Hummingbird Studies Raise Questions About Birdsong Evolution

According to Duke University Medical Center neurobiologist Erich Jarvis, the paper's lead author, while most of the 23 orders of birds can vocalize, like the rooster that crows, these vocalizations are not learned but are genetically hardwired sounds. Only three orders of birds the songbirds, parrots and hummingbirds have the ability to learn songs from their adult tutors and repeat them in the right context, said Jarvis. This type of vocal learning is very similar to the way that humans learn to speak.

Surprisingly, these singing birds are not closely related to each other, and in fact have close relatives that cannot learn song, he said. Despite their evolutionary distance, the new research indicates that hummingbirds use the same seven structures in the brain that parrots and songbirds use when they are singing their learned vocalizations structures that aren't even present in non- vocal learning orders of birds.

The finding raises the evolutionary question of whether the three orders of birds developed the ability to learn song independently, and each time developed similar brain structures to serve this purpose. Alternately, there may have been a common ancestor with the ability to learn song, and only a few of the descendants retained this ability along with the specialized brain regions.

Jarvis said that all the evidence supports the former explanation (that vocal learning developed independently three times) and points to another example of independent evolution the similar development of wings from limb structures in pterosaurs (ancient flying dinosaurs), bats and birds.

"The reason why wings evolved in a similar way is because there is an environmental constraint, the center of gravity, placed on how animals can fly," he explained. "Here, I think there is an interaction between the environment and the brain, and Mother Nature has a basic constraint, even instructions, on how you can develop brain structures for a complex behavior such as vocal learning."

In applying the bird findings to mammals, Jarvis pointed out that, like the bird family, mammals have only a few members who demonstrate learned vocalizations humans, bats, whales and dolphins. The evolutionary implications of the bird research may impact the study of how and why these few mammals have developed the ability to learn speech or sing.

Jarvis and the other lead collaborator, Claudio Mello of Rockefeller University, gathered much of their data for this study in Espírito Santo, Brazil, where a living museum and nature reserve of the Atlantic Tropical Forest dedicated to hummingbirds was established. They observed birds in their natural state, using binoculars, video cameras and sensitive audio recorders to record natural hummingbird behavior and song complexity. A few of the birds were captured and their brains were studied to define the regions that become activated while singing.

Collaborators on the study were Sidarta Ribeiro from Rockefeller University and Maria Luisa da Silva, Dora Ventura and Jaques Vielliard from the University of Sa Paulo in Brazil.