Total body or exchangeable potassium is used as an important indicator of body cell mass in the study of body composition. Body composition studies have been used extensively in the study of nutrition but recent work has questioned the validity of using changes in total body potassium as a measure of protein or nitrogen variation. To investigate the relationship between tissue nitrogen and potassium during nutritional manipulation 382 tissue samples from 100 surgical patients were analyzed by Kjeldahl analysis for nitrogen content and flame photometric analysis for potassium content. Nitrogen was related to potassium in parenchymous or cellular tissues by the relationship N (mg/g) = 14.7 + 0.17 K (microEq/g). The tissue content of the two elements was highly correlated (r = 0.80, p less than or equal to 0.001). For skeletal muscle a similar relationship existed N (mg/g) = 16.6 + 0.15 K (microEq/g) (r = 0.76, p less than or equal to 0.001). These relationships held for all nutritional states and degrees of trauma and nutritional manipulation. In sequential parenchymous tissue samples obtained from 15 subjects delta N/delta K = 0.17 +/- 0.03 mg/microEq. Nonparenchymous or acellular tissue nitrogen and potassium were poorly correlated. The potassium content was very low in these tissues. Exchangeable potassium was a valid indicator of parenchymous tissue nitrogen and as such a legitimate measure of nutritional status.