the same sex, even though on average one would expect to see
twin males, mixed sex twins and twin females in the ratio of 1:2:1. The
observation was not significant statistically, but it set me wondering wheth-er
it was more than mere coincidence.

The answer came from research that was begun for an entirely different
purpose. In 1984, I joined forces with psychologist Steve Glickman to begin
breeding spotted hyenas in captivity at the University of California at
Berkeley. The project was supported by a grant from the National Institute
of Mental Health, which hoped it would throw light on the links between
aggression and hormones. We wanted to see if the female hyena’s anatomy
and aggression were caused by unusually high levels of androgens, the male
hormones that stimulate mammals to develop masculine characteristics.

With the permission of Kenya’s Department of Wildlife Conservation and
Management, and with help from several Masai friends, we collected infant
hyenas from dens outside the Mara reserve. The very youngest cubs were remarkably
aggressive. One called Talek refused to take his bottle until he had attacked
his smaller brother. More disturbing still was the plight of an emaciated
10-day-old called Rabbit, who was suffering from seriously infected wounds
inflicted by her fat sister, Owl. Clearly, Owl had prevented Rabbit from
nursing for many days, and Rabbit was near death when we found them. She
was saved only by constant nursing by Glickman’s wife, Krista.

Sibling rivalry

The first thing these observations revealed was the likely cause of
death of hyena cubs in the wild. Because the cubs are confined to a narrow
burrow which the mother cannot enter, the dominant sibling can prevent the
subordinate from getting to the entrance, where it could be nursed. The
result is starvation, complicated by infection of the wounds that develop
from repeated attacks.

Handling these tiny charges led us to discover something else that was
to revolutionise our field work – a difference in the shape of the end of
the erect penis and clitoris which, despite all the years spent watching
spotted hyenas, had somehow been overlooked. Young cubs have erections whenever
they interact with older animals at the communal den, so we could sex infants
for the first time.

Information on the sex of wild cubs now accumulated rapidly. In 1988,
zoologists Kay Holekamp and Laura Smale, then at the California Academy
of Sciences and now at the Michigan State University, started studying the
animals in the Talek clan to discover how youngsters acquire their mother’s
rank. By 1990 we had sex data on 33 litters. Of these, 28 were mixed sex
and 5 were twin males. In the case of the single cubs, both sexes were equally
likely when born to low-ranking females. But high-ranking females produced
an excess of single male cubs. Why the absence of twin females in the wild?
Why were twin males so rare? And why were half the litters reduced to singletons?

Vital clues have come from our captive breeding programme at Berkeley.
By 1987, the animals we had caught in the wild had reached breeding age,
so for the first time we could observe spotted hyena twins being born, and
monitor their subsequent behaviour. So far we have found no reason to think
that prenatal events bias sex ratios: the hyenas have produced litters of
mixed and single-sex twins in the expected proportions. For both mixed and
single-sex twins, fighting is most intense on the day of birth, falling
away rapidly after one cub (usually the female in mixed sex twins) gains
dominance in the first day or two of life. Later, aggression begins to grade
into play. It could be that play replaces aggression earlier when the cubs
are of opposite sex. If so, that may explain why the subordinate survives
in mixed-sex litters but rarely in same sex twins.

These observations do not explain the evolutionary role of this behaviour,
or indeed why the mixed-sex litters are less aggressive. Could hunger play
a part? While spotted hyenas are the only mammals known to practise neonatal
siblicide, the behaviour is quite common among some large predatory birds
– eagles, herons and egrets – where it evolved for reasons to do with the
food supply. If food is scarce, the dominant fledglings kill their subordinates;
if food is plentiful, all survive. Better to have a few strong fledglings
than a nestful of weaklings.

But the reasons for siblicide in the hyenas cannot be so simple. Access
to food may be one factor. In the nearby Serengeti, cubs sometimes starve
while their mothers travel long distances for food, according to field studies
by Heribert Hofer and Marion East of the Max Planck Institute for Behavioural
Physiology, near Munich. Even in the Mara, where prey is abundant, singleton
cubs grow faster than members of twin litters. But if food is limiting,
the low-ranking females should produce more singletons, not the better-fed
high-ranking females.

Food supply and growth rate cannot be the only factors; if they were,
all dominant cubs would be expected to kill their siblings, regardless of
their sex. In reality, an opposing biological imperative comes into play.
In all animals, evolution favours behaviour that balances the individual’s
welfare with the welfare of close relatives, who carry many of the same
genes. From this point of view, killing off a sibling always carries a
penalty, and is only worthwhile if there is a very large gain to oneself.
In the case of the hyenas, the gain from eliminating a sibling of the same
sex appears to be greater than the gain from killing one of the opposite
sex.

One possible reason, for females at least, may be intense competition
for rank. The top-ranked female is eventually replaced by her daughter
at the top, and high rank is strongly linked with more successful reproduction.
It could be that twin sisters seek to eliminate each other as infants to
avoid going on to compete intensely as adults. Males, who leave the clan
young, do not pose such a threat, and so can be spared.

But there is a snag with this theory. Field studies show that while
close female kin do indeed compete in adult life, they will also join ranks
to compete against other families. Indeed, coalitions among close female
relatives appear to help maintain the remarkably stable relationships characteristic
of a hyena clan.

Another part of the answer may lie in a different aspect of hyena social
life, according to a theory developed some years ago by Robert Trivers and
D. E. Willard, who were then at Harvard University. For many animals, males
compete more intensely for mates than do females. Even poor-quality females
usually have a chance to breed, but only the better males produce many
offspring. Field studies of mammals show that larger, better-fed females
produce higher-quality offspring and that physical quality often results
from high social rank. So, according to Trivers and Willard, high-quality
or high-ranking females should produce more sons, while low-quality or lower
ranking females should concentrate instead on daughters.

