To watch my recent seminar at the California Academy of Sciences, click here.
Case study 1: Adaptive radiation on empirical fitness landscapes in Caribbean pupfishes
During my dissertation research, I noticed a remarkable pattern of adaptation in Cyprinodon pupfishes: despite their wide distribution from Massachusetts to Venezuela, adaptive radiation has occurred in only two places: the tiny Bahamian island of San Salvador and an isolated Yucatan lake, Laguna Chichancanab. In each of these locations, multiple coexisting pupfish species with unique and specialized feeding strategies have evolved from a generalist common ancestor within the past 10,000 years. For example, I documented the rapid evolution of a predatory species which feeds on the scales of other pupfishes, a novel niche separated by 168 million years from the most closely related species with convergent scale-eating ecology.
I found that ecological novelty in these two adaptive radiations is driving exceptional rates of morphological diversification relative to background rates in allopatric species. I estimated a time-calibrated phylogeny for Cyprinodontidae and compared the likelihood of Brownian motion models with single or multiple rates of morphological diversification for 16 functional traits. Morphological diversification rates within the two sympatric radiations containing specialists were exceptional outliers, reaching up to 130 times faster than background rates. These high rates were not due only to clade age, but instead to the presence of ecological novelty.
Furthermore, I found that different ecological niches within this radiation are driving ecological speciation at different rates. I used genotyping by sequencing to generate a dataset of over 13,000 SNPs and compared patterns of genetic differentiation between trophic specialists. This study suggests an unexpected connection between the rarity of a niche and an accelerated speciation rate, perhaps driven by larger fitness valleys separating rare ecological specialists from their sister species on neighboring peaks (Martin and Feinstein, unpublished data).
I next asked what mechanisms drive such extreme rates of morphological diversification and ecological novelty by measuring the fitness landscape for pupfishes on San Salvador. To test the role of competition, I measured the growth and survival of 1,865 F2 hybrids of all three species placed in low- and high-density field enclosures in two lakes on San Salvador where all three species coexist. I found multiple coexisting fitness peaks driven by increased competition at high densities, supporting the early burst model of adaptive radiation. Hybrids resembling the abundant generalist species were isolated on a local fitness peak separated by a valley from a higher fitness peak corresponding to the phenotype of the hard-shelled prey specialist. This study is, so far, the most comprehensive estimate of the adaptive landscape across multiple traits and species. It also provides the first experimental evidence for one of the key predictions of adaptive radiation theory: multiple coexisting fitness peaks should correspond to species observed in the wild. This complex adaptive landscape could also explain the rarity of trophic specialists across many similar environments in the Caribbean if there is widespread stabilizing selection on generalist populations.
Case Study 2: Measuring the parameters for sympatric speciation
Cameroon crater lake cichlids are the flagship example of sympatric speciation in nature. These dramatic examples of in situ speciation in uniform crater lakes inspired theoretical models exploring the parameter ranges necessary for this controversial process to occur. However, these models have never been tested in a putative example of sympatric speciation in nature.
I conducted the first explicit tests of theoretical predictions from sympatric speciation models in two of the Cameroon cichlid radiations. First, I measured the strength of disruptive selection on trophic morphology using scale growth rings as a proxy for growth rate. Unexpectedly, my first discovery was that ecological and phenotypic divergence was incomplete within young species complexes in these lakes, lagging behind highly divergent sexual coloration. Many different speciation models predict that speciation can become stalled if the strength of disruptive selection and assortative mating are not sufficiently strong. Indeed, I found significant evidence of disruptive selection in both species complexes, but it was weak relative to model predictions and Kingsolver’s classic survey. Additional measurements of mated pairs in one species complex supported strong assortative mating by diet, color, size, and trophic morphology. Despite such strong assortative mating, speciation in both species complexes appears to be stalled due to weak disruptive selection, whereas other species in these same radiations exhibit complete phenotypic divergence. Combined, these studies provide the first evidence for the dependence of sympatric speciation on both strong assortative mating and strong disruptive selection as predicted by theory in a compelling example in nature.