Single doses of naloxone (0.025 to 0.5 mg/kg) or of one of four quaternary narcotic antagonists (i.e. nalorphine allobromide, nalorphine methobromide, naloxone methobromide or naltrexone methobromide, 1 to 60 mg/kg) were given s.c. to rats before morphine, 5 mg/kg i.v. In the absence of antagonists morphine reduced G.I. transit of a charcoal meal to about 15% of drug-free controls and consistently delayed nociceptive reactions (55 degrees C hot plate) in all animals. Doses of antagonists slightly reducing morphine antinociception (centrally effective = A) and restoring G.I. transit to about 50% of drug-free rats (peripherally effective = B) were estimated. The A:B ratio, indicating peripheral selectivity, was at least 8 for any of the quaternary antagonists given 10 min before morphine, but prolonging this interval may have resulted in a lower figure (i.e. less peripheral selectivity) because of reduced A and increased B. This was definitely so for naltrexone methobromide (A:B, greater than 60 at 10 min, about 1 at 80 min) and was not apparent for nalorphrine methobromide according to available data, which for nalorphrine allobromide and to a lesser extent for naloxone methobromide showed only an increase in B at intervals longer than 10 min. Both morphine-induced antinociception and inhibition of G.I. transit were reduced by naloxone at the lower doses tested and were fully prevented at the higher. These findings indicated that, unlike naloxone, the investigated quaternary narcotic antagonists are interesting prototype drugs for selective blockade of opiate receptors outside the CNS, although certain critical aspects, possibly biological N-dealkylation to the corresponding tertiary antagonists, condition peripheral selectivity.