Data derived from a long-term grazing trial were used to determine short-term effects of mixed grazing by cattle and sheep in Highland Sourveld. Five cattle to sheep ratio treatments (viz. 1 :0, 3:1, 1 :1, 1 :3, 0:1) were applied, each at three stocking rates (viz. 0.5 (low), 0.71 (medium) and 1.0 (high) animal unit equivalents (AUE) ha¯¹). Ratios were expressed in terms of AUE cattle: AUE sheep. Stocking rate and ratio treatments were balanced at the start of each grazing season. Fourteen-month old Hereford steers and 'two-tooth' Merino wethers were used as experimental animals. The trial comprised two components, viz. an animal production component and a simulated component. The animal production component was conducted only in the medium stocking rate treatment where the effect cattle to sheep ratio on the performance of cattle and sheep was determined for each of four grazing seasons (viz. 1989/90, four-paddock rotational grazing The low and high stocking rate 1990/91, 1991/92, 1992/93). A management system was applied. treatments were implemented by simulating a four-paddock rotational grazing system and using a single paddock for each stocking rate/ratio combination. Grazing of the simulated treatments coincided with the grazing of a specific 'test' paddock in each ratio of the medium stocking rate treatment. These 'test' paddocks and the simulated treatment paddocks were monitored to determine the impact of mixed grazing by cattle and sheep on individual grass plants and the sward. The whole of the experimental area was rested during the growing season prior to the start of the trial and all paddocks were burnt in the dormant season just before the start of the first grazing season. The trial therefore commenced with a with a uniform sward of immature herbage in all paddocks allocated to each group of animals. ln the second and subsequent growing seasons only those paddocks which had been rested in the previous season were burned prior to the onset of growth in early spring. Rainfall for the first three grazing seasons was similar to the long-term mean of 790mm whilst the last season was considered 'dry' with 554mm recorded during 1992/93. In all grazing seasons, as the proportion of cattle in the species mix increased, sheep performance increased. A decline in sheep performance was recorded in each ratio treatment from the first to the third season. This decline was attributed to the increased maturity and thus lower quality of herbage on offer to the sheep, and the fact that only one paddock available to the animals in the second and third grazing seasons had been burned prior to the start of the season. Sheep performed best during the 'dry' season where herbage quality was maintained for longer into the grazing season than in previous seasons. In contrast, cattle performance was affected by the stocking rate (animals ha¯¹) of cattle rather than the presence of sheep. As the quantity of herbage on offer per steer declined steer performance declined. Animal performance data were also used to predict the effects of adding cattle to a sheep production enterprise and vice versa. The general trends were that the introduction of cattle into a low stocking rate, sheep-only production enterprise would allow for an increase in the stocking rate of sheep while maintaining the performance of the sheep. In this way the carrying capacity of a farm may be improved. Stocki ng rate and ratio treatments varied from those established at the start of each grazing season due to the differential performance of the cattle and sheep in each treatment. As the proportion of cattle in the species mix increased, stocking rate increased and the ratio widened in favour of cattle. Stocking rate (AUE ha¯¹), calculated a posteriori for each season, was the major influence on the severity of grazing on individual plants and within patches. As stocking rate of cattle and sheep at the various ratios increased, the extent and severity of grazing increased. At stocking rates in excess of O.8AUE ha¯¹ however, sheep-only grazing resulted in a greater proportion of plants, per species and per area, being grazed more severely than was the case for an equivalent stocking rate (AUE ha¯¹ ) of cattle. Furthermore, cattle and sheep had similar effects on patch size distribution when stocked at the same number of AUE ha¯¹. There were no measurable effects of stocking rate and ratio on proportional species composition and basal cover over a two year monitoring period. A technique for estimating basal cover in tufted grasslands was developed and is presented as an appendix to the thesis. Data were also used to evaluate the use of AUE as an integral part of the grazing capaci ty concept. Results indicated that cattle and sheep cannot be equated in terms of AUE when referring to the grazing impact. It is suggested that the definition of grazing capacity should include the species of livestock and assume a grazing management system appropriate to the grazing habit of the animals concerned. Resul ts of the trial provide strong indications that, in the long-term, the current recommendations of grazing cattle together with sheep in order to prevent the degradation or loss of veld condition which occurs in sheep-only grazing systems, will not succeed. A four-paddock rotation grazing system does not appear to be an appropriate veld management system for sustainable sheep production in sour grassveld. An alternative approach to veld management is suggested in which the sheep are confined to only those areas of the farm which were burnt at the start of the grazing season. Ideally, sheep should not allocated to the same paddock for two consecutive grazing seasons. / Thesis (Ph.D.)-University of Natal, Pietermaritzburg, 1994.