The effect of rapid massive transfusion upon platelet counts has been studied retrospectively in 24 patients treated for severe trauma. Multiple linear regression analysis showed that a platelet dilution model with an exponential form described the quantitative relationship between the pre- and posttransfusion platelet counts and the volumes of blood products administered. Intravenous salt solutions had very little effect upon the platelet count even when large volumes were infused. Red blood cell transfusions decreased the platelet count to an extent two and one-half times greater than that of colloid infusions. The differences in the magnitudes of the dilution effects are attributable to the different physiologic volumes of distribution of the various classes of infusates. If only red blood cells (RBC) and fresh frozen plasma (FFP) are transfused, two useful equations obtain. First, the current platelet count (Pt) can be predicted knowing the initial platelet count (Po) and the number (#) of units of RBC and FFP administered: Pt = Po (0.634) e-0.046 [#RBC + #FFP/8]. Second, the point during massive transfusion when platelet transfusions will be necessary can be predicted knowing the initial platelet count.