CD36 is one of the major glycoproteins of platelets and known as GPIV. Besides platelets, CD36 is distributed in megakaryocytes, monocytes, capillary endothelium and mammary epithelial cells. In vitro analyses, CD36 is reported to act as receptors to a variety of ligands including collagen, thrombospondin, malaria-infected erythrocytes and oxidized LDL. However, it remains unclear to which of these functions CD36 is critical in vivo. CD36-deficient individuals can be the key to answer this question. In calcium-deficient state, CD36-deficient platelets exhibited a delay and decline of irreversible aggregation on agonist stimulation. Irreversible aggregation of platelets depends on intake of arachidonic acid once-secreted from platelets and production of its metabolite Eps/TxA2. The calcium influx in response to U46619 (TxA2 analogue) of CD36-deficient platelets was not different from normal platelets in the presence of indomethacin and ETYA. Defective aggregation of CD36-deficient platelets in calcium-deficient state seemed to be derived from defective intake of arachidonic acid. This assumption was verified by our results that inhibitory effect of arachidonic acid in aggregation depended on the presence of platelet CD36. Intake of arachidonic acid through CD36 may have an effect in low concentration state of arachidonic acid. The CD36 deficiency is present in several % in Japanese and approximately 0.3% in Caucasians and is divided in type I (deficient in platelets and monocytes) and type II (deficient only in platelets). Analyses of CD36 cDNA revealed that codon 90 (proline/serine) was critical as to the surface expression of CD36 protein. By analyses of CD36 genomic DNA, the CD36 gene could be classified; 1) serine90 type that was not translated as CD36 protein, 2) proline90 type that was not transcribed to mRNA, 3) proline90 type that was transcribed only in monocytes and not in platelets, 4) proline90 type that was transcribed in platelets but in very small amounts and 5) wild type proline 90. The results of family studies were consistent with the assumption described above.