Both chemical and physical effects of red cells have been implicated in the spontaneous aggregation of platelets in sheared whole blood (WB). To determine whether the chemical effect is due to ADP leaking from the red cells, a previously described technique for measuring the concentration and size of single platelets and aggregates was used to study the shear-induced aggregation of platelets in WB flowing through 1.19-mm-diameter polyethylene tubing in the presence and absence of the ADP scavenger enzyme system phosphocreatine-creatine phosphokinase (CP-CPK). Significant spontaneous aggregation was observed at mean tube shear rates, (G) = 41.9 and 335 s-1 (42% and 13% decrease in single platelets after a mean transit time (t) = 43 s, compared to 89 and 95% decrease with 0.2 microM ADP). The addition of CP-CPK, either at the time of, or 30 min before each run, completely abolished aggregation. In the presence of 0.2 microM ADP, CP-CPK caused a reversal of aggregation at (t) = 17 s after 30% of single cells had aggregated. To determine whether red cells exert a physical effect by increasing the time of interaction of two colliding platelets (thereby increasing the proportion of collisions resulting in the formation of aggregates), an optically transparent suspension of 40% reconstituted red cell ghosts in serum containing 2.5-micron-diameter latex spheres (3 x 10(5)/microliters) flowing through 100-microns-diameter tubes was used as a model of platelets in blood, and the results were compared with those obtained in a control suspension of latex spheres in serum alone. Two-body collisions between microspheres in the interior of the flowing ghost cell or serum suspensions at shear rates from 5 to 90 s-1 were recorded on cine film. The films were subsequently analyzed, and the measured doublet lifetime, tau meas, was compared with that predicted by theory in the absence of interactions with other particles, tau theor. The mean (tau meas/tau theor) for doublets in ghost cell suspensions was 1.614 +/- 1.795 (SD; n = 320), compared to a value of 1.001 +/- 0.312 (n = 90) for doublets in serum. Whereas 11% of doublets in ghost cell suspensions had lifetimes from 2.5 to 5 times greater than predicted, in serum, no doublets had lifetimes greater than 1.91 times that predicted. There was no statistically significant correlation between tau meas/tau theor and shear rate, but the values of tau meas/tau theor for low-angle collisions in ghost cell suspensions were significantly greater than for high-angle collisions.