Comprehensive or nonselective schools have been in existence in the United Kingdom for over 40 years, having gradually replaced the old selective system of grammar schools for able pupils and secondary modern schools for less able pupils. In spite of their growth in numbers, comprehensive schools have been subject to continual controversy, and current government policies favor a return to forms of pupil selection and differentiation. This paper therefore begins with a review of the history of comprehensive schooling and the rationale for its adoption, and then assesses existing research into the effectiveness of selective and nonselective systems. The third section evaluates some of the effects of comprehensive schools on their pupils by examining an "experiment of nature" that occurred when a community in Wales partially reorganized its schools, leaving a third of its pupils in a selective system and putting two-thirds into new nonselective comprehensive schools. A wide range of data were collected on the pupils entering the two systems, including personality characteristics, verbal ability, mathematical ability, and nonverbal ability or intelligence. Findings are tabulated and discussed for (1) cohort intake data for selective and comprehensive systems; (2) actual and predicted outcome scores for the two systems; (3) residuals and standard deviations; (4) catchment areas of the two systems; and (5) outcome results by ability groupings. Results indicate that the significantly poorer academic and social performance of comprehensive schools is largely due to poor results with the middle third of the ability range who in the selective system attended the top streams of the secondary modern schools. A bibliography is included. (TE)