Insights into an Evolutionary Radiation : Causes And Consequences of Diversification in the Western Ghats Bush Frogs

Abstract

One of the biggest attractions of the tropics is its stunning diversity; the origin and determinants of this diversity are amongst the most important questions in evolution and biogeography. There are two ways of looking at diversity: diversity as observed in clades on the tree of life and diversity as distributed in space. Various processes drive patterns in clades leading to a gradient of evolutionary radiations. These radiations are amongst the major contributors to Earth's biological diversity and act as good model systems to test the relative significance of processes governing macroevolutionary patterns of diversification. My research presented here revolves around one such evolutionary radiation of frogs that we uncovered from the Western Ghats, a tropical mountain system in Peninsular India. The broad aim of my thesis is to understand the causes and consequences of diversification, the primary process that drives evolutionary radiations and diversity differences in clades. The ultimate goal is to use a model system in a regional setting to explore diversification process¬es in clades against the backdrop of existing theory and ultimately paint a broader picture to further our understanding of diversity patterns in space. I have organized this thesis into six chapters. The first chapter serves as an introduction to the concepts, models and problems addressed, while chapters two to five form the crux of this thesis, and the final chapter pro¬vides a synthesis of results and directions to carry this research forward. In the first chapter, I have provided a conceptual link between diversity in Glades, evolutionary radiations and diversification and its drivers, all of which form the backbone for the following chapters. I review the potential historical processes that acted independently or in combination to give rise to one of the global biodiversity hotspots, the Western Ghats of the Indian sub-continent. I also introduce the model system, bush frogs, and provide a brief overview of their current taxonomic and systematic status. Finally, I have described the goal and questions addressed in this thesis. In chapter 2, I address the problem of delimiting lineages and provide a base for the remaining chapters 3, 4 SE 5, which form the core of this research. In this chapter, I address two major issues in biogeography. namely the 'Linnean shortfall' (the problem of undescribed lineages) and the 'Wallacean shortfall' (the lack of distribution data), and use an extensive sampling regime and a phylogenetic framework to delimit lineages. When I began my sampling, 21 lineages had been described from the Western Ghats. Three years of sampling across 13 major massifs has led to an increase in the number of the lineages to 70. These lineages were delimited using a hierarchical multi-criteria approach using a haplotype phylogenetic tree, genetic distance, geography and morphology. Lineages recovered in this study range from shallow to deep divergences and are spatially distributed from lowlands to

Massif summits. Further, inclusion of geography as a variable in the hierarchical approach increased efficacy in delimiting lineages. This approach was used to establish a working hypothesis of lineages for later chapters, where I address questions related to the drivers behind the diversification of these Lineages. The goal in chapter 3 was to examine the evolutionary and biogeographic processes underlying in-situ Glade diversification in the bush frog clade at broad temporal and spatial scales. First, I examined the Glade for characteristics of an evolutionary radiation. One of the fundamental challenges in characterizing evolutionary radiations is to discern independent colonization from in-situ diversification. Phylogenetic reconstructions shows that bush frog lineages in the Western Ghats belong to two clades: a larger Glade composed of 67 lineages, with their ancestral origin centered in the Western Ghats Escarpment, and a set of three lineages with their distribution in Western Ghats, but phylogenetically nested within the Sri La.nkan bush frog Glade. The larger Western Ghats clade recovered here lends strong support for in-situ radiation of bush frogs, the origin of which can be dated to Eocene - Oligocene boundary. The bush frogs are a relatively young radiation compared to recent reports of ancient lineages and potential radiations whose origins date to the Cretaceous. Within the Western Ghats, bush frogs make up the largest vertebrate radiation, and when considered with the sister clade in Sri Lanka, constitute the largest vertebrate radiation in South Asia. These clades attain greater significance because of the limited geographical space in which the whole radiation has occurred. Temporal patterns of line-age diversification in the Western Ghats bush frog Glade show strong signatures o f an early burst, considered as one of the characteristic features of evolutionary radiations, especially adaptive radiations. Considering that a similar pattern could arise from the initial formation of geographical isolates, I tested the Glade for the role of regional biogeographic barriers in the initial diversification of the Glade. Ancestral range reconstruction showed that the initial diversification of the bush frog Glade is coincident with the Palghat Gap, a prominent geographical break in the Escarpment. Two sister clades, North and South, could be discerned, with their respective centres of origin north and south of the Palghat Gap. There was limited dispersal across the gap and a number of these dispersal events resulted in diversification in the adjoining regions. These results, for the first time, lend strong evidence for the biogeographic significance of the Palghat gap for amphibian diversification. The discovery of the North and South sister-clades, apart from elucidating the significance of Palghat gap as a barrier in the initial diversification of the Glade, provided a unique setting to gain further insights into diversification processes.

