Biologists from Washington University in St. Louis collaborated with a large number of butterfly and plant specialists to reconstruct the origin and global spread of butterflies.
Working with researchers from dozens of countries, Michael Landis and Mariana P. Braga in Arts & Sciences helped create the world's largest butterfly tree of life, assembled with DNA from more than 2,000 species representing all butterfly families.
Using this framework as a guide, the researchers traced the movements and feeding habits of butterflies through time in a four-dimensional puzzle that led back to North and Central America. According to their results, published May 15 in the journal Nature Ecology and Evolution, this is where the first butterflies took flight. The study was led by Akito Kawahara, curator of lepidoptera at the Florida Museum of Natural History.
As part of the research team, the Washington University scientists were particularly focused on finding the best way to estimate which host plants were devoured by the first butterfly caterpillars and their descendants.
"The study is fundamentally at a global scale," said Landis, assistant professor of biology. "Butterflies occur almost everywhere across the continents, but it was still unclear when and where major butterfly genera first originated, when and how they dispersed among landmasses and when they first gained the ability to munch on different groups of host plants as larval caterpillars."
There are some 19,000 butterfly species. Piecing together the 100 million-year history of this group of insects required information about their modern distributions and associated host plants. Prior to this study, there was no single resource that researchers could query to access that type of data.
"The challenge was to deal with the large number of plants that are eaten by at least one of the many butterfly species included in this study," said Braga, who worked in Landis' lab at Washington University and now is a postdoctoral researcher at the Swedish University of Agricultural Sciences. "This number is much larger than what the latest methods are able to process in a reasonable amount of time.
"My solution was based on my previous work at WashU," Braga said. "We published a study that shows that the interactions between butterflies and host plants often evolve as groups of interactions and not in isolation. Thus, it was reasonable to first identify these groups and then estimate how old each group was.
"That's how we came to the result that the group containing bean plants is the most likely ancestral host of all butterflies," she said.
This study includes a very large number of species. More importantly, it includes 92% of all butterfly genera, which means it paints a nearly complete picture of all butterfly lineages, the scientists noted.
"Seeing how ecological relationships among butterflies and plants ebbed and flowed as they co-evolved reminds us that ecosystems are always in flux," Landis said.
