Several breeds of dog are at high risk of developing diabetes mellitus. Diabetes can result in blindness and other severe illnesses within one year of diagnosis, mostly due to inadequate glycemic control. The standard evaluation of therapy relies on a 12-hour, in-hospital glucose curve. This test does not reflect day-to-day and nocturnal glycemic variation in humans. Glucose curves also show a day-to-day variability that has been observed in dogs. If changes observed in people also occur in dogs, in-hospital glucose curves may not be the best way to evaluate diabetes therapy. We will utilize a continuous glucose monitoring system (CGMS) to measure glucose concentration every 5 minutes for 72 hours in 20 stable diabetic dogs. Comparisons will be made with results obtained in-hospital and at-home, during the day and at night, and on day 1 verse day 2 at home. Nadir, peak glucose, and mean glucose will be evaluated as well as recommendations regarding insulin does and type for each curve. Assessing the limitations of in-hospital glucose curse will improve management of diabetes in dogs, improving glycemic control. Long-term control of glucose is the most important factor in preventing complications in patients with diabetes mellitus.