A long-term grazing trial was started in 1977 at Armoedsvlakte Research Station, about 10km west of Vryburg, in Tarchonanthus veld of the Ghaap’s Plateau, which is a variation of the Kalahari Thornveld veld type. The main aim of this study was to use the extensive veld condition and animal production data set to investigate the effects and interactions of stocking rate, grazing system applied and seasonal rainfall on veld condition and cattle production. The grazing trial has changed three times since its inception resulting in three different phases. The main changes in veld condition during phase one (1977-1991) was due to density independent effects (e.g. seasonal rainfall) and not density dependent effects (e.g. stocking rate). A major change occurred in 1985 following a multiple year drought. The drought resulted in adverse changes in species composition, basal cover and residual biomass of all treatments. The system did not recover from the drought during phase one, despite well above mean seasonal rainfall for a number of years after the drought. During phase two (1992-1999) and phase three (2000 to present) completely different vegetation dynamics occurred than what was experienced during phase one. Density dependent effects (e.g. stocking rate) were more important in explaining variation in veld condition during these two phases. High stocking rates resulted in adverse changes in species composition, poor basal cover and a low residual biomass production. It is however important to note that seasonal rainfall did explain a significant additional amount of variation in veld condition. This suggests that a continuum of non-equilibrium and equilibrium vegetation dynamics occurred in these two phases. The residual biomass and seasonal rainfall model for phase one indicate completely different results for the gain per animal data. In the seasonal rainfall model, stocking rate does not have a significant effect on gain per animal, but seasonal rainfall and the interaction of stocking rate with seasonal rainfall explains most of the variation in gain per animal. This suggest a continuum of non-equilibrium and equilibrium dynamics and that animal production is more sensitive to seasonal rainfall than to stocking rate, although the significant interaction of stocking rate with seasonal rainfall suggest that the seasonal rainfall effect on animal production is dependant on stocking rate. The residual biomass model however indicates that stocking rate is more important than rainfall in explaining variation in the mass gains per animal. The stocking rate effect on gain per animal was significant and indicated that as stocking rate increased, that gain per animal decreases. Seasonal rainfall and the interaction of stocking rate with seasonal rainfall had no significant effect on gain per animal. The amount of variation explained by the seasonal rainfall model was larger than the residual biomass model and this indicates that rainfall explains more variation in gain per animal, than residual biomass does. This possibly indicates that non-equilibrium effects are stronger than the equilibrium effects, but it is important to notice that stocking rate had a significant effect in some cases. The gain per hectare models (seasonal rainfall and residual biomass) for phase one indicates that stocking rate has a significant effect on gain per hectare. Increasing stocking rates resulted in higher gain per hectare, which suggests that the turning point of the typical “Jones and Sandland model” has not been reached and this might be due to light stocking rates applied during the duration of phase one. The seasonal rainfall model however has significant effects of seasonal rainfall and interactions of stocking rate with seasonal rainfall on gain per hectare. This suggests that the effect of stocking rate is dependent on seasonal rainfall and that seasonal rainfall explain an additional amount of variation in gain per hectare. In general, it appreared that the optimal stocking rate for animal production was higher than those applied during the duration of the trial, but this is due to lower than planned actual stocking rates applied during all three phases of the trial. It is very difficult to determine a generic optimal stocking rate for different rainfall volumes and it is recommended that the actual stocking rate for different ecological zones be determined based on rainfall, biomass, species compos[i]tion, basal cover and available browse and not just on the provisional recommendations. The type of grazing system applied did not show any statistically significant effects on both gain per animal and gain per hectare for the animal production data during phase one. This result is interesting and contradictive to most of the scientific literature where some authors concluded from their studies that rotational grazing systems produce higher animal production than continuous grazing systems, whereas others researchers state that continuous grazing systems produce higher animal production than rotational grazing systems. In phase two both the residual biomass and seasonal rainfall models for phase two did not show any significant effects and interactions of stocking rate, seasonal rainfall level and/or residual biomass on both gain per animal and gain per hectare. Both the residual biomass and seasonal rainfall models for phase three did not show any significant effects and interactions of stocking rate, seasonal rainfall level and/or residual biomass on animal gains per animal. The seasonal rainfall model did not show any any significant effects and interactions of stocking rate, seasonal rainfall level and/or residual biomass on animal gains per hectare. However, the residual biomass model indicated that stocking rate had a significant effect on gain per hectare and the production closely followed the Jones and Sandland (1974) model as at low stocking rates, gain per hectare increases at a rapid rate, but as stocking rates increases to high stocking rates, the rate of increase in gain per hectare declines, until it eventually reaches a turning point, where after gain per hectare declines with increasing stocking rates. Stocking rate only had a significant effect on the condition score of cows during phase two and phase three, as high stocking rates resulted in poor animal condition in both phases. No significant effects and interactions of stocking rate and seasonal rainfall were indicated on calving percentage, weaning percentage, conception rates and percentage of desirable meat produced during phase two. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, [2011].