The aggregation of human blood platelets is inhibited when membrane disulphide bonds are reduced (Born, 1967). Exposure of blood platelets to U.V. light induces platelet aggregation (Dickson et al., 1971) possibly also by the reduction of disulphide bonds (Doery et al., 1973). By using dithiothreitol as a reducing agent, we have confirmed that low concentrations inhibit platelet aggregation but at higher concentrations dithiothreitol itself induces aggregation. Platelet-rich plasma was prepared from human and rat blood by differential centrifugation, and platelet aggregation was measured photometrically in 0.1 ml samples (Gordon & Drummond, 1974). Suspensions of washed human platelets were prepared from 2ml volumes of platelet-rich plasma by using an albumin density gradient (Walsh, 1972). Total adenine nucleotides and 5-hydroxytryptamine in platelets and in plasma or suspending medium were measured fluorimetrically (Gordon & Drummond, 1974; Drummond & Gordon, 1974) before and after platelet aggregation to determine the extent of the platelet-release reaction. Lactate dehydrogenase was measured by the method of Wroblewski & La Due (1955) to check for possible lysis of platelets. Platelets aggregate when collagen or ADP is added to stirred platelet-rich plasma. The response to collagen is characterized by a lag of several seconds duration, followed by irreversible aggregation caused by release of amines and nucleotides from the platelets. In contrast, the aggregation response to low concentrations of ADP is rapid, reversible and not associated with the release of platelet constituents. The addition of dithiothreitol to platelet-rich plasma samples inhibited aggregation induced by both ADP and collagen, but dithiothreitol was less effective against ADP. The I.C.50 value (concentration of dithiothreitol inhibiting aggregation by 50%) was 0 .45m~ for collagen and 2 .15m~ for ADP. In each case, a half-maximal aggregating stimulus was used and dithiothreitol was preincubated in platelet-rich plasma for 2min at 37°C before adding the aggregating agent. The addition of 1 m-dithiothreitol abolished collagen-induced aggregation in human platelet-rich plasma and also completely inhibited the accompanying plateletrelease reaction : collagen released 34 % of platelet 5-hydroxytryptamine in control samples, but caused no release in samples containing 1 m-dithiothreitol. The inhibitory potency of dithiothreitol was similar in human and in rat platelet-rich plasma. When samples of platelet-rich plasma were preincubated with dithiothreitol at 37°C before addition of the aggregating agent, inhibition increased with time, reaching a maximum after 10-2Omin. When dithiothreitol was added to human platelet-rich plasma in concentrations of 3 m ~ and above, it caused platelet aggregation after a prolonged lag phase (lamin). The duration of the lag decreased and the rate of aggregation increased with increasing dithiothreitol concentrations. In suspensions of washed human platelets dithiothreitol did not produce aggregation unless fibrinogen was also added. Even when fibrinogen was present there was always a lag of about 2min between the addition of dithiothreitol and the start of aggregation. When dithiothreitol had been incubated with the platelet suspension for 2min or longer in the absence of fibrinogen, the subsequent addition of fibrinogen caused immediate aggregation. The lag before dithiothreitol-induced aggregation was due to an action of dithiothreitol on the platelets and not on the fibrinogen, since preincubation of dithiothreitol with fibrinogen did not alter the duration of the lag (Fig. 1). Aggregation of human platelets by dithiothreitol was associated with the release of 10-15 % of platelet 5-hydroxytryptamine and adenine nucleotides. No release of lactate dehydrogenase from the platelets could be detected during aggregation by dithiothreitol, suggesting that the release of amines and nucleotides was selective and not caused by platelet lysis. When disulphide bonds are reduced by dithiothreitol, there is concomitant formation