Dextrans are known to inhibit platelet aggregation. However, conflicting results have been reported with hydroxyethyl starch (HES), usually administered as a single dose. To clarify this, we studied the effects of HES on platelet number, aggregation and volume during the course of long-term hemodilution therapy. Twelve patients with disturbances of cerebrovascular perfusion were randomized and divided into two groups of 6 patients each. One group was treated with 10% HES 200/0.62, the other with 6% HES 200/0.62. The patients in group 1 received a loading dose of 500 ml, followed by 500 ml daily for 10 days. Group 2 patients were infused a loading dose of 500 ml, followed by 1000 ml on days 1 to 4 and 500 ml on days 5 to 10. 10% HES had a larger volume effect. The hematocrit was lowered by 19.0%, whereas 6% HES caused a decrease of only 15.6%. 10% HES lowered the platelet number significantly (p < 0.05) by 5.8%, 6% HES did not lower the platelet number significantly. Platelet volume decreased significantly during therapy (p < 0.05) in both groups. 10% HES reduced the platelet volume by 13.9%, the reduction with 6% HES was 9.0%. Under therapy with 10% HES, platelet aggregation declined slightly, but significantly (p < 0.05), whilst no effect was seen with 6% HES. During long-term hemodilution therapy with HES, the initial decrease in platelet number, which is due to dilution, is quickly compensated. The continuous decline in mean platelet volume during therapy is probably due to colloid-osmotic shrinkage. Since there is a positive correlation between platelet size and function, the slight inhibition of platelet aggregation caused by 10% HES is possibly due to the observed decline in platelet volume.