SUMMARY The effectiveness of two commonly available liquid diets was assessed in 40 severely malnourished black African patients. All patients were shown to have normal xylose absorption. The diets were given according to the manufacturer's recommendations. One diet was lactose containing (LC diet) (150 g/d) and high protein (112 g/d), the other normal protein and lactose free (LF diet) (protein 67 g/d), total energy content being similar. Patients were randomly divided into two equal groups and allocated (blind) to one of the diets. Tolerance and nitrogen balance were assessed over two three day periods on half and then full strength formulations. Severe intolerant symptoms were observed in 50% of patients on half strength and 94% of patients on full strength lactose containing diet with evidence of malabsorption of fluid, nitrogen, and fat. Despite high stool nitrogen losses (3.75±1 04 g/d), however, positive nitrogen balance was achieved in most patients receiving the full strength LC formulation. On the other hand, the full strength LF diet was generally well tolerated and was associated with significantly lower faecal losses and positive nitrogen balance. The results indicate that high density lactose containing liquid formulae are poorly tolerated by severely malnourished black African patients, while lactose free formulae containing approximately 10 g nitrogen/d are well tolerated and result in positive nitrogen balance.A recent survey has shown a high incidence of malnutrition among our hospitalised medical patients: over 80% of male and 50% of female patients had significant losses of protein and fat stores.' As most of these patients have relatively normal gut absorption, nutritional support can be based on enteral regimens. There are at present only two such regimens locally available that are suitable for fine nasogastric infusion; one being milk based and therefore lactose containing, the other lactose free. A further study of ours has revealed the population to be predominantly lactose intolerant,2 with 90% of our medical ward patients showing a flat blood glucose response to oral lactose. Consequently, we might have anticipated problems with the milk based diet and designed the study to assess the practical importance of this.