Nicotine and other nicotinic receptor agonists have been found in a variety of studies to improve memory, while nicotinic receptor blockade can impair memory. The critical neural mechanisms for nicotinic involvement with memory are still under investigation. Initial evidence supports the involvement of the ventral hippocampus. Lesions in this area block nicotine-induced memory improvement and mecamylamine-induced impairment. Local ventral hippocampal application of the nicotinic channel blocker mecamylamine impairs memory in the 8-arm radial maze. Both alpha 4 beta 2 and alpha 7 nicotinic receptors seem to be involved. Ventral hippocampal infusions of high doses of the alpha 4 beta 2 nicotinic antagonist dihydro-beta-erythrodine (DH beta E) and the alpha 7 nicotinic antagonist methyllycaconitine (MLA) impair memory performance on the 8-arm radial maze. However, high doses of these drugs may limit specificity and they cause preconvulsant effects, which in themselves may affect memory. The current study used the more challenging 16-arm radial maze to determine the effects of lower doses of these drugs on memory and to differentiate effects on working and reference memory. Adult female Sprague-Dawley rats were trained on a working and reference memory task in the 16-arm radial maze and then were implanted with bilateral chronic guide cannulae directed to the ventral hippocampus. After recovery from surgery, the rats received acute intrahippocampal infusions of dose combinations of DH beta E and MLA. In the first study, DH beta E (0 and 6.75 microg/side) and MLA (0, 6.75, 13.5 and 27 microg/side) were administered in a counter-balanced order. In the second study, lower doses of DH beta E (0, 1.6375, 3.275 and 6.75 microg/side) were administered alone or with MLA (0 and 6.75 microg/side) in a counter-balanced order. In the first study, DH beta E caused a significant increase in both working and reference memory errors. MLA at a dose of 27 microg/side caused a significant increase in working memory errors, but this dose had no significant effect on reference memory errors. Interestingly, no additive effects were seen with combined administration of DH beta E and MLA in this study, and at the doses used, no effects were seen on response latency. In the second study, lower doses of DH beta E did not cause a significant deficit in working memory performance. Co-administration of MLA with these subthreshold doses did precipitate a memory impairment. The current results confirm the specificity of the memory deficits caused by these drugs. These results support the involvement of alpha 4 beta 2 and alpha 7 nicotinic receptors in the ventral hippocampus as being critical for memory function.