Oral delta 9-tetrahydrocannabinol (THC) in gelatin capsules is under evaluation as an antiemetic agent in cancer patients, but knowledge concerning its bioavailability is incomplete and, furthermore, alternative routes of administration may be desirable. In this study, the disposition of THC was determined in four rhesus monkeys given 2.5-mg/kg doses using the following routes of administration and formulations: intravenous (iv); orally (po) on a cookie and in gelatin capsules; intramuscularly (im) in Tween-80 and in Emulphor-EL620; rectally in various suppository bases. Serum THC concentrations were measured by RIA and analyzed by weighted nonlinear regression. Serum concentrations were best described by a sum of two exponentials with alpha and beta half-lives (mean +/- SD) of 0.74 +/- 0.59 and 14.9 +/- 12.5 h. Apparent bioavailability (%F +/- SD) of various formulations of THC were: gelatin capsules, 26 +/- 14; cookie, 89 +/- 16; intramuscularly in Tween-80 and in Emulphor, 39 +/- 13 and 102 +/- 15, respectively. Using the method of statistical moments, mean residence times in the body (h +/- SD) were: intravenous, 6.08 +/- 1.60; cookie, 21.92 +/- 3.11; gelatin capsule, 26.80 +/- 23.61; intramuscularly in Emulphor, 10.92 +/- 3.46 (in Tween-80, not calculated). THC was not bioavailable by the rectal route. We conclude from this study that THC formulated as a gelatin capsule exhibits a low and variable extent of bioavailability and that intramuscular THC may be a useful alternative route of administration since it is more completely bioavailable.