Cholecystokinin octapeptide (CCK-8), administered intracerebroventricularly (i.c.v.), will suppress feeding. The aim of the present study was to determine the pharmacological characteristics of this satiety inducing effect in rats. For this purpose, we employed a feeding bioassay model in 24 h fasted rats and examined the effects of CCK-8 and a variety of structurally related analogs on latency to feed after i.c.v. injection and on the amount of food and water consumed as measured after the initiation of feeding in sequential 20-min epochs for 1 h. CCK-8, given in doses of 0.1, 1 and 10 nmol, produced a dose-dependent increase in feeding latency and a reduction of food intake during the first 20 min after initiation of feeding. Food intake during the next 40 min and water consumption were not altered. Plasma levels of CCK-like immunoreactivity after an i.c.v. injection of a dose of CCK-8 which blocked feeding (10 nmol) rose insignificantly from 117 to 125 pg/ml. In contrast, at the minimally effective dose of CCK-8 after i.v. administration (10 nmol), which also produced an inhibition of feeding, the plasma level was 1430 pg/ml. This difference indicates that plasma levels of CCK after i.c.v. CCK-8 are not adequate to produce the observed feeding suppression and suggests that the effects of i.c.v. CCK-8 are not mediated by a peripheral redistribution. Systematic dose response studies revealed the following rank order of potencies: CCK-8 greater than or equal to G-17 II much greater than CCK-8 NS = G-17 I greater than or equal to CCK-4 = CCK 26-29 = 0. Only gastrin-17 II (sulfated) produced an effect comparably significant to CCK-8. I.c.v. proglumide at 2500 nmol failed to modify the effects of CCK-8 at 10 nmol after i.c.v. injection. These data demonstrate that the structural requirements for feeding suppressive activity in rat brain are the carboxyterminus with a sulfated tyrosine residue, located 6 to 7 residues from the carboxyterminus, as present in CCK-8 and gastrin-17 II.