Four out of 15 dogs were successfully induced diabetes mellitus with intravenous iqiection of 30 mg of streptozotgin and 50 mg of alloxan monohydrate per kilogram body weight and maintained more than 9 weeks without iqiection of insulin or oral hypoglycemic sgent Histopathologicallyi these four dogs have typical diabetic lesions such as degeneration and vacuolation of pancreatic islet cells, and fatty change of liver at necropsy in the end of study. Serum glucose level increased dramatically at 24 hours post-injection but serum fructosamine level increased gradually and reached plateau at 31-41 days post-injection of streptozotocin and alloxan. Serum fructosamine concert%lion correlated very well with serum glucose concentration of preceding 4-7 weeks in experimentally induced diabetic dogs. Our data suggest that serum fructosamine reflects mean glucose concentration of preceding 4-7 weeks in experimentally induced diabetic dogs.