AbstractRats were trained in a water maze in a dark room with the extramaze cues restricted to only dimly back‐lit shapes. We used lidocaine to reversibly lesion the dorsal hippocampus and this controlledcue room in order to examine interhippocampal synthesis of lateralized place engrams. Experiment 1 showed that lidocaine injected into both hippocampi effectively abolished place navigation for up to 25 min but not at 45 min. In experiment 2, each day under lidocaine blockade of one hippocampus, pretrained rats were trained in the water maze to locate the target according to two cues (e. g., AB). Two hours later, the contralateral hippocampus was inactivated and the rats were trained to the same location with two other cues (CD). On day 5, intact brain retrieval was tested in one of three conditions: ACQ (e. g., AB), one of the pairs of cues used in acquisition training; SYNTH (e. g., AC), one cue from each of the pairs used in acquisition; CONT (e. g., AE), one cue that was used in acquisition training and a novel cue. The results show that the hippocampi learned the two tasks independently and similarly [latency (L) at the asymptote = 7 s]. Retrieval performance was at the asymptote for ACQ (AB) and SYNTH (AC) (L = 6 and 7, respectively) but was disrupted for CONT (L = 12). In experiment 3 as in experiment 2, the rats were trained, under unilateral blockade, to a new place for 4 days. On day 5, retrieval with the trained hippocampus blocked was worse (L = 11) than with the untrained side blocked (L = 5). We conclude that interhippocampal synthesis of lateralized place engrams is automatic and highly efficient. © 1995 Wiley‐Liss,