RELATIONSHIPS BETWEEN DISPERSAL, GEOGRAPHIC RANGE, AND DIVERSITY IN THE ROSALES: EVIDENCE FOR LINEAGE SELECTION?

The origin – and demise – of biological diversity is a question that has long interested biologists and paleontologists alike. Many recent approaches to this problem focus on systematic comparisons, although ecological comparison remains a common practice and the fossil record remains the best source of information on past diversity available. Relationships between geographic range, diversity, and taxon survivorship have been demonstrated in paleontological studies as well as in extant populations, but the mechanisms that link the ecological traits of a species with its geographic range are still not well understood. I here use the Rosales, a diverse group of flowering plants possessing a wide range of life history traits and dispersal modes, as well as a good fossil record, in order to quantify the relationships between different modes of dispersal, life history traits, and geographic range using herbarium specimens and museum collection databases. In addition to uniting modern phylogenetic methods with a paleobotanical prospective, this study also compares the three-way relationship between ecology (dispersal), macroecology (geographic range size), and lineage selection and diversification together possibly for the first time. Phylogenetic correlation is controlled for using the QuaSSE method (Maddison & Maddison), and also in order to elucidate differences between character state proliferation resultant from differential net diversification rates (lineage selection) as distinguished from selection at the individual level (character state change rates). Phylogenies used are APG III and Potter et al.’s (2007) phylogeny of the Rosaceae, with Rosaceae sequences acquired from Potter et al and other Rosales augmented via GenBank. Preliminary results using Mesquite (Maddison & Maddison 2011) and MuSSE (Fitzjohn) produce conflicting results, possibly resulting from bias in preliminary GenBank data. More rigorous analysis of a more complete dataset including over 90% of Rosalean genera using QuaSSE implemented through Diversitree (Maddison & Maddison) is ongoing, and will be presented at the meeting.