In the ectosylvian cortex of 24 barbiturate-anesthetized cats, area AI was identified by its frequency organization and the responses to tonal stimuli of single neurons in that field were examined using sealed stimulating systems incorporating calibrated probe microphone assemblies. The responsiveness to monaural and binaural best-frequency stimuli was examined quantitatively for 282 single units in AI. One hundred thirty-nine cells (49%) were excited by independent stimulation of only one ear and were classified as EO cells. In general, the effective monaural excitatory input was derived from the contralateral ear. One hundred ten (39%) neurons were excited by independent stimulation of each ear and were classified as EE units. For these neurons, the contralateral responses were generally stronger, shorter in latency, and lower in threshold than were their ipsilateral responses. Thirty-three cells (12%) gave weak or no responses to monaural stimuli but responded securely to binaural stimuli. These cells were classified as predominately binaural (PB). Binaural interactions were examined by comparisons of the response to binaural, equally intense stimuli to the stronger monaural response. Among EO cells suppression was the most common form of interaction, while for EE cells summation was the more common. Less than 8% of cells were found to be monaural. In electrode penetrations radial to the cortex surface, cells received their stronger or sole monaural excitatory input from a common ear, generally the contralateral. Within such penetrations, however, cells commonly differed with regard to the nature of their input from the other ear and/or in their binaural interactions. Comparison of these data with data previously reported for subcortical auditory nuclei revealed that AI preserves many of the stimulus specificity characteristics of the lower nuclei. The reasons for the preservation of these characteristics at the cortex and the implications of the present data for the binaural column hypothesis are discussed.