The Family Lives of Elephants

In 1961, Irven Buss published a scientific paper in which he proposed that the fundamental social unit among African elephants could best be described as a family group. While this put the subject on a firm scientific footing for the first time, he was hardly the first person to have noticed. Even white Europeans had at least mentioned the idea going back to the 19th century, particularly with regard to the Asian elephants living in the British Raj. 

Not to mention that the locals had surely noticed, and just not written scientific papers about it.

Nonetheless, Buss, like the big game hunters before him, assumed that these families were led by the largest and strongest males. That must have seemed natural enough, and would fit with the behaviour many other mammals, but even in the 1960s, studies on elephant culls in Uganda began to reveal that the real leaders were the oldest and most experienced females. Maternity testing since then has confirmed the extent to which the families are, like those of humans, composed of very closely related individuals, and that these are organised in at least two levels of higher social grouping, each less related than the one below it.

The oldest female in the group, the matriarch, is of pivotal importance to the family she leads. Studies have shown that the older the matriarch, the more infants the group is able to raise, and the more other families will want to associate with them. The loss of a matriarch can temporarily weaken a family, and, in severe cases, lead to a localised social collapse with potentially dire consequences for all involved.

This combination of regular family life and a matriarchy is relatively rare among mammals. There are, of course, many pack and herd animals that associate together in large groups, and the females may often be the ones that lead them towards new food sources when they move. But true matriarchies, where the oldest female forms the linchpin of the closely related social group, are found only among elephants, cetaceans, and two species of primates: bonobos and (some) humans.

There are three living species of elephants: the Asian elephant, the African bush elephant, and the African forest elephant; the latter not being identified as a distinct species until 2000. So far as we can tell, all three display similar social organisation, and, for what it's worth, this implies that much the same was probably true of their closest extinct relatives, such as the woolly mammoth.

The bush elephants (Loxodonta africana) of the Amboseli National Park, in southern Kenya have been studied since the early 1970s, providing us with the basis for an unusually long-term population study. While bush elephants (like the other two species) are formally listed as an endangered species globally, they have been doing relatively well in this particular area. When the study started in 1972, researchers estimated the total population as around 600, but in 2025, we knew of 2,000 elephants in the area. This is due to a combination of high tourist activity, active game wardens (fewer elephants means less tourist dollars), and, to some extent, protection by the local Maasai people.

Crucially, every elephant that has lived nearby during that time has been catalogued as part of the ongoing research project. We obviously don't have genetic data on all of them, but what we do have confirms that the maternity records based on simple observation are accurate. Paternity, as one might expect, has proved much harder to establish. But, bearing that obvious limitation in mind, a study published earlier this year gathered together many of these old records, supported by genetic data, to give a detailed insight into the family lives of elephants in Amboseli over the last 50 years.

In total, 84 families of elephants have been identified over the years at Amboseli, 67 of which are still around. Some are small, with as few as five members, but many are much larger, reaching up to 60 individuals. They are, from a human perspective, extended families, rather than a direct lineage of grandmother to mother to daughter. In fact, the average relatedness between any two members of a given family is that of aunt to niece, suggesting that each is most likely to be started by a core group of sisters.

Obviously, given the time involved, none of the matriarchs at the beginning of the study are still with us. Indeed, each family had, on average, four matriarchs over the 50-year period. Typically, female elephants become matriarchs at around 39 years of age, and very few manage it before of 20 - those that do are typically orphans, with no related older females around to fall back on. The length of time for which they hold the position is hard to evaluate directly, because we have to exclude all of those who were either already in place in 1972, or still were as of 2025. But the best guess from the study is an average of about 19 years, which is not insignificant, even for an elephant.

On fifteen occasions, new families arose when an existing one split into two. This is, perhaps, an inevitable consequence of the steadily increasing population of elephants locally over the decades, with some of the groups becoming too unwieldy to continue. However, they were also more common when times were harsh, such as during a drought. At these times, perhaps, splitting up into smaller groups that could forage separately without exhausting the same stock of food made sense.

One might expect that such splits would tend to be fairly even, but more often than not, the new group was much smaller than the one continuing under the existing matriarch. In a typical split, for example, only a third of the female elephants would lead with their new matriarch. Such splits occurred along family lines themselves, with those most distantly related to the old matriarch likely to leave to follow a closer relative. This would suggest a deep level of bonding among the animals, with them aware, on some level, of which members of their family they share a deeper kinship with, perhaps reinforced by their behaviour towards one another down the years.

But that's not a universal rule. On occasion, a female from one family left to join up with another one. This sort of event, while rare, is still not entirely random, since the departing elephant will at least pick a new family she has some relationship with, but it does illustrate that something - social compatibility, we might say - can override family loyalty. 

The authors point out that it doesn't follow that the departing members didn't keep in touch with one another. We know that elephants can communicate over long distances, using a mix of scent marking, and calls that can even be transmitted subsonically through the ground. 

There are also instances where an entirely new family group, not especially related to one another, forms when two or more orphans gather together for companionship, with one (in one case, as little as eight years old) becoming a matriarch by default. This is different to what happens when families split, where the new matriarch is usually about the same age as the existing one, a sister or close cousin who has similar experience and now gets the opportunity to lead on her own.

What all of this suggests is that elephants are making decisions based on relatively sophisticated information. Whether they leave or join a group is not a random matter, based on who happens to be closest, but an active choice. New matriarchs have to display certain leadership qualities, and, at least according to this study, there are times when purely social considerations of who you want to spend your time with, and who you want to rely on to lead, can overrule kinship alone.

[Photo by Bernard DuPont, from Wikimedia Commons.]