1. The oral sedative potencies of cannabis herb, crude ethanolic and petroleum-ether fractions, were assayed against delta'-trans-tetrahydrocannabinol (THC) administered orally to mice, by measuring spontaneous motor activity over 30 min periods, at selected times, up to 6 h. 2. The THC contents of the extracts were determined chemically by gas-liquid chromatography analysis and the B/C ratio (biological activity divided by chemical activity) calculated for each. The B/C values for cannabis herb, which contained THC but no CBD, was 4.47 and for ethanolic and petroleum-ether extracts, 5.26 and 4.39, respectively. 3. The sedative potency expressed as SDA50, the dose required to give 50% effect over 6 h, was 1.06 (0.98 to 1.15) mg/kg for THC; 4.72 (4.22 to 5.27) mg/kg for cannabidiol and 1.26 (1.22 to 1.80) mg/kg for chlorpromazine. 4. An infusion of cannabis herb made with boiling water was shown to have sedative activity of very low potency. 5. When the cannabinoids were completely extracted from a sample of herb with petroleum-ether the aqueous and ethanolic extracts of the marc had some sedative activity; but the 70% ethanolic fraction had none. 6. The sedative activity of THC, cannabis herb and a water soluble fraction is blocked by aspirin, a cyclo-oxygenase inhibitor, and restored by prostaglandin E2 (PGE2). 7. The sedative effect of chlorpromazine is not blocked by aspirin.