Six adults with non-insulin dependent diabetes mellitus participated in a laboratory procedure to determine the effects of an acute stressor, mental arithmetic, on blood glucose, cardiovascular, and subjective indices of stress. They then completed 12 days home monitoring of stressful events, subjective stress, and blood glucose. Diet and activity were controlled to evaluate the direct effects of stress on blood glucose in the natural environment. Laboratory results showed significant increases in blood glucose, cardiovascular (HR and SBP), and subjective stress ratings during the mental arithmetic task when compared to a resting condition. Home monitoring data were consistent with the laboratory findings; blood glucose range tended to be greater on high vs low stress days, especially when the difference between high and low stress was greatest. These findings suggest that the laboratory stress induction procedure was externally valid and that in the natural environment, stress has a hyperglycemic effect on blood glucose.