Pigeons' rates of responding on simple schedules appearing alone or as components of signal-key multiple schedules were not systematically different early in training, but were different later in training. This suggests that a simple schedule may be an appropriate baseline from which to measure behavioral contrast. Positive behavioral contrast, like the present differences between simple and multiple schedule responding, does not appear when naive subjects respond on signal-key multiple schedules, but does appear when experimentally experienced subjects are used.