1. Platelet survival in 27 insulin-dependent diabetic patients with severe retinopathy was studied in a double-blind cross-over trial using placebo, aspirin (990 mg/day) and a combination of dipyridamole (225 mg/day) with aspirin at two dosage levels (330 mg and 990 mg/day). 2. Twenty patients (group I) had 51 Cr-labelled-platelet survival after treatment with placebo and the high-dose-aspirin/dipyridamole combination. The remaining seven patients (group II) had platelet-regeneration times measured after each of the four treatment periods. 3. Treatment of group I patients with the high-dose-aspirin/dipyridamole combination resulted in significant ( P < 0·001) prolongation of platelet survival from 7·3 ± 0·2 (mean ± sem ) days to 8·4 ± 0·1 days. 4. In group II patients, when compared with the mean placebo result of 7·2 ± 0·2 days, the mean aspirin-labelled-platelet-regeneration time was significantly ( P < 0·01) longer only after high-dose-aspirin/dipyridamole (9·8 ± 0·5 days) but not after low-dose-aspirin/dipyridamole (8·3 ± 0·5 days) or aspirin alone (7·3 ± 0·3 days). 5. These results suggest that it may be premature to consider reducing the dose of aspirin in aspirin/dipyridamole combinations below 1 g/day when used as antithrombotic therapy.