Effects of herbivory, fire, and tree harvesting on Acacia drepanolobium were studied using plant population dynamics as the philosophical basis of research. Specifically, growth rates, chrono-sequence of re-growth, biomass and charcoal yield, herbivory, flowering, seed production, germination, mortality and the ants of Acacia drepanolobium were studied in the black cotton ecosystem of Mpala Research Centre, Laikipia, Kenya, between September 1995 and December 2000. Acacia drepanolobium was the most abundant tree or shrub with densities ranging from 80% to 98% of all the overstorey species, but it was the least browsed of all the trees and shrubs in the black cotton ecosystem, ranging from a mean of 7.2% to 9% of the individuals browsed. The tree is inhabited by four Acacia ant species, Crematogaster mimosae, Crematogaster sjoestedti, Crematogaster nigriceps, and Tetraponera penzigi, which are believed to be obligate, and which probably play a role in the low browsing rates observed. Six herbivore treatments replicated three times (no herbivores - O; only cattle - C, all herbivores allowed - MWC {control}, mega-herbivores {elephants and giraffe} and wildlife {W} – MW only, wildlife – W - only {all wildlife except mega-herbivores} and wildlife and cattle only - WC) was the main experimental design used in understanding the dynamics of the tree species under influence of different herbivores. Mean annual height growth of A. drepanolobium trees was 24.9 cm yr-1, while the mean Relative Growth Rates ranged from 14.6 x 10 –3 cm cm-1 yr-1 to 18.7 x 10 –3 cm cm-1 yr-1. Growth rates were different among the herbivore treatments and between seasons. Shoots of the tree grew by a mean range of 6.8 cm to 9.1 cm, were similar among the treatments but differed among the seasons. Canopy volume increased over time although it fluctuated with seasons, suggesting an increase in bushiness of A. drepanolobium in the ecosystem. Trees occupied by different ant species showed differences in shoot density (number of new shoots per twig), being greater in Crematogaster nigriceps occupied trees compared with the other ant species. Swollen thorn (gall) density per unit of twig length was greatest in treatments with megaherbivores; these galls were significantly larger on trees occupied by the ant Crematogaster nigriceps. Treatments with herbivores were more spinescent than the total exclusion treatment. Spine lengths ranged from 0.8 to 2.4 cm, and recorded a progressive reduction of up to 36.36.7% in treatments without browsers suggesting a relaxation of induced defence in A. drepanolobium. Flowering in A. drepanolobium was low and staggered over the study period ranging from 0.8% to 2.0% of the trees with no differences among the treatments suggesting that the level of herbivory was not sufficient to influence reproduction of the tree in the experimental site. Consequently, seedling recruitment was very low within the experimental site. However, a nearby site recorded flowering of between 22.7% and 93.5%. Mean pod production, mean number of seeds per tree and mean weight of pods and seeds had a positive linear relationship with tree density (R2=0.77, 0.81 and 0.81 respectively). Trees occupied by Crematogaster mimosae were the most likely to flower (68%) compared with C. nigriceps (5.8%), again suggesting that ants had an effect on the tree’s reproduction. Mortality of A. drepanolobium trees averaged 0.9% to 4.2% over the study period, being significantly greater in treatments with mega-herbivores. Seedling survival ranged from 42% to 75%, being greatest in the cattle only treatment. Between 30% and 100% (mean 67.2%), of A. drepanolobium seeds were attacked by a bruchid beetle (Bruchus sp.). Seeds attacked by bruchid beetles had significantly lower germination rates. Similarly, seeds passed through a fire also recoded significantly low germination rates compared with normal seeds. Fire (3.4%) and bruchid beetles (20.7%) germination compared with (control) undamaged seeds (84%) play an important role in the population dynamics of A. drepanolobium. Fewer A. drepanolobium seeds (33%) were recovered from the surface compared with buried (72%) seeds after a fire, indicating seed loss from the effect of fire and predation. In the burn experiment, fire top-killed 16% of A. drepanolobium trees but no tree or seedling was killed. On the other hand, fire significantly reduced the density of non-A. drepanolobium trees by between 50% and 100%, with none of them showing signs of coppicing after the fire unlike top-killed A. drepanolobium trees. Woody biomass from A. drepanolobium was strongly related to stem diameter (Y = 3.77x + 1.17, R2 = 0.96, P < 0.001). Mean charcoal production from earthen kilns was 2.83 Mg ha- 1. Height and stem diameter in coppicing stands increased at a mean rate of 28.6 cm yr-1 and 0.7 cm yr-1 respectively. Biomass in coppicing stands accumulated at a mean rate of 1.3 Mg ha-1 yr-1 in a 14-year period, yielding dry biomass of 18.26 Mg ha-1 useable wood that can produce a minimum of 3.0 Mg ha-1 of charcoal. This study shows that Acacia drepanolobium populations are affected by several factors including herbivory, fire and ants. The population dynamics of this tree shows that it can be harvested for sustainable charcoal yield over a 14-year cycle. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2007.

Herbivory is a key process that mediates the abundance of primary producers and community composition in both terrestrial and aquatic ecosystems. On tropical coral reefs, changes in herbivory are often related to phase shifts between coral-dominance and dominance by seaweeds, or foliose macroalgae. Resilience or capacity to resist and reverse such phase shifts is, therefore, viewed as a critical function on coral reefs. This thesis used grazer exclusion and assay experiments at six sites within three different fisheries management regimes in Kenya to identify the impacts of herbivores (sea urchins and fishes) on algal dynamics in the context of coral reef resilience. First, I examined the grazing rates necessary to prevent phase shifts by quantifying consumption and algal production. Here, I found that, over a 390-day experiment, at least 50 percent of algal production must be consumed to avoid accumulation of algal biomass. Using video observations, I also showed that scraping parrotfishes remove more algae (per unit of fish biomass) than previously assumed, and that sea urchins, if released from predation, have similar impacts to fishes. Then I focused on algal succession, and found that sea urchins and fishes have different effects that are mediated by their abundances and species composition. Where sea urchins were less abundant and parrotfishes absent (e.g. young fisheries closures), progression of algae from turfs to early and then late successional macroalgae occurred rapidly and within 100 days. I then turned my focus to the removal of already established macroalgae (grown for > 1 yr in the absence of herbivores) and showed that sea urchins and browsing fishes were able to remove significant amounts of macroalgae where either herbivore was abundant. However, using multiple-choice selectivity assays and in situ video recordings, I found that browsing fishes fed very selectively with low overlap in diet among species, leading to low functional redundancy within a high diversity system. Finally, using long-term survey data (from 28 sites) to build a 43-year chronosequence, I showed that it is possible that the effects of herbivory will not be constant across transitions from open fishing to fishery closures through non-linear grazing intensity. Therefore, increases in herbivory within fisheries closures may not be immediate and may allow a window of opportunity for algae to go from turf to unpalatable macroalgae until scraping and browsing fishes fully recover from fishing (~ 20 years). The findings in this thesis are novel and raise concern over the potential implications of the slow recovery of parrotfishes or, given lower than expected functional redundancy in grazing effects, the absence of even one browsing fish species in fisheries closures. Overall, this thesis highlights the importance of herbivore community dynamics in mediating interactions among algae, and provides new insights for conservation and management actions that attempt to bolster the resilience of coral reefs.