Five experiments examined the factors that determine whether or nor summation will occur when experimental stimuli are presented on a television screen for autoshaping with pigeons. In Experiment 1 there were periods when the television screen was illuminated white and periods when it was dark. When conditioning was conducted during the white periods, then summation was observed if the test trials were conducted with a white screen during the intertrial interval. Summation was not found when conditioning was conducted during the dark periods, no matter how testing was conducted. The remaining experiments demonstrated that summation will occur if at least some of the conditioning trials are conducted with the experimental stimuli being accompanied by white illumination in the remainder of the television screen. It was not necessary for this illumination to persist into the interval between trials. These findings indicate that summation is influenced by the similarity of the patterns of stimulation that are used for training and testing.