An indirect platelet radioactive antiglobulin test was used in a prospective crossmatch study in 29 patients for selecting compatible random donors for platelet transfusion. Patients received crossmatch-compatible, crossmatch-incompatible and uncrossmatched platelets. A successful outcome for transfusion was defined as a 24 h post-transfusion corrected platelet increment of greater than 4.5 X 10(9)/l. Of 47 predictions based on crossmatch results, the technique successfully predicted the outcome of the transfusion in 90% of cases. In all patients, transfusion of crossmatch-compatible platelets resulted in a significantly higher 24 h corrected platelet increment than was obtained following transfusion of incompatible platelets. The mean +/- SEM 24 h corrected post-transfusion increment (X 10(9)/l) was 17.79 +/- 2.01 in 26 patients who received crossmatch-compatible pooled random donor platelets; in 21 patients who received crossmatch-incompatible platelets, it was 1.19 +/- 0.56; and in 25 patients who received uncrossmatched platelets, 4.42 +/- 0.97. The differences were highly significant. There was an 83% correlation of results with those obtained using an indirect platelet suspension immunofluorescence technique.