Eighty-three cells in the superior temporal gyrus of awake squirrel monkeys were tested with 12 different species-specific vocalizations; 75 (90%) responded to one or more calls. A wide range of selectivity and response probabilities were seen. Eighty-nine percent of the cells responded to more than half of the vocalizations. In only one case did a cell respond to only one call type ('peep'). This cell also responded to steady tone bursts, but only in a narrow range of frequencies approximating the major band of acoustic energy in peeps. In most cases, no obvious acoustic relationships were seen between calls effective for a given cell. Likewise, in most cases the frequency range over which a cell responded to steady tone bursts did not readily explain responses to vocalizations. More than half of the cells discriminated between one or more pairs of acoustically similar vocalizations. The response patterns of different cells to the same vocalization showed considerable variety. This was especially true for responses to the temporally more complex calls. The variety of responses given by different cells to the same vocalization suggests that no vocalization is coded by one unique discharge pattern. Our results suggest that many neurons in the STG do not select between different classes of vocalizations according to presence or absence of simple acoustic features.