The effects of various 5-HT3 receptor antagonists were examined in the social interaction (SI) test following discrete microinjection into either the dorsal raphe nucleus (DRN) or amygdala of the rat. Following DRN injection, ondansetron, ICS205-930, and MDL72222 (5-500 ng) all failed to modify SI under high light/unfamiliar (HLU) test conditions relative to vehicle pretreated controls. The 5-HT3 receptor agonist, 2-Me 5-HT (100-2500 ng), was similarly ineffective under both HLU and low light/familiar (LLF) conditions, although 5-HT (20-100 ng) increased SI under the HLU paradigm. After amygdaloid injection, ondansetron (10-100 ng), granisetron (1-10 ng), ICS205-930 (10-100 ng), GR 65630 (1-10 ng), and MDL72222 (100-1000 ng) all significantly increased SI under the HLU but not LLF condition. Furthermore, a detailed behavioural analysis revealed that the behaviours underlying this increase were similar to those seen in vehicle pretreated animals tested in the LLF compared to HLU condition. The benzodiazepine, flurazepam (200 ng), increased both SI (HLU condition) and punished responding in a modified water-lick conflict model, after amygdaloid injection. Both ondansetron (10-1000 ng) and ICS205-930 (1-100 ng) were ineffective in the conflict test. Finally, 2-Me 5-HT and 5-HT (100-10,000 ng) reduced SI under the LLF test condition with no concomitant change in locomotor activity.(ABSTRACT TRUNCATED AT 250 WORDS)