1. A comparison was made between the influences of supramammillary (SUM) and medial septal (MS) nuclei on hippocampal physiology in Nembutal-anesthetized rats. Specifically, the effects of prestimulation of the SUM or MS on the perforant path-dentate field potential, on spontaneous activity of single units, and on perforant path-induced unit activation were assessed. Another series of experiments addressed the issue of whether the SUM and MS effects on the perforant path-dentate field response are independent. 2. Prestimulation of the SUM or MS significantly facilitated the perforant path-dentate population spike with no clear effect on the field excitatory postsynaptic potential (EPSP) recorded in the subgranular zone of the dentate hilus. Prestimulation of either nucleus also reduced the threshold for spike onset. The major differences between the two spike facilitation effects were the magnitude of the change and possibly the optimal interstimulus intervals required to obtain the effects. 3. Acute transection of the ipsilateral column of fornix or dorsal fornix eliminated the SUM population spike facilitation effect. MS lesion or dorsal fornix/fimbria transection eliminated the MS spike facilitation effect. The MS lesion did not alter the effects of SUM prestimulation. Cingulum or medial forebrain bundle transection affected neither SUM- nor MS-mediated spike facilitation. Thus the SUM and MS influences on the dentate field response appear to be independent of one another. The relevant SUM afferents travel through the ipsilateral column of fornix and dorsal fornix, whereas MS afferents project through the dorsal fornix/fimbria. 4. Single units recorded in stratum granulosum (SG) were assessed with respect to several parameters. These included the mean firing rate, whether or not excitation occurred prior to the field population spike and at lower threshold, and whether or not a driven unit responded to a second perforant path stimulus delivered at short latency following the first (during the period of population spike depression). The latter parameter in particular appeared to separate SG cells into two classes. The cells that were not activated during the second field potential were classified as granule cells, whereas those that were activated were classified as basket cells. Based on this distinction, significant differences were also found between the two cell classes on the other parameters. In particular, cells classified as granule cells often had very low firing rates.(ABSTRACT TRUNCATED AT 400 WORDS)