In a biometrically planned, randomised double blind crossover study the influence of a glucosidase inhibitor, acarbose, (Bay g 5421, pseudotetrasaccharide) on the blood glucose in a total of 10 dieting obese NIDDM patients was compared with placebo. There were no significant differences between the groups with regard to age, height, body weight and duration of diabetes. After a 2-week preliminary period 5 patients received Bay g 5421 for 4 weeks and then placebo for a further 4 weeks. After the same preliminary period the other 5 patients received placebo for 4 weeks and then Bay g 5421 for the second period of 4 weeks. On average, the blood glucose profiles were lower after Bay g 5421 than after placebo. This was statistically almost significant (p less than or equal to 0.09) for the area of the profiles, i.e., the average value of the 13 points of the daily profiles. The fraction of glycosylated hemoglobins (Hb A1a-c) was also lower during the period when the patients were treated with Bay g 5421 than with placebo. There was, however, no statistically significant difference. In contrast, the glucosuria-decrease during the Bay g 5421 period was significant. The serum cholesterol and serum triglycerides did not show any therapy-dependent difference. Thus it may be concluded that Bay g 5421 effectively reduces the blood glucose concentration, in particular, postprandial hyperglycemia in dieting diabetic patients type 2.