The sniff was identified by a brief episode of increased respiratory rate, usually with a well-defined time of onset. It was detected against the background of respiratory activity in rabbits simply, reliably, and noninvasively by statistical evaluation of digitized pneumograph records. The basal rate of exploratory sniffing was controlled by familiarization. Upon conditioning to olfactory cues, the rate of sniffing for CS+ increased sharply above the basal rate during the first 10 trials and was maintained at high levels by continued reinforcement. During extinction with discrimination between olfactory cues, the rate for CS- fell sharply at first and then more slowly toward the basal rate. With pseudoconditioning, the rabbits responded to an unpaired odor after several sessions; the rates of response acquisition and extinction and the maintained level of responding were lower than with a paired odor in classical delayed conditioning, and the response was not discriminative in respect to another novel odor given during extinction. The sniff displayed a prominent sensory bias for olfactory cues. The relative frequencies of sniffing and respiratory slowing were measured as conditioned responses by screening procedures with a small computer.