The aim of this thesis was to describe the structure and dynamics of ant and plant communities in the southern Karoo and to assess mechanisms of species coexistence in ant and plant communities. The role of species interactions in structuring natural communities was emphasised. Diversity indices were used to determine the importance of habitat in maintaining ant species diversity. Ant species diversity was not predicted by measurements of plant species diversity or vegetation structural diversity. Ant species richness was correlated with vegetation structural diversity but not with plant species diversity. Ant species appeared to respond to aspects of vegetation height. Although vegetation complexity influences ant species richness, competitive effects, particularly of dominant ant species, appear to suppress sub-ordinate ant species influencing measures of richness and diversity. Aggressive dominant ant species determine the distribution and abundance of sub-ordinate ant species. Interference competition for space was prevalent between dominant ant species and competitive success was a function of vegetation. It was not clear whether ants respond directly to physical conditions created by vegetation which then influences foraging activity or, alternatively, whether ants respond to productivity gradients which are affected by vegetation. Understanding vegetation dynamics is critical to interpreting patterns of ant species distribution and abundance. A patchy habitat disrupts the competitive dominance of aggressive dominant species, removing the potential towards habitat monopolisation, and therefore maintaining ant species diversity. In the Portulacaria afra rangeland, Pheidole sp. 2 was superior in well-shaded microhabitats but Messor capensis nested successfully under woody shrubs and several ant species (Tetramorium peringueyi, T. quadrispinosum, Monomorium alamarum, Ocymyrmex barbiger) persisted on bare nutrient-rich patches. Rangeland dominated by grass would favour Pheidole sp. 2 at the expense of other ant species which would be unable to establish successfully. There is no evidence supporting the notion that ant communities are at equilibrium. Rather, dominance hierarchies lead to the replacement of species over vegetation gradients with the tendency towards the aggressive acquisition and monopolisation of space. The coexistence of dominant ant species at study sites in the southern Karoo was a combination of territorial strategy and colonisation ability. In the P. afra rangeland most ant species escaped the severe competitive effect of Pheidole sp. 2 by persisting as fugitives on bare areas of local disturbance where Pheidole sp. 2 was less successful. At Tierberg, competitively inferior ant species with a decentralised territorial system coexisted with competitively superior ant species in an unstable equilibrium by pre-empting newly available space through the lateral expansion of territories. Competitively superior species, however, replace the inferior competitor (yet, superior coloniser) in time. Plant succession in the Portulacaria afra rangeland follows a deterministic pattern of cyclical replacement. No single mechanism adequately explains the coexistence of all plant species in this community. Nurse-plant effects were important for the establishment of P. afra and later-successional trees below woody shrubs and P. afra respectively. Soil moisture and nutrient levels below nurse plants were favourable for the successful establishment of seedlings but nurse plants also disrupt the inhibitory effect of grass on seedlings and young plants. Linked to the facilitative process is the directed dispersal of seeds by biotic vectors. Seed dispersal by ants and birds moves propagules away from the harsh competitive environment of established adult plants (particularly grass) to safe establishment sites below nurse plants. Complementary root systems of seedlings and nurse plants may facilitate the establishment of young plants but ultimately competition will reduce nurse plant vigour leading inevitably to the death of the nurse plant. At least two plant species (Lycium cinereum and Psilocaulon absimile) persist in the P. afra rangeland as fugitives on areas of local disturbance. The relatively cool and moist environment below P. afra clumps supports a variety of detritivorous taxa that enrich the soil. Once the P. afra clump collapses and dies, bare nutrient rich patches that favour the establishment of L. cinereum and P. absimile, ahead of the competitively superior grass component, remain. However, the replacement of the fugitive plant species by grass is inevitable as the soil nutrient levels decline. The competitive superiority of grass controls the dynamics of the P. afra rangeland system. Disruption of the inhibitory effect of grasses by elevated nutrient levels (e.g. detritivore activity, ant nests) permits plant species to establish that would otherwise be unable to in the absence of