Hippocampal field potentials evoked by paired-pulse perforant path stimulation were used to identify normal feedforward and feedback inhibitory influences on hippocampal principal cells. Three distinct aspects of inhibitory function were identified in the dentate gyrus. They are: (1) first spike amplitude-dependent inhibition of the second spike, which at low stimulus frequency is primarily feedback in nature; (2) frequency-dependent inhibition of a single spike or the first spike of a pair, which occurs as stimulus frequency is increased from 0.1 to 1.0 Hz and which is primarily a reflection of feedforward inhibition; and (3) frequency-dependent inhibition of the second spike that is relatively independent of first spike amplitude and probably due to a combination of feedforward and feedback mechanisms. The results indicate that granule cell recurrent inhibition alone, evoked at low stimulus frequency, is relatively brief and weak. At higher frequencies, probably more relevant to physiological activity, feedforward inhibitory activity is recruited. The combination of feedforward and feedback mechanisms results in strong, maximal duration, granule cell inhibition. Similar frequency dependence of inhibition was not seen in area CA1 in response to ipsilateral perforant path stimulation since low frequency stimulation did not evoke CA1 spikes. CA3 stimulation did evoke large contralateral CA1 population spikes, but paired-pulse inhibition was weaker than that evoked by ipsilateral perforant path stimulation in terms of the duration of inhibition and the ability to suppress the development of epileptiform behavior. The identification of simple tests that reflect distinct inhibitory processes in vivo permits similar studies to be conducted in vitro to determine how to preserve inhibitory processes for cellular studies of normal and human epileptic tissue in which the state of excitatory--inhibitory balance is the subject. These results also form the basis for the interpretation of the following study (Sloviter, 1991), which addresses the relationship between selective dentate interneuron loss and the pathophysiology that accompanies it.