using the sister clades. The South Glade shows support for the models of early burst and subsequent diversity-dependent decline, while the North Glade did not differ from a constant pure-birth model of diversification. The sister clades also showed strong differences in the number of lineages and in the ecological and morphological diversity of component lineages. In terms of the ecological diversity of the clades, the lineages in the South Glade are predominantly closed canopy dwellers and occupy all the vertical strata of the forest from understorey to canopy. In contrast, the North Glade is characterized by the absence of canopy lineages and also in the relative increase in the number of open canopy lineages. I excluded area as a causal factor, considering the relatively smaller area of the region of the South Glade, south of Palghat Gap, compared to the North Glade. Taken together, an early burst in lineage diversification, high species richness and Glade-level endemism, narrow geographical range of the Glade, restriction of canopy lineages to this Glade and persistence of a historical rainforest refugium in this area suggested a potential role for ecological opportunity in the adaptive radiation of the South Glade. The results from Chapter 3 brought out the strong contrast in the ecological diversity of extant lineages among the bush frog sister clades occurring in this region. One major ecological pattern is the predominance of closed canopy lineages in the South Blade, while the North Glade was characterized by relatively more open canopy lineages. However, what caused the origin of open canopy lineages and the predominance of such lineages in the North Glade was not known. In chapter 4, I explore whether signatures of the global climate mediated emergence of open canopy biomes can be discerned in the habitat shift and in the observed contrast in ecological disparity between the sister clades. The results show multiple habitat shifts among bush frog lineages in the Western Ghats, with a strong contrast between the sister clades; the North Glade exhibited a greater number of habitat shifts compared to the South Blade. A unique shift to the bamboo reed Ochlandra could be discerned in the South Glade. The study provided the first evidence for independent shifts to grass associated habitats among bush frogs and frogs in general. Examination of ecomorphology brought out further interesting insights. For the same body size, lineages occurring in grass-associated habitats, Oehiandra reeds and grasslands have independently evolved smaller hind limbs compared to lineages that occur in the forest and shrub-land habitats. This convergence in limb-size and a. pattern of reduced limb-size provides support for the adaptive nature of this trait in relation to grass associated habitat shifts among bush frog lineages. Examination of this morphological trait in time, using disparity through time plots, showed high disparity within the North Glade as expected from the patterns of greater habitat shift. These results lend support for orbitally forced habitat dynamics as a driver for the intra-cladal patterns of ecological disparity in bush frog sister clades.

Researchers in the past have largely tested evolutionary radiations for adapt signatures under the assumption of ecological opportunity. One of the drawbacks of temporal models of diversification is the exclusion of geography. Mountain systems slid-the -Western Ghats provide an ideal setting to explore diversification in space. In chapter took a biogeographic approach and tested the relative importance of Quaternary glaciatic ecological gradients and Earth related processes on the spatial patterns of lineage and s dada! diversification in a mountain setting. This was done using a set of a-priori predict', on the phylogenetic, spatial and temporal patterns of lineage diversification based on geological and climatic history of the Western ghats. These were tested using sister-lines and latitudinal and elevational range data obtained from field sampling across 13 mass Independent methods were used to test for the alternative processes. A majority of recer diverged bush frog sister - lineages were characterized by non-overlapping geographi ranges, and were isolated on adjacent massifs, indicating allopatric speciation eves Independent transitions along elevational gradients among sub-clades during the Mioc' lend support for diversification mediated by the uplift of the escarpment. In-situ diversificat in the montane zone of the two highest massifs shows the effect of past - climate media forest-grassland dynamics, but divergence times lend less support for the role of Quatern glaciations. These results highlight the important role of geological processes in the line; diversification of bush frogs. By examining patterns in a large radiation, I was also able bring out the temporal and spatial significance of different processes. Diversification in s Glades shows that recent non-adaptive processes have masked the effect of initial adapt radiation. Based on these findings, we recognize the massifs in this Escarpment as cent of lineage diversification and generators of diversity and the lowland and medium elevat habitats in the southern Western Ghats, with deeply divergent lineages, as museums. Through this research, I have provided a number of insights — many for the first time — into the macroevolutionary patterns and historical processes behind this large vertebrate radiation in a global biodiversity hotspot. The thesis also sets a foundation to explore the processes driving ecological and biogeographic patterns — species richness, endemism and community assembly — in space.