It has for long been known that in industry tasks involving the prolonged repetition of a single short operation throughout the day tend to lead first of all to an increased variation in the time taken to perform the task, and later to a general decrease in rate of work (Wyatt and Fraser, 1929, 1937). The rate tends to be lowest at the middle of the working spell, and to increase towards the end. Feelings of boredom are often experienced. But these effects could be due to the monotony of the actions performed, rather than to perceptual repetition (see McGrath, 1963b). The effects of repetition on perception were first investigated experimentally by Mackworth (1950). He presented to his observers, for a period of two hours, a type of clock face with a pointer moving round it in small regular jumps. At fairly infrequent and irregular intervals, the pointer made a double jump; and the observer was required to signal each occurrence of a double jump by pressing a key. After about half-an-hour observers began to miss the double jumps; and the number of omissions increased throughout the two-hour period. Several varieties of task produced similar results: observing a regular series of flashes of light, with an occasional brighter flash (Bakan, 1955);

detecting signals lit up by a sweeping line of light on a simulated radar screen (Mackworth, 1950; Baker, 1956); reporting the appearance of a circle of unusual size among a number of other circles (Fraser, 1957)·

The decrease in efficiency in performing these tasks was attributed to a decline in the observer's vigilance or attention to the task. Though it occurred in several different tasks, decline in efficiency varied considerably with the nature of the task. In general, it was less when discrimination between signals and non-signals was easy, perhaps because arousal is greater when there is a strong contrast with surroundings. Thus using his type of radar task, Mackworth (1950) found little decline in performance with relatively bright signals. Fraser (1957) obtained a similar result when there was a large difference in size between signal and non-signal circles. Fraser also found that there was no decline in performance when the exposure of the signals was increased from one to two seconds.