Although estimates of brain weight are useless in prediction of individual intelligence in man, the available data (none of them entirely adequate) suggest that the real correlation may be as high as 0.3. Evaluation of causes suggests the participation of a direct effect. Natural selection on intelligence at a current estimated intensity suffices to explain the rapid rate of increase of brain size in human evolution. Selection on birth weight may also suffice for this. It does not seem overly difficult to estimate directly the relationships among brain weight, intelligence, and fitness.