Date of Award


Degree Name

Doctor of Philosophy


Biological Sciences

First Advisor

Todd Barkman, Ph.D.

Second Advisor

Darin Penneys, Ph.D.

Third Advisor

Devin Bloom, Ph.D.

Fourth Advisor

Steve Kohler, Ph.D.


Biogeography, heteranthery, medinilla, melastomataceae, sonerileae, systematics


Medinilla is a large paleotropical genus in the plant family Melastomataceae with a long history of systematic uncertainty. The first objective of this research is to test the monophyly of Medinilla. A species tree is estimated using 599 nuclear gene trees, and 87 concatenated chloroplast genes are used to estimate the plastid phylogeny. Sampling includes 272 Medinilla and outgroups. Pachycentria and Plethiandra are resolved within Medinilla. Other fleshy-fruitedgenera long associated with Medinilla (i.e., Catanthera, Heteroblemma, Kendrickia) belong to other clades. With this new evolutionary framework, an infrageneric classification system is proposed for Medinilla consisting of eight sections, 12 informal alliances, and all 398 accepted species (plus four undescribed species). Each group is characterized and an identification key is provided. Four replacement names and 11 new name combinations are proposed to accommodate taxonomic issues. This is the next definitive step in the centuries-long effort toward a natural classification of one of the largest and most problematic groups in Melastomataceae.

With a better understanding of Medinilla and its phylogeny, the second objective is to investigate the biogeographic origins of the entire genus, as well as those that have colonized the Philippines. Previous studies of Medinilla hypothesized a Miocene origin in Sundaland, but the results of these analyses indicate Indo-Burma is a more likely birthplace for the genus. Nevertheless, Sundaland played an important role in the early diversification of Medinilla. From there, one major clade moved west, reaching as far as Africa. Other clades moved east, reaching as far as Polynesia. Both Sundaland and Indo-Burma have been major evolutionary hotspots of Indomalayan and Australasian biodiversity since the Miocene, and these findings further support this conclusion. The Philippines is another major center of diversity, hosting more described Medinilla species than anywhere else. The nature of this diversity was unclear. It is situated at the crossroads between two major biotas of Asia and Australia. It is also isolated by water in one of the most geologically active places on Earth. Analysis results suggest approximately 90% of Philippine Medinilla species originated from a single colonization event in the Miocene. Moreover, this clade has been a source of diversity to all of the surrounding areas, reaffirming the evolutionary importance of the Philippine islands.

The phylogenetic framework also allowed for the investigation of the evolution of a widespread floral trait: heteranthery. Heteranthery is known to facilitate the division of labor between anthers that provide a pollen reward to pollinators, and those that supply pollen for pollination. However, why this phenotype exists in some flowers and not others is poorly understood. Because of repeated observations of large flowers with heteranthery relative to smaller-flowered, non-heterantherous congeners, the hypothesis that flower size and heteranthery are related is tested. Comparative analyses show that the rate of evolution of heteranthery is positively correlated with larger flower-size in Medinilla. Therefore, size dependence helps explain why heteranthery is not more widespread. It could be related to mating system, target pollinator size, and/or another factor yet to be determined.

Access Setting

Dissertation-Open Access

Included in

Botany Commons