Erythrocyte components released from blood clot after subarachnoid hemorrhage are thought to be involved in the pathogenesis of cerebral vasospasm. The effect of erythrocyte components on smooth muscle has been investigated extensively; however, their action on endothelium remains unclear. We studied the effects of different erythrocyte components and bloody cerebrospinal fluid (CSF) on [CA++]i in cultured bovine pulmonary and cerebral arterial endothelial cells using fura-2 [Ca++]i microfluorimetry. Erythrocyte lysate and bloody CSF produced a biphasic [Ca++]i response, a peak and a plateau, and the effect of erythrocyte lysate on [Ca++]i was attenuated by the endoplasmic reticulum Ca++ pump inhibitors thapsigargin and cyclopiazonic acid, by the voltage-independent Ca++ channel blocker SK&F96365, by the P450 cytochrome inhibitors econazole and miconazole and by the inorganic Ca++ pathway blockers lanthanum, nickel and cobalt. This suggests that erythrocyte lysate releases Ca++ from internal stores and promotes Ca++ influx from voltage from voltage-independent Ca++ channel. Erythrocyte lysate was then fractionated into different molecular weight fractions with different pore-size membranes and each fraction was tested separately. The component that increased [Ca++]i had a molecular weight < 1 kdalton. Fractions that contained oxyhemoglobin did not affect [Ca++]i. Adenosine nucleotides mimicked the effect of erythrocyte lysate and bloody CSF on [Ca++]i. The P2-purinoceptor antagonist suramin attenuated the effect of ATP, erythrocyte lysate, the fraction < 1 kdalton and bloody CSF. We concluded that adenosine nucleotides are the component of erythrocyte lysate and bloody CSF that increase [Ca+2]i in endothelial cells and that the effect involves P2-purinoceptors. However, no evidence is presented that this increase in endothelial [Ca+2]i by erythrocyte lysate causes cerebral vasospasm.