Using a conditioned suppression procedure, the effects of three contingent relationships between conditioned (CS) and unconditioned (US) stimuli were investigated. A traditional positive (if CS-then US) contingency suppressed response rate during the CS relative to responding during stimulus-free minutes of the session. A negative (if CS-then no US) contingency resulted in suppressed responding during CS-off minutes, and rate increases during the CS. A no-contingency control procedure, during which CS and US were randomly related, almost totally suppressed responding throughout the session and showed no differential effects of the CS on response rate. An analysis of changes in response rate during the minute after US-offset revealed acceleration under the no-contingency condition and, to a somewhat lesser degree, under the negative contingency. Both conditioned suppression and non-suppression are analyzed in terms of the temporal relationship between CS and US.In a recent theoretical paper, Rescorla <1967) argued that the contingency between the conditioned stimulus (CS) and unconditioned stimulus (US), rather than the pairing or simple contiguity between these two events, is a critical determinant of Pavlovian conditioning. In the language of the system proposed by Rescorla, the majority of Pavlovian conditioning research involves a positive contingency (i.e., if CS-then US) stimulus arrangement and results in excitatory conditioning. A Thompson and McConnell, 1955;Bitterman, 1964), but rather in presenting the CS and US in random relationship to each other so that no CS-US contingency exists. Under this control procedure, CS and US may occasionally be paired on a particular trial, but such a pairing occurs strictly under a random distribution of events and predicts nothing about the order of events on subsequent trials. Furthermore, such a procedure provides a control for both excitatory and inhibitory aspects of Pavlovian conditioning.A recent survey (Davis, 1968) of research involving the conditioned suppression or conditioned emotional response procedure of Estes and Skinner (1941) indicates that this procedure shares the emphasis on the positive contingency and excitatory conditioning typical of much Pavlovian conditioning research in this country. Conditioned suppression is frequently analyzed in terms of the elicitation of respondents (the "fear" CR) during the CS. Fear is presumed to be incompatible with the baseline operant, which is thereby suppressed (e.g., Hunt and Brady, 1951;Kamin and Schaub, 1963