Triple microinjections of neurotensin (10 nmol each), which occupied a large volume of tissue within the nucleus reticularis gigantocellularis (Gi), produced an inhibition of the tail-flick reflex in awake rats. This effect was less than that previously observed by a single injection (10 nmol) into the nucleus raphe magnus (RMg) (see ref. ). Bilateral injections of neurotensin (10 nmol each) into the nucleus reticularis paragigantocellularis lateralis (LPGi) had no effect. The neurotensin antagonist [D-Trp11]-neurotensin (3 pmol) was previously found to enhance morphine, but not beta-endorphin antinociception from the periaqueductal gray (PAG) when injected into the RMg. A similar enhancement of morphine, but not beta-endorphin antinociception from the PAG was observed in the current study by [D-Trp11]-neurotensin injections into the bilateral LPGi, bilateral Gi, or medial Gi. These data suggest that neurotensinergic projections from the PAG function in an antianalgesic manner throughout the RVM during morphine, but not beta-endorphin antinociception. The antinociceptive effect of neurotensin, on the other hand, is more localized.