M.Sc. / The catchment area of the Olifants River has, over a number of years, been exposed to extensive mining, agriculture and urbanisation activities in the Witbank-Middelburg and Phalaborwa regions which has largely contributed to the deterioration of water quality. Of major concern is the influence anthropogenic activities have on the aquatic ecosystem of the Olifants River within the Kruger National Park and how me health of fish residing In these waters is affected. To determine fish health, a biological monitoring method, the Health Assessment Index (H41), was tested for the first time In South Africa. Four surveys were conducted at Mamba and Balule In the Olifants River, Kruger National Park during 1994 (February, May, July and November). An additional survey was conducted in February 1995 at Loskop Dam. Samples of water and sediment were taken for analysis of metals and physical and chemical water parameters. A maximum of 20 Clarias gariepinus fish were sampled at each location. Evaluation of the fish was done according to guidelines set in the HAI and parasite population composition (prevalence, abundance, mean intensity) was determined. Organ and tissue samples including gills, liver, muscle and skin were analysed for the bioaccumulation of chromium, copper, Iron, manganese, nickel, lead, strontium and zinc, using atomic absorption spectrophotomeby. Metal concentrations in the water at Mamba and Balule were within guideline limits, whereas concentrations at Loskop Dam were above guidelines. As reflected by the application of the Aquatic toxicity Index, Loskop Dam presented with the poorest water quality followed by Mamba then Balule. It was found that certain physical and chemical variables namely fluoride, potassium, sulphate and total dissolved solids concentrations at Mamba and Balule were relatively high, particularly during drier months. Metals accumulated in organs and tissues, with the highest concentrations In the gills followed by the liver, skin and muscle. The discriminant analysis, utilising metal bloaccumulation, discriminates between water quality at Mamba and Balule, revealing a 100 % classification probability for each survey. Values obtained In the application of the HAI indicated that variables with good predictor accuracy were plasma protein, all parasites, endoparasites, liver, white blood cell counts, ectoparasites, skin, fins and gills. The discriminant function for the HAI generally Indicated variables similar to those exhibiting high predictor accuracy. The discriminant function showed relatively low classification probability for each survey. In case 1, where separate endo- and ectoparasite variables were Included in the determination of me discriminant function, probability for me entire study ranged between 47.5 % and 84.2 %. In case 2, where endo- and ectoparasite variables were given a refined score rating system, probability ranged between 62.5 % and 100 %. The low classification probability Indicates either the Importance of repetitive testing for this technique or a total departure from it. Results showed that fish populations with higher HAI values are found in water of poorer quality (Mamba), while healthier fish populations i.e. with lower RAI values are found In water of better quality (Balule). Parasite data shows a similar tendency by indicating that ectoparasites are abundant in water of higher quality, while endoparasftes increase in water of poorer quality. The HAI reflects the condition of fish populations in relation to their aquatic environment, therefore, the HA/ gives an indication of water quality and should be used as a first level screening tool. If complemented by a parasite survey, distinguishing between endoand ectoparasites, the results from the HAI will be enhanced.