Interactive Effects of Fixed-Ratio Size and Number of Food Pellets Per Fixed Ratio on Rats’ Food-Reinforced Behavior

Abstract

In three experiments, albino rats’ lever pressing was reinforced with food pellets under fixed-ratio reinforcement schedules during daily 3-hr sessions. In Experiment 1, fixed-ratio size was varied while one pellet was delivered at the completion of each ratio. Overall response rate was an inverted U-shaped function of fixed-ratio size. As fixed-ratio size was increased, the number of pellets obtained per session remained fairly constant at low and moderate response requirements, then decreased at higher requirements. In Experiment 2, the number of food pellets available under a continuous reinforcement schedule following completion of each fixed-ratio component was varied, while fixed-ratio size was held constant at the highest value reached by each rat during Experiment 1. With increases in the number of pellets available (i.e., In reinforcer magnitude) per fixed ratio, mean numbers of pellets delivered per session increased while mean overall response rate decreased. In Experiment 3, reinforcer magnitudes ranging up to 512 pellets per fixed ratio were delivered at fixed-ratio sizes ranging from 128 to 2048 responses. At each fixed-ratio size, mean overall response rates were inverted U-shaped functions of reinforcer magnitude, while mean number of pellets delivered was related directly to reinforcer magnitude. As fixed-ratio size was increased, progressively larger reinforcer magnitudes maintained the maximum overall response rate. Running rates were uniform across a wide range of reinforcer magnitudes and fixed-ratio sizes. Finally, across each doubling of fixed-ratio size, the relative persistence of behavior increased directly with the magnitude of the reinforcer. The results of Experiment 3 demonstrated that reinforcing effects increased with increasing reinforcer magnitude. Differences in the response-rate results of the three experiments suggest that in many cases reinforcing effects may be more accurately assessed by nonresponse-rate measures.