An experiment to assess the value of a shadowgraph simulator for driver training is described. Two groups of 36 pupils with no previous driving experience, and matched for age and sex, were taught to drive up to the standard of the DOE driving test. The pupils in one of the groups received their first three hours instruction on a shadowgraph simulator. The results of the experiment showed that there was no significant difference between the time required by those pupils who had started on the simulator, and those who had had all their lessons in a car. There was a wide variation in the amount of time required to pass the driving test: with extremes of 14 and well over 60 hours. The average number of hours of instruction required was 31. An intelligence test, a test of Mechanical Comprehension and a Personality Inventory were administered to each pupil. There was some evidence to suggest that the more intelligent pupils learned to drive more quickly. (Author)