We studied the effects of moderate changes in red blood cell RBC aggregation on blood flow in the vasodilated vascular bed of an isolated rat heart. We compared a non-aggregating RBC suspension (in Krebs-albumin medium) with RBC suspensions in 1% and 2% Dextran 70 (MW 70000), exhibiting two different degrees of moderate aggregation. Degrees of aggregation were precisely estimated by in vitro laser aggregometry. Each heart was perfused by the non-aggregating RBC suspension and by one aggregating RBC suspensions. Blood flow was measured in a range of perfusion pressure from 40 to 80 mm Hg. For the three RBC suspensions, linear pressure/flow relationships were found. From the comparison between the pressure/flow relationships obtained with Krebs albumin medium and either 1% or 2% Dextran, it was possible to compare in vivo the contribution of RBC to the viscosity (i.e., the relative apparent viscosity) in the 2 aggregating RBC suspensions with that of the non-aggregating RBC suspension. The contribution of RBC to the viscosity was found to be 20% to 25% lower in the 1% RBC suspension than in the non-aggregating RBC suspension. With 2% Dextran which induced a higher degree of aggregation no differences were found between the relative apparent viscosities of the aggregating and the non-aggregating suspension. From the comparison between RBC in 1% Dextran and Krebs-albumin, we concluded that in vivo a moderate RBC aggregation reduces viscous resistance due to the presence of blood in a vascular network. Since no more effect of RBC aggregation per se was found when the degree of aggregation was higher (with RBC in 2% Dextran), this suggests that, in this case, aggregation induces opposite effects along the myocardial vascular network which cancel each other out, thus inducing a nil net balance.