Four rats and four pigeons responded for food delivered by variable in
terval schedules that provided programmed rates of reinforcement rangi
ng from 15 to 480 reinforcers per hour. Rate of responding increased,
decreased, or increased and then decreased within sessions. The within
-session pattern of responding changed with changes in the programmed
rate of reinforcement and with the species of subject. Finding within-
session changes in responding during variable interval schedules exten
ds the generality of these changes to another schedule. It implies tha
t variable interval schedules should be used cautiously as baselines f
or assessing the effects of other variables, such as drugs. Finally, t
he results suggest that systematic pauses in responding during the ses
sion may contribute to the decreases in the average rates of respondin
g with increases in the rates of reinforcement that are sometimes obse
rved when subjects respond on variable interval schedules that deliver
high rates of reinforcement.