disturbance. Degradation of the grass component by overgrazing at the P. afra rangeland site has the same effect as disturbance by soil-nutrient enrichment and results in an increase in plant species diversity. All stages of the cyclical process would be relatively uncommon in a grass dominated system because of the inability of plants to establish successfully in grass. Strong feedback links exist between the ant and plant communities studied. Habitat patchiness maintains ant species diversity by disrupting dominance hierarchies. Ants maintain plant species diversity by moving seeds out of competitively harsh microhabitats into safe sites (often below nurse plants) for establishment. Soil nutrient enrichment (ant nests and various detritivore taxa) create disturbances which also facilitates plant establishment. Overgrazing by domestic stock causes the replacement of palatable by unpalatable plant species. Unpalatable plant species, released of suppressive effects, develop into a monospecific stand that inhibits the further establishment of seedlings. Loss of habitat patchiness results in habitat monopolisation by one (or two) dominant ant species. Habitat degradation therefore severs the complex feedback links between the plant community and the important invertebrate component (dispersal agents and detritivores) which disrupts the dynamic processes driving the system. / Thesis (Ph.D.)-University of Natal, Pietermaritzburg, 1998.

A grazing trial investigating the effect of season of grazing and stocking rate initiated at the Nortier Experimental Farm in 1988 provided an opportunity to assess the response of the veld to both grazing and environmental influences in an arid environment. The trial allowed an assessment of the relative influence of internal (equilibrium) and external (non-equilibrium) forces on the dynamics of an arid rangeland. This study involved the analysis of a nine year data set stretching from 1988 to 1996 and served to provide evidence supporting the existence of an equilibrium/non-equilibrium continuum in rangeland dynamics. The most significant implication of this result is that rangeland systems should not be classified as either equilibrial or non-equilibrial, but rather according to a continuum extending between equilibrium and non-equilibrium poles. The exact position of any system on this continuum is a function of the relative influence of internal and external forces on its species dynamics. The dynamics of the veld at the Nortier Experimental Farm showed significant response to both grazing and environmental variables suggesting conformity to both equilibrial and non-equilibrial paradigms. Both ordination and analysis of variance highlighted the importance of rainfall particularly in the fluctuations of the predominant grass species, Ehrharta calycina, which increased in abundance with rainfall. Partial ordination enabled the assessment of species variation following the removal of variation associated with rainfalL Partial ordinations revealed the gradual, directional movement of samples through multivariate space in response to grazing treatments. Individual plant species were also shown to be responding to grazing, the extent of which was influenced by season of grazing and stocking rate. Both the partial ordinations and the ANOVA showed Melothria sp., Tetragonia fruiticosa and Hermannia scordifolia as increasing and Ruschia caroli as decreasing in absolute abundance in response to grazing. Season of grazing was shown to significantly influence the abundance of H. scordifolia over time. The 'shrublherb complex', which constitutes the 'key resource' at the Nortier Experimental Farm displayed an increase in absolute abundance over the duration of the trial. This increase in absolute abundance was accompanied by an increase in the relative abundance of the palatable component of this resource. The application of medium to heavy stocking rates during spring, summer and autumn and low stocking rates during winter resulted in elevated absolute abundances of palatable plants. Furthermore, low stocking rates, when averaged across all season of grazing treatments, resulted in a significantly higher absolute abundance of unpalatable plants. These findings provide the basis for the development of management principles for the Strandveld Vegetation Type. The application of medium to heavy stocking rates within a rotational grazing system, as recommended by the literature dealing with grazing systems in the Karoo, is supported by the results of the Nortier grazing trial. Medium to heavy stocking rates should be applied during spring, summer and autumn and low stocking rates during the winter months. Furthermore, it is recommended that rests of between 12 and 14 months should be afforded to portions of the veld periodically due to the variability in growth, flowering and fruiting times ofdifferent plants in the Karoo. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2005.