The day-to-day variability of blood glucose concentrations in juvenile diabetes means that it is often more reasonable to aim to achieve a generally good pattern of blood glucose control, rather than regularly to assess the next insulin dose after each blood glucose measurement. This means that immediate assessment by the patient of his blood glucose concentrations is not always necessary. We have investigated control in 22 insulin-requiring diabetic patients by means of a monthly series of four blood samples taken during a day into collector bottles and transported to a laboratory for blood glucose assay. The overall means before breakfast, before lunch, before dinner, and before bed were 6.1, 5.8, 7.3, and 7.2 mmol/L, respectively. In many patients, sufficiently good control can be obtained by this method so that it is not necessary to ask them to measure their own blood glucose concentrations or to ask them to obtain the fairly expensive meters for reading glucose oxidase strips. Control would then probably be best assessed by a series of three daily profiles taken once per month, with, if necessary, the results being discussed with the patient. On the other hand, in more unstable diabetes, home assessment by patients of blood glucose measurements is indicated.