94 diabetic patients established on treatment with porcine (n = 47) or bovine (n = 47) insulin took part in a double-blind crossover trial, in which 6-week periods of treatment with the appropriate animal insulin were compared with periods of treatment with biosynthetic human insulin (BHI). 6 patients withdrew during the trial, in 3 cases because of hypoglycaemia while taking BHI. In bovine-insulin-treated patients, the mean glucose level (mean of seven capillary-blood samples over 1 day), the modified M index, and total daily insulin requirement were the same on BHI and bovine-insulin treatment. For porcine-insulin-treated patients, mean glucose level and the modified M index were slightly higher on BHI than on porcine-insulin treatment (9.7 vs 9.0 mmol/l and 79.6 vs 65.0, respectively), despite an average increase of 2.3 units/day of BHI after 6 weeks of such treatment. Hypoglycaemic episodes were not significantly more or less frequent on BHI in either group of patients. In both groups fasting blood glucose was higher during BHI treatment than during animal-insulin treatment (14.2 vs 12.8 mmol/l [bovine group]; 12.1 vs 9.6 mmol/l [porcine group]). In bovine-insulin-treated patients blood glucose before the evening insulin injection was higher on BHI than on bovine insulin (11.6 vs 10.0 mmol/l). BHI appears to be a safe alternative to porcine or bovine insulin. Differences in the pharmacokinetics of BHI may account for the observed differences in blood-glucose responses.