Interaural time differences (ITDs) were measured from 400--7000 Hz on cats in order to provide quantitative data for use in physiological/behavioral studies on sound localization. ITDs derived from clicks and the initial portion of tone bursts showed a pronounced roughness and frequency dependence. This frequency dependence is most evident at higher angles of incidence and indicates that a single ITD will not always represent a single position on the azimuth. Controls demonstrate that most of the roughness in these functions was due to reflections off the surface supporting the animal and that the measured ITDs corresponded to predictions made by steady-state theory. Measurements made with and without the pinnae in position indicate that they have relatively little effect on these ITD functions, particularly for frequencies below 2500 Hz and for small angles of incidence. In spite of acoustic limitations exemplified by the roughness and frequency dependence of these functions, ITDs generated by sound sources situated close to the midline provide reliable localization cues that are much better than those derived from sources well out on the azimuth. Finally, it is noted that another ITD, the group ITD, can be ascribed to an acoustic signal. Calculations based on the measured steady-state ITDs show differences between the group and steady-state ITDs over a given range of frequencies. Differences between the group and steady-state ITD can be significant, and it is argued that: (1) The group ITD can provide a localization cue to the auditory system that is distinct from the steady-state ITD; and (2) it is possible these group ITDs are used by the nervous system to localize sound sources in realistic situations.