Researchers from the University of Sussex found out that African elephants can recognize the voice of different tribal groups and use this information to react to their surroundings.

The scientists also discovered that the elephants can identify the voice of a man, woman, or child. This auditory information helps them delineate whether the nearby humans are a threat or not.

The researchers staged an experiment inside the Kenya Amboseli National Park where 1,500 elephants are located. The park is home to two tribal groups, a cattle-herding tribe called Maasai used to attack the elephants, and the Kamba tribe, which are considered as no threat to the elephants.

They used loudspeakers to broadcast the voice of a Maasai man saying, "Look, look over there, a group of elephants coming." When the elephants heard this, they huddled closer together and sniffed the air in anticipation for an attack. By contrast, when they heard the voice of a Kamba man saying the same sentece, they were relaxed. The elephants were also calm when they heard the voice of Maasai boys and women, who did not participate in the attacks.

The next part of the test involved broadcasting a Maasai man's voice mimicking the voice of a woman. The reaction of the elephants was the same as when they hear the voice of a Maasai man, showing that they use gender signals more efficiently than human beings do.

"They're using vocal information from another species - us - and they're using that to discern threat," behavioral ecologist at Britain's University of Sussex and study co-author Graeme Shannon told USA Today. "That takes really advanced cognitive abilities. ... These are subtle differences these elephants are attending to."

According to a 2011 research by Shannon and Karen McComb, also from the University of Sussex, the oldest elephants can also distinguish the roar of a male lion from the roar of the female. They can also acknowledge the roar of other elephant groups, including groups that they haven't interacted with for many years.

This study was published in the issue of Proceedings of the National Academy of Sciences.