The soil seed bank plays a dynamic role in the regeneration of plant communities after natural and anthropogenic disturbance. In this thesis, I addressed how disturbances influence the vegetation and seed bank of Florida rosemary scrub. In Chapter One I evaluated changes in species composition and spatiotemporal pattern of the vegetation and seed bank along a gradient of disturbance. During the summers and winters of 2007-2009 percent ground cover and seed bank species composition were assessed among replicates of three vegetation types subjected to minimal, moderate, and extreme anthropogenic disturbance (native rosemary scrub, degraded scrub, and agriculturally improved pasture, respectively). These vegetation types shared the same soil and topographic characteristics but differed in disturbance history. I found that species composition and spatial pattern varied with disturbance. In pastures the compositional and structural characteristics of rosemary scrub were lost and only native scrub species able to evade herbivory persisted in this community. Native and degraded scrub differed most from each other in species abundances and spatial pattern. Degraded scrub showed highest abundance of subshrubs and a spike moss species, while rosemary scrub was dominated by shrubs. The seed banks of scrub herbs in degraded scrub had a tendency towards a random spatial distribution that lacked association with aboveground cover. Conversely, rosemary scrub seed banks tended to have an aggregated distribution and were associated with occurrence of conspecific species aboveground, litter, and shrub cover. These results indicated a change in the spatial heterogeneity of the seed banks of scrub herbs in degraded scrub. In Chapter Two I evaluated changes in seed bank density with time-since-fire in native rosemary scrub. Due to large pulses of recruitment immediately after fire and population decline iii with time-since-fire, I predicted seed density with time-since-fire would follow a unimodal function with low density in early and late years post-fire, and highest density at intermediate time-since-fire. I compared seed density data among sites with different time-since-fire: two sites each of three, six, ten and 24 years time-since-fire and three long-unburned sites ( > 24 years). Variability in seed bank composition and density increased with time-since-fire and only recently burned stands were distinctly different from the other time-since-fire age classes. Some species and functional groups did exhibit a quadratic or cubic association to time-since-fire (ruderal herbs, subshrubs, Ceratiola ericoides, Lechea cernua, Paronychia chartacea, Phyllanthus tenellus); however, timing of the peak in seed density varied depending on life span and age of reproductive maturity. Scrub herbs were the most abundant functional group in the seed bank and showed highest density in the first ten years post-fire. This pattern corresponds to the pattern of aboveground species abundance and suggests abundances above- and belowground are closely linked. Understanding the dynamics of the seed bank in both naturally and anthropogenically disturbed communities in Florida rosemary scrub is important for the restoration of scrub habitat and management of existing populations of endangered and threatened scrub species endemic to the Lake Wales Ridge in central Florida.

Although capable of sexual reproduction, many plants also rely heavily on clonal reproduction. The formation of multiple, physiologically-independent units with the same genotype has important implications for spatial genetic structure and genetic diversity in these plants. The endangered scrub-dwelling perennial, Polygonella myriophylla is known to reproduce both sexually and clonally but no study to date has been able to investigate the spatial genetic patterns that occur in this species. I use microsatellite markers to investigate questions about clonal structure and genetic diversity in five populations of P. myriophylla and address some of the implications of my findings for conservation of this species: Overall, I find that 57% of sampled clusters of P. myriophylla are composed of a single genet (genetic individual) with multiple physiological units (ramets) while the remainder are made up of two or more genets. I found differences in both clonal reproduction and genetic diversity among populations. I also found evidence of limited gene flow even over small spatial scales (less than 10 km) and for at least 4 genetic clusters occurring within the species range. Despite high levels of genetic diversity overall, there is evidence of reduced genetic diversity in two populations My results suggest that high levels of clonality may be important in maintaining genetic diversity in P. myriophylla. I also provide evidence that dirt roadsides may not represent a refuge for this species.