My observations of Talek’s clan appear to fit this profile. High-ranking
males monopolise mating, and most top-ranked females produce a large excess
of singleton sons. The lower ranking females, on the other hand, produce
more equal numbers of male and female cubs.

For the high ranking females, single sons hold two big advantages.
In the long term, they are more likely to sire many grandchildren. And in
the short term, they seem to be less costly to rear: a high-ranking wild
female will produce her next litter much sooner after having a singleton
son than after having a singleton daughter (mixed twins are intermediate).
Holekamp and Smale have been following the fates of males as they mature
and disperse to join new clans. We hope eventually to have enough data on
their mating success to test the hypothesis.

We already know that some females produce statistically unlikely types
of litters, and there is no evidence these are the result of events in the
womb. But there is some evidence that a hyena mother intervenes in fights
between her offspring to tailor the sex ratio of her litter according to
her own rank. Late last year, for instance, graduate student Paul Barber
and I detected such evidence in a quite unexpected finding – a female in
the Talek clan, named Deejay, with triplets. One of these cubs, the dominant
sibling, was unscathed, while the second bore the scarred shoulders typical
of a subordinate cub. The smallest was terribly injured, the skin virtually
stripped from its back by its siblings. We saw Deejay alternating between
two den holes, in one of which she seemed to have stowed the runt. Once
we saw all three cubs together. Deejay was persistently snapping at the
dominant cub, trying to prevent it from attacking the injured runt as it
tried to nurse.

But the most striking evidence that females manipulate the outcome of
fighting has come from the Mara. In 1990, herdsmen poisoned an entire clan
which had been living just outside the reserve. Over the following months,
Holekamp and Smale watched as several mid-ranking females from Talek’s
clan moved out of their traditional home range to start a new clan in the
vacant territory. As a result, Talek’s clan was reduced to about 15 breeding
females, two-thirds its normal size. Suddenly, the highest-ranking females
stopped producing singleton males and began to produce more daughters than
sons so that over the past three years we have witnessed something never
seen before in any hyena clan: the survival of five female-female litters.

Sexual favours

Why the sudden switch? The most likely reason is that the females are
trying to build up the maternal core of the clan as quickly as possible
– and at a time when the competition for food is low because of the sudden
drop in numbers. For the time being at least, daughters are more valuable
than sons.

I am not suggesting that fighting in the cradle evolved originally to
allow mothers to tailor the sex of their offspring. It is more likely that
siblicide, like the hyenas’ bizarre female genitalia, originated as a side
effect of masculinising hormones. These, in turn, may have evolved in the
females because they made them more aggressive as adults.

This leaves the question of why female hyenas need to be so aggressive.
Cubs would fare poorly if they had to fend for themselves. My early studies
showed that cubs of high-ranking females were able to eat their fill under
the protection of their mothers; low ranking cubs often go hungry. Female
aggressiveness may have become important when the spotted hyena evolved
as an active predator, hunting communally, from a scavenging ancestor. Embryos
that are subjected to high levels of androgens will go on to develop into
aggressive adults that are better able to protect their cubs. These same
androgens lead to aggressive newborns, followed by siblicide and, finally,
the tendency of mothers to exploit siblicide to adjust sex ratios.

If this reasoning is correct, the female spotted hyenas are capitalising
on an extreme form of infant behaviour that is largely beyond their control.
It offers a vivid example of how violence and subtlety can work hand in
hand in evolution.

Laurence Frank is research associate at the Field Station for Behavioral
Research, University of California at Berkeley.

* * *

Life before birth

What makes newborn hyenas so aggressive? The answer may lie with testosterone.
In the womb, hyenas are exposed to unusually high levels of testosterone
in the second half of gestation, according to studies of captive hyenas
by Paul Licht, Steven Glickman and colleagues at the University of California
at Berkeley and others at the University of California, San Francisco. The
masculinising hormone comes from the mother.

All mammalian ovaries produce the hormone andro-stenedione, which is
converted by enzymes in the ovary into oestrogen. Pregnant hyenas produce
an excess of androstenedione, which is converted in the placenta to testosterone
and sent to the fetus. Levels of testosterone measured in hyena fetuses
are comparable to the highest levels detected in adult male hyenas.

Furthermore, this excess of testosterone can stimulate the fetus’s genitalia
and brain to develop masculine characteristics. Tamer Yalcinkaya and Pentti
Siiteri at the University of California, San Francisco, have discovered
that a hyena fetus, unlike its mother, produces only low levels of a protein
called steroid binding globulin, which normally mops up circulating androgens
and so reduces their ability to masculinise the genitalia and brain.

At birth, cubs of both sexes have high levels of androgens, with males
higher in testosterone and females in andro-stenedione. These fall after
birth, in line with the drop in aggressiveness over the first weeks of life.
Female hyenas have high levels of androstenedione throughout life, which
possibly contributes to their aggressive nature. Interestingly, female hyenas
behave normally in reproduction and are excellent mothers, showing that
the essential ‘female’ parts of the brain are protected from the androgens
that masculinise their aggressive behaviour.

Aggression is difficult to study because it always takes place in a
social context. For instance, simply increasing levels of androgens does
not necessarily make an individual more dominant. Indeed, it may result
in a loss of rank by causing him to pick fights he cannot win. Neonatal
siblicide is a uniquely ‘pure’ form of aggression because it occurs in the
absence of any social history. It therefore offers an opportunity to study
the influence of hormones independently of the influence of social complexities.