Credit: Charles Linkem

A patch of forest next to a shopping mall is just one example of forest fragmentation, or the division of forest habitat by human residential communities, farmland, or other developed areas. As forests become more and more fragmented, biodiversity is lost and studying the biodiversity that remains grows more urgent. Matt Fujita and Adam Leache, both former graduate students in UC Berkeley’s Department of Integrative Biology, decided to study Hemidactylus fasciatus, a gecko species distributed over a fragmented forest habitat. They found that what was originally thought to be one species is actually four. The pair sequenced DNA regions from 51 geckos representing ten populations from rainforest fragments in Equatorial Guinea, Gabon, Ghana, Nigeria, Cameroon, and Congo. Using a statistical approach called Bayesian species delimitation, they then assessed the number of species present in the sample. This method first poses a “guide tree” model of possible speciation events, and then evaluates the probability of each node in the guide tree (the probability of each speciation event) by taking into account multiple factors including genetic sequence data, population size, and divergence times. Although traditional methods of classifying species are based on morphology, this statistical approach is valuable because it provides an objective and quantitative assessment. Indeed, Fujita and Leache hope their study will spark more discussion about using this method to evaluate visually indistinguishable diversity revealed by genetic studies. For now, their study is encouraging because it indicates that there may be more diversity in forests than previously expected. As Fujita says, “There is a lot more to discover, and we need to do that soon, before it is too late and the forests are gone.”

This article is part of the Spring 2010 issue.