The effect of altering the frequency of self-monitoring of blood glucose upon blood glucose control was assessed in 25 Type 1 diabetic patients. Patients entered a balanced, crossover study in which they were randomly allocated to three 12-week periods. During each period the frequency at which patients self-monitored their blood glucose was: (1) four-point profiles on 2 days per week; (2) four-point profiles on 1 day per week; (3) two-point profiles on every day of the week. There were no significant differences in mean blood glucose concentration, fructosamine or glycosylated haemoglobin concentrations during the three periods. There was no correlation between the frequency at which patients altered insulin doses and measures of metabolic control. Altering the frequency of blood glucose self-monitoring had a variable and uncertain influence upon the frequency at which patients altered insulin doses. Some patients felt unsafe monitoring blood glucose only four times per week. There was however no patient consensus on the best-liked regimen.