An all-black Chestnut-bellied Flycatcher found on small satellite islands to the north and southeast of the island of Makira in the Solomon Island chain.

Same traits, different island: study highlights evolution quirks

Two populations of flycatchers that evolved on different remote islands separately developed the same trait – all-black feathers – according to a new study that used machine learning to understand the process that shaped the birds’ genome.

“The chestnut-bellied flycatcher is not as well-known as Darwin’s finches,” said Leonardo Campagna, an evolutionary geneticist at the Cornell Lab of Ornithology and lead author of “Selective Sweeps on Different Pigmentation Genes Mediate Convergent Evolution of Island Melanism in Two Incipient Bird Species,” which published Nov. 1 in PLOS Genetics.

“But this complex of birds has also gone through many evolutionary changes, many of which involve changes in the coloration and patterning of their plumage,” he said.

Originally, a large population of chestnut-bellied birds lived on one of the bigger Solomon Islands, in the Pacific Ocean. From there, some birds established new populations on a couple of smaller islands. Over time, birds on the two smaller islands lost their chestnut bellies and became all black. But the birds on each island developed black plumage at different times, from different genetic mutations which moved rapidly through the small island populations. One of these mutations spread during the last 1,000 years – a mere blink in evolutionary time.

“Clearly there’s something advantageous about having all-black plumage,” Campagna said. “We’ve traced this trait back through time by sequencing the entire chestnut-bellied flycatcher genome for the first time. The two mutations that lead to black plumage appeared at different times, on different islands, and on different genes related to melanin pigment production. That level of convergence is wild."

The various flycatcher populations are in the early stages of speciation – splitting off to form new species – but they have not yet diverged much genetically and they can interbreed. But they rarely do, producing a few hybrids. Field experiments have shown the chestnut-bellied birds and the all-black birds each react aggressively toward a perceived interloper with their own plumage color, but do not respond the same way to the members of their species with a different color.

And it turns out the flycatcher genome is still evolving. “We’re finding there’s a third melanic (all black) population of flycatchers among islands about 300 miles away from the original island,” said senior co-author Al Uy, a biology professor at the University of Rochester. “The mutation governing their plumage color is different yet again from those on the other two islands we studied.”

Uy has been studying the Solomon Islands flycatchers for about 15 years, aided by a group of Indigenous islanders. “I think the emerging pattern is that there’s something about small islands that's favoring these all-black birds – in the more distant archipelago were melanism has evolved for the third time, we found that melanic and chestnut-bellied birds still coexist within each island but as islands get smaller, the frequency of melanic birds goes up.”

There are multiple theories about what's driving the switch to back plumage, including female preference, the greater durability of black feathers, and even a possible link to genes that govern other advantageous behaviors.

The study authors include computer scientists Ziyi Mo and Adam Siepel from Cold Spring Harbor Laboratory who wrote the machine learning program that helped the researchers dig deeper into the past and measure mutation patterns in the flycatcher family tree.

“The use of machine learning is an exciting new development in the field of population genetics,” Campagna said. “We train the computer to recognize specific evolutionary patterns for when a particular genetic trait started, how strong natural or sexual selection was, and how quickly it moved through a population. We can then ask the trained algorithm to tell us the most likely scenario that generated the data that we observe in the present populations. It's like going back in time.”

Pat Leonard is a writer for the Cornell Lab of Ornithology.

Media Contact

Becka Bowyer