Self-identification by racial/ethnic membership was requested, pursuant to a court order, from 1,373 applicants for professional and administrative employment in the federal service. Of the 1,348 for whom test answer sheets could be identified (760 males, 586 females, 2 unspecified), 139 declined to provide information about racial/ethnic membership and 155 omitted responding. Abilites assessed were verbal, judgment, induction, deduction, and number. Test scores were analyzed by sex for the eight largest groups identified. The results were in accord with most recent research on this subject: distinctively higher scores on induction, deduction, and number than on verbal and judgment for Asians; fairly flat profiles for whites, blacks, and Mexican Americans; and lower scores for females on judgment and number. Of interest was the similarity between black and white profiles, between two Asian profiles, and the superior performance of those who explicitly declined to state their racial/ethnic membership.In recent years the performance of different racial/ethnic and sex groups on a variety of cognitive ability tests has come under extensive investigation. Much of the increased attention has emphasized the meaning or validity of such test scores as they are used for selection purposes in education and employment (e.g., Jaeger, 1976). However, meanings of test scores can be sought in other areas. Distinctions may be made by type of cognitive ability, such as verbal and nonverbal, since different groups may have different profiles of abilities. For example, it is widely held that females score higher on tests of verbal ability, whereas males tend to score higherThe author wishes to thank Barbara M. Wooby for finding and retaining the answer sheets, Anthony J. Mento for sorting them, and Ilean R. Cook for rescoring and reanalyzing them. Comments by T. J.