Tropical Intercontinental Disjunctions: Gondwana Breakup, Immigration from the Boreotropics, and Transoceanic Dispersal

Abstract

With the advent of molecular data, and particularly the use of molecular clocks when warranted, biogeographers have begun to reexamine the roles of Gondwana breakup, immigration from the Laurasian tropics, and transoceanic dispersal in shaping similarities among the world’s tropical floras (Givnish et al. 1999, 2000; Lavin et al. 2000, 2005; Renner et al. 2000, 2001; Azuma et al. 2001; Chanderbali et al. 2001; Renner and Meyer 2001; Conti et al. 2002; Davis et al. 2002; Jobson et al. 2003; Zhang and Renner 2003; Berry et al. 2004). As increasing numbers of continental disjunctions are analyzed, it has also become possible to ask whether there are regular patterns in the directions and timing of transoceanic dispersal events (S. S. Renner, in this issue). Wind and ocean surface currents (as well as tectonics) all affect dispersal and are not randomly distributed in space and time; they should thus leave an evolutionary trace provided that they have been stable long enough to override lineage-specific differences in dispersal and establishment capability (e.g., see Munoz et al. 2004). Tropical intercontinental disjunctions raise issues analogous to those involving disjunct temperate floras in the Northern Hemisphere (see IJPS symposium edited by Manos and Donoghue [2001]), but the ocean gulfs are often greater and the latitudinal distributions of taxa less dynamic. Similar issues also emerge in studying the origin of island lineages (Wagner and Funk 1995; Givnish 1998; Emerson 2002). Insular groups pose special problems, especially when extraordinary divergence due to adaptive radiation makes it difficult to identify mainland ancestors or relatives. Yet intercontinental disjunctions can be even more complex to decipher, given the frequent lack of asymmetries between potential source and target areas (but see T. J. Givnish et al.; K. J. Sytsma et al.), the existence of direct connections to other landmasses at various times, and the longer and more complex tectonic histories of the terrains. The papers collected here represent the product of a symposium held in Madison during August 2002, sponsored by the Association for Tropical Biology and the Association for Systematic Botany and honoring Robert Thorne for his outstanding contributions to biogeography and especially to the analysis of tropical plant disjunctions (Thorne 1972, 1973; R. Thorne’s ‘‘reflections’’ at the end of this volume provide an entry into his biogeographic contributions). We invited researchers to present analyses based on molecular data for a variety of disjunctly distributed tropical lineages that, in aggregate, display a diversity of continental disjunctions, mechanisms of dispersal, and geological age. We asked each contributor to address three questions: (1) How much information about the timing and nature of the events causing disjunctions can be obtained from molecular data? (2) What appear to be the relative roles of dispersal and vicariance, and how do these relate to events in Earth history? and (3) How can we cross-check and validate our inferences using fossils or other data? All of the contributions calibrate the lengths of branches in molecular trees (measured as the numbers of inferred nucleotide substitutions) against the apparent age of one or more ancestral taxa (inferred from the geological age of the oldest known fossil attributed to those taxa). The absolute dates of cladogenetic events within a lineage are then inferred using one or more of several calibration models and attributed to apparent dispersal or vicariance events, which have themselves been inferred from the distributions of modern-day taxa and their phylogenetic relationships to each other. The timing of range expansions is finally tied to shifts in geography, climate, or likely pathways for dispersal during geological time, and conclusions are drawn about the likely mechanisms underlying a particular disjunct distribution. The validity of such conclusions thus hinges on the completeness of our knowledge of the fossil record, both in areal extent and in time; on uncertainties in the branch lengths inferred from particular molecular markers; on the validity of chronological calibrations based on particular models and a particular set of fossil dates; and on uncertainties regarding the distributions of ancestral taxa inferred from a molecular phylogeny, distributions of modern-day taxa, and particular models for shifts in geographic range. Nevertheless, the approaches taken by papers in this symposium, involving the complementary use of molecular and paleobotanical data, offer some of the most powerful techniques now available for assessing the age—and, hence, the mechanism—associated with a shift in the range of a lineage. Without such approaches, or comprehensive and thus perhaps unattainable fossil data bearing on the timing of a range shift, one might indiscriminately invoke long-distance dispersal, continental drift, or boreotropical migration to explain essentially any tropical disjunction. The surprising conclusion from careful analyses in this symposium and elsewhere (e.g., Cronn and Wendel 2004; Lavin et al. 2005) is that long-distance dispersal may be more pervasive than all but its most ardent advocates (e.g., Iltis 1967) had previously proposed.