From pattern to process: ecology and evolution of host specificity in the fig-pollinator mutualism.

Abstract

One of the greatest challenges in the study of coevolution, indeed, for biology in general, is to understand how evolutionary and ecological processes shape patterns in nature. Ecologists routinely observe patterns of association among organisms, such as parasites infecting hosts or insects pollinating flowers and systematists routinely infer patterns of phylogenetic relationship. Such patterns invite explanation and suggest hypotheses about the evolutionary process, but it is difficult to investigate contemporary processes, including natural selection and, of course, impossible to directly observe historical processes. Observation of patterns in various ecological contexts, inference of phylogenetic patterns, model and simulation of processes, and direct experimentation aim to test specific predictions about the role of ecology in shaping evolutionary trajectories, and evolutionary processes in shaping ecological associations. The fig-wasp pollinator mutualism provides a unique opportunity to examine fundamental processes of coevolution, namely, reciprocal adaptation where interacting partners are the agents of selection. Because pollinating wasp reproduction is directly linked to host plant reproduction, it is possible to estimate the fitness consequences of interaction for both partners simultaneously. By manipulation of interacting individuals and species, or by examination of natural variation within and among populations, it may be possible to estimate the strength and direction of selection on each mutualistic partner. This work employs molecular genetic patterns, ecological observations, and direct experimentation to investigate host specificity in Ceratosolen (Agaonidae, Hymenoptera) pollinators of Ficus subgenus Sycomorus (Moraceae) and potential processes affecting the origin and evolution of species diversity in this system. The first chapter examines genetic variation in Ceratosolen pollinators of widespread Ficus across the geographic range of several host species. Deep mitochondrial DNA sequence divergence between host-specific populations distributed across Wallacea suggests host conservatism during ancient range expansion and subsequent isolation by distance. Geographic patterns of sequence divergence and host association are more consistent with a model of allopatric speciation than speciation by host switching. The second chapter investigates pollinator host choice by morphotyping and DNA barcoding of floral visitors in a community of closely related and sympatric fig species. Host specificity was very high, but rare pollinator sharing among sympatric fig species was observed at a rate of 1-2%. Even such rare events could be evolutionarily significant and pose challenges for species delimitation. The third chapter examines fitness consequences of pollinator sharing by experiment. A new method of manipulating fig pollinators investigated the reproductive consequences of intra- and interspecific pollinator visitation for both mutualistic partners. When pollinators were introduced to a novel host species, hybrid seed set was comparable to results of conspecific crosses. Hybrids germinated, established, and grew at rates comparable to non-hybrids. Pollinator fitness, however, was compromised after oviposition in the novel host. Although heterospecific pollinators induced gall formation, offspring did not develop to maturity in the new host. Microsatellite genotypes of a New Guinea fig community indicated a substantial but not absolute barrier to gene flow among sympatric species. That hybrids constituted fewer than 2% of individuals in populations may be explained by selection against pollinator host switching in this system. Collectively, these studies suggest that the extreme species-specificity of associations between Ceratosolen pollinators and Sycomorus figs is maintained by the fitness cost of colonizing new hosts. At the same time, hybridization resulting from rare instances of pollinator sharing in even the most extremely specialized of pollination mutualisms has the potential to influence diversification and coevolution.