African elephants use names to call each other, study suggests
Adres URL jest ukryty dla gości, należy się zarejestrować i zalogować, aby zobaczyć hiperłącze
Wild African elephants may address each other using individualized calls that resemble the personal names used by humans, a new study suggests.

While dolphins are known to call one another by mimicking the signature whistle of the dolphin they want to address, and parrots have been found to address each other in a similar way, African elephants in Kenya may go a step further in identifying one another.

These elephants learn, recognize and use individualized name-like calls to address others of their kind, seemingly without using imitation, according to the study published Monday in the journal Nature Ecology and Evolution.
The most common type of elephant call is a rumble, of which there are three sub-categories. So-called contact rumbles are used to call another elephant that is far away or out of sight. Greeting rumbles are used when another elephant is within touching distance. Caregiver rumbles are used by an adolescent or adult female toward a calf she is caring for, according to the study.

The researchers looked at these three types of rumbles, using a machine-learning model to analyze recordings of 469 calls made by wild groups of females and calves in Amboseli National Park and Samburu and Buffalo Springs National Reserves between 1986 and 2022. All the elephants could be individually identified by the shape of their ears, as they had been monitored continuously for decades, according to the study.

The idea was that “if the calls contained something like a name, then you should be able to figure out who the call was addressed to just from the acoustic features of the call itself,” said lead study author Mickey Pardo, an animal behaviorist and postdoctoral fellow at Cornell University in New York.

The researchers found that the acoustic structure of calls varied depending on who the target of the call was.

The machine-learning model correctly identified the recipient of 27.5% of calls analyzed, “which may not sound like that much, but it was significantly more than what the model would have been able to do if we had just fed it random data,” Pardo told CNN.

“So that suggests that there’s something in the calls that’s allowing the model to identify who the intended receiver of the call was,” he added.