Behavioral contrast can be defined as an inverse relation between rate
of responding in one component of a multiple schedule and the conditi
ons of reinforcement in the other component. The present study was an
attempt to produce contrast by changing the probability of reinforceme
nt. Four pigeons pecked a key for mixed-grain reinforcers delivered by
a two-component multiple schedule. In each component, reinforcers wer
e scheduled by a variable-interval schedule and were delivered at a ce
rtain probability (i.e., a percentage reinforcement procedure). In con
trast conditions, the probability of reinforcement in the second compo
nent was either increased or decreased. In some contrast conditions, n
onreinforcement in the second component was signaled and in other cond
itions it was not. Nonreinforcement was always signaled in the first c
omponent. The changes in probability of reinforcement produced contras
t. However, the contrast produced by changes in probability of reinfor
cement could not be differentiated from the contrast produced by chang
es in the rate at which reinforcers were delivered. The present result
s demonstrate that behavioral contrast can occur when a percentage rei
nforcement procedure is employed. Furthermore, although the results do
not suggest that probability of reinforcement can be varied independe
ntly, they do extend the generality of contrast and its potential prac
tical implications. (C) 1997 Academic Press.