Plasma concentrations of THC were measured by gas-liquid chromatography and mass spectrometry following three routes of administration and correlated with clinical effects. Plasma concentrations peaked at 3 minutes after intravenous injection and then sharply declined. The peak "high" occurred at 30 minutes while plasma concentrations were declining. This lag between plasma concentration and "high" continued during most of the span of the drug's effects. The situation was quite similar following smoking, except that peak plasma concentrations were lower. After oral administration of THC, absorption was slow, with peak concentration occurring at 1 to 2 hours. Plasma concentrations were much lower. Correlations between plasma concentrations of drug and "high" were significant but not impressive. The degree of "high" was quite variable in relation to the prevailing plasma concentration. Conjunctival injection was found so long as plasma concentration of THC could be measured. Pulse rate increases occurred at lower concentration after oral administration than after the other two routes. It is unlikely that a range of plasma concentrations can be reliably equated with impaired performance. The mode of administration will become important should THC or some homolog become a therapeutic agent.