AFTER REVIEWING SOME CRITICISMS OF THE TWIN METHOD, PARTIAL REPLIES TO THESE ARE SUGGESTED. STATISTICAL METHODS FOR THE ANALYSIS OF TWIN DATA ARE DISCUSSED, INCLUDING SOME NEW IDEAS ABOUT MULTIVARIATE ANALYSIS OF TWIN DATA. THE REMAINDER OF THE PAPER SUMMARIZES RECENT TWIN STUDIES ON PHYSIOLOGICAL AND SENSORY VARIABLES, PERCEPTUAL TASKS, MOTOR SKILL, COGNITIVE VARIABLES, AND PERSONALITY. RECOMMENDATIONS FOR FUTURE RESEARCH IN HUMAN BEHAVIOR GENETICS ARE GIVEN. (4 P. REF.)