Luke J. Harmon - Research

Current research

Luke J. Harmon, University of British Columbia

1. Predictable patterns of evolutionary radiation

Does adaptive radiation proceed idiosyncratically, according to the particular starting conditions and constraints present in an evolving lineage? Or might there be repeatable patterns of radiation that can be identified in diversifying groups?

Phelsuma ornata, Mauritius

A. Community structure of diurnal arboreal island lizards

Anolis lizards in the Carribean are a classic model for evolutionary radiation. However, there is still some question as to which aspects of this system are general features of adaptive radiation, and which are idiosyncratic to anoles. Day geckos in the Indian Ocean provide a unique opportunity to address this question; distantly related but ecologically similar to anoles, these lizards have radiated on island archipelagos in the Indian Ocean. My research seeks to test the hypothesis that anoles and day geckos represent an example of repeated radiation - the repeated evolution of similar communities independently. For communities of day geckos in Mauritius, Madagascar, and the Seychelles, my results have been intriguing; I was able to predict several key aspects of day gecko community ecology from previous work on anoles. I am now extending this work to compare patterns of ecomorphological correlations between these two groups of lizards.

B. Comparative study of adaptive radiations

In a previous paper, we showed a relationship between patterns of morphological evolution and diversification among four large lizard clades: groups with elevated rates of diversification early in their history also tended to have much of their morphological disparity associated with old, deep clades in the tree. This suggests that there may be general patterns that can be found by comparing a large set of radiations. We are currently doing just that in association with NCEAS, the National Center for Ecological Analysis and Synthesis in Santa Barbara, CA. We have compiled a large data set of more than 50 radiations that includes many of the classic examples of adaptive radiation, including anoles, Galapagos finches, African cichlids, and Hawaiian spiders, among others. We are now testing several hypotheses about the dynamics of diversification using this data set.

2. The effects of diversification on ecosystems

Paxton Lake, Texada Island, BC

Much attention has been paid to the potential causes of adaptive radiations, but little is known about their effects. As organisms diversify and evolve new morphologies, however, this is likely to have interesting consequences for ecosystems. We are seeking to understand what these changes might be using the three-spined stickleback.

To do this, we have carried out a mesocosm experiment using 1000 L cattle tanks. In these tanks, we placed different combinations of sticklebacks, including intermediate, generalized forms from single species lakes as well as benthic and limnetic forms from the famous species pair lakes here in British Columbia. We then measured several aspects of these mesocosm systems, including productivity and zooplankton community structure. Any differences among our treatments may represent the ecological consequences of adaptive speciation in sticklebacks.

3. Statistical methods for analyzing diversification.

Plot of reconstructed rates of speciation
(white dots) and morphological evolution
(black dots) for Caribbean Anolis.

Several major initiatives are helping to recover more and more of the tree of life. With that tree, we will be able to test hypotheses about the tempo and mode of radiation. Phylogenies are difficult to construct partially because they are information-rich, and I believe that we still lack some of the basic tools needed to extract that information. Additionally, many of the tools that we do have are poorly characterized in terms of their power, bias, and other basic statistical properties. To that end, I am using simulations to more fully characterize the statistical properties of the methods that are currently in use. For example, in recent papers my colleagues and I have described the error rates of correlation tests using independent contrasts when species are measured with error, and investigated biased inferences about patterns of species diversification arising from using incorrect models to infer phylogenies. I am also undergoing a project to describe the statistical properties of estimates of speciation and extinction rates under a variety of conditions.

I have developed several new statistical approaches to investigate the dynamics of diversification using phylogenies. These include a method to infer the patterns of morphological disparity among clades, a test for a correlation between rates of speciation and morphological evolution in phylogenies, and improved likelihood-based methods for testing models of morphological evolution.

4. Evolutionary consequences of neutral macroecological theory

Spatial metacommunity model after 2000
generations, with different species denoted
by unique colors

Metacommunity models are proving useful in macroecology. The most famous of these models, Hubbell's Unified Neutral Theory, has generated much controversy. much of this work has focused on ecological predictions of this model, such as species-abundance curves and patterns of spatial turnover in communities. However, relatively little attention has been paid to the evolutionary aspects of this model, despite the fact that it includes speciation and thus makes clear evolutionary predictions. For example, Hubbell's model makes predictions about the shape of phylogenetic trees derived from neutral metacommunities, and the way that those trees should relate to the species-abundance distribution. I am using simulations to investigate those predictions further; the first set of results from these simulations are in press. In collaboration with a number of other researchers, we are also compiling data to test the evolutionary predictions of the UNT.