Summary.1.—This experiment was undertaken to test the hypothesis that the attainments, attitudes, behaviour and personalities of a group of pupils involved in an organization based upon streaming are influenced by that organization.2.—The experiment involved two groups, each of ninety pupils, entering the same selective central school at the age of eleven years. The control group was organized into three forms whose membership did not change during the two‐year period following entry to the school. The experimental group was organized into three streams and pupils were transferred between streams after each half‐yearly examination.3.—Taken together the results of the ability tests showed no significant differences between the groups attributable to the organization based on streaming.4.—The attitude tests towards examinations, school lessons and school life in general yielded no significant differences between groups.5.—The samples of classroom behaviour revealed that in the group organized into streams fewer social contributions to lessons were made by pupils, that there was more aggressive behaviour and less attention to work.6.—The estimates of personality made by the teachers revealed no significant differences between the groups. The pupils' self‐estimates revealed an extensive, but probably temporary, deterioration in personality following re‐grouping.7.—Studies of several pupils indicated that the more lasting effects of transfer are a highly individual matter. No general long‐term effects attributable to st