Determinants of pigeons' waiting time: Effects of interreinforcement interval and food delay.

Manabe (1990) worked with pigeons on variable-interval schedules. The team changed two things at once: how long the bird waited for food after a peck, and how often food could arrive no matter what.

Each bird lived in a small box with one colored key. Pecks on that key sometimes paid off. The payoff could come right away or after a short delay. The next chance for food could arrive soon or later. The researchers timed how long each bird waited between bursts of pecking.

No single rule explained the waiting times. One pigeon waited longer when food delays grew, but only on certain schedules. Another bird barely changed its timing at all.

The mix of delay and overall food rate shaped each bird’s pattern. The authors call the result “subject-dependent.” In plain words: you have to watch the individual, not just the schedule.

Davis et al. (1972) saw response rates drop when any delay was added. Manabe (1990) agrees, but adds that the drop depends on the wider food schedule, not just the delay itself.

Corrigan et al. (1998) later showed that unsignaled delays hurt rates because birds stare at the hopper instead of pecking. Manabe (1990) set the stage by showing the delay effect is already messy before you remove the signal.

Barnard et al. (1977) pinned the safe window at four seconds. Manabe (1990) says even that window slides around once you factor in how often food can arrive for free.

When you write a delay-based program—token boards, points, or edible delivered later—do not trust a single fixed rule. Start with short delays, but also track how often reinforcement can occur overall. Watch the learner, not the clock. If waiting times drift, shorten the delay or tighten the overall schedule before behavior breaks down.