Effects of ecology on Ficus population genetic structure in southern Africa

Abstract

South African forests have become fragmented due to both historical events and current human activities. Fragmentation may have increased population differentiation by reducing the gene flow between plants isolated in fragments. The genetic effects of fragmentation on species, especially plants, have received little attention in South Africa. Fig trees are keystone species in South African forests that provide food for many frugivores. Their population genetic structure has never been studied in South Africa. In my study, the population genetic structures of three forest fig trees (Ficus bizanae, Ficus craterostoma and Ficus sur) with different habitat specificity, geographical distribution, and level of endemicity were studied to explore the ecological factors (both non-biotic and biotic) that may contribute to their population differentiation. First, I did a phylogeographic and population genetic study on the forest specialist Ficus species, F. craterostoma, from forests covering the species' range in South Africa. Chapter two reports this work. Three putative refugia were inferred, which played an important role in maintaining genetic diversity and facilitating recolonization by F. craterostoma after the LGM. In addition, comparing the genetic structures of nuclear and cytoplasmic DNA I found that nuclear DNA in F. craterostoma may be resistant to fragmentation while cytoplasmic elements are more vulnerable to fragmentation. This difference stems from the reduced effective population size of cytoplasmic DNA, but seed dispersal must also be limited whereas pollen dispersal, facilitated by a mutualistic pollinator fig wasp, is very extensive. In fact, a novel analysis that compared genetic variability with a combined forest size, suggested that forests within a 61 km radius of one another are genetically connected via the tree’s fig wasps. Differences in the biology of species can lead to differences in their responses to habitat fragmentation. Therefore, a comparison of genetic structure among species with different biologies may provide an understanding of the effect of fragmentation. I compared the population genetic structures of three Ficus species (F. bizanae, F. craterostoma and F. sur) on both large and fine scales and report on these in my third and fourth chapters respectively. At a large scale, the two forest specialists have a stronger genetic structure compared to the habitat generalist (F. sur). Within the forest specialists, the species with a limited distribution (F. bizanae) showed stronger genetic structure, indicating local pollen and seed dispersal. At the fine scale, I found significant spatial genetic structure (SGS) in all three Ficus species, which may be due to limited seed dispersal between fragmented forests. Among them, F. bizanae had the highest positive kinship coefficient within a short distance as well as the highest sp statistic. This indicates that in addition to limited seed dispersal, there may also be limited pollen dispersal in F. bizanae. An increasing SGS of F. craterostoma was detected from saplings to adults at Ingeli forest (ING), which suggests that recent population fragmentation has stopped such long-distance immigration. Non-random mortality caused by microenvironmental selection may also contribute to the contrasting age-related fine scale SGS.