The objective was to investigate in vivo the rheological effects of plasma substitutes (fluid gelatin, dextran 40, dextran 60, hydroxyethylstarch). The study was performed during a hypervolemic hemodilution in 60 adult volunteers who underwent an odontological surgery. Two control groups were infused with Ringer-Lactate or 4% human albumin. Each patient was infused with 500 ml of plasma substitute or control fluid. Blood samples were obtained at four different times: prior to (t0) and immediately after (t1) the infusion, then 3 hours (t2) and 24 hours (t3) after the hemodilution. At t1 and t2, plasma viscosity did not present important variations, except for the group infused with dextran 60 where a significant increase in plasma viscosity was observed. Blood viscosity and erythrocyte aggregation measurements were performed at hematocrit 40% to determine the intrinsic properties of substitute and eliminate the effects of hemodilution. Blood viscosity at high shear rates underwent minor variations in all groups. In contrast, blood viscosity at low shear rate and aggregation parameters exhibited important variations which markedly depended on the used substitute. The infusion of Ringer-Lactate, albumin and dextran 40 induced a significant decrease in blood viscosity and in erythrocyte aggregation. With dextran 60 and gelatin, blood viscosity and erythrocyte aggregation increased. Finally, the infusion with hydroxyethylstarch did not present important differences as compared to the results recorded at t0. Twenty for hours after the infusion, it should be noted an increase in plasma fibrinogen, probably due to surgery, which caused an increase in erythrocyte aggregation in all groups.(ABSTRACT TRUNCATED AT 250 WORDS)