Food-deprived pigeons were exposed to a one-key observing response pro
cedure. Discrete trials ended either with or without response-independ
ent food presentation (p = 0.5); trials were prolonged until no respon
ding had occurred for at least 3 s. Key pecks (observing responses) co
uld produce discriminative stimuli correlated with the trial outcome,
either S+ (food) or S- (no food). Under some conditions, pecks produce
d both stimuli on a 15-s variable-interval (VI) schedule. Under other
conditions, a minimum 6-s interresponse time was required after comple
tion of the VI (tand VI DRL) to produce one of the stimuli. For all bi
rds, the tand VI DRL contingency resulted in a much greater and more p
rolonged decrease in stimulus production when it was applied to S-tria
ls than when it was applied to S+ trials. The results are consistent w
ith conditioned reinforcement theories such as the delay-reduction the
ory. (C) 1998 Academic Press.