Platelet thrombospondin and an unidentified but biologically similar plasma protein were shown to inhibit the gelatin-binding activity of fibronectin. Inhibition of fibronectin gelatin-binding activity was identified and quantitated by using latex-fibronectin particles in combination with latex-gelatin particles in a new competitive aggregation assay. Inhibition was expressed as the reciprocal of the dilution of test sample required to produce a 50% return of baseline control aggregation rate (inhibitor units). Serum and plasma from healthy donors (n = 60) showed similar reductions in fibronectin gelatin-binding activity (47.9 +/- 12.9 and 49.4 +/- 12.7 inhibitor units per milliliter, respectively). However, serum fibronectin gelatin-binding activity per milligram of fibronectin was significantly less than that of plasma. The addition of calcium chloride to platelet-rich plasma resulted in a similar reduction in fibronectin gelatin-binding activity per milligram of fibronectin. No change was observed after recalcification of platelet-poor plasma. Washed platelets (1 X 10(9)/ml) in Tris HCl buffer released 18 +/- 8 fibronectin inhibitor units per milliliter after calcium ionophore A23187 addition. When inhibitor-rich preparations from platelets and plasma were chromatographed on Sepharose CL-6B, the inhibitors eluted at the same location. Inhibitor-rich eluates from both sources bound to heparin-Sepharose and eluted with 0.45 mol/L NaCl. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of purified inhibitor preparations demonstrated a major protein band with an approximate molecular weight of 185 kd. Western blot analyses using antiplatelet thrombospondin identified the platelet-derived inhibitor as thrombospondin but failed to react with the plasma-derived inhibitor. These data demonstrated that platelet-released thrombospondin was responsible for the reduction in fibronectin gelatin-binding activity seen in serum. An unidentified plasma factor also inhibits fibronectin gelatin binding.