In this study we applied clustering procedures to a subgroup of 341 participants from the WAIS–R standardization sample. These individuals were selected by virtue of their having full-scale profiles characterized by scatter of greater than 8 scaled scores. Using a hierarchical clustering algorithm, a multistage procedure was used to establish and evaluate a cluster solution. The subject selection and clustering methods were successful in revealing a set of 9 profile types characterized by unique profile shapes. All profiles were associated with FSIQs that were at least in the average range. Seven of the profiles were characterized by specific subtest strengths, only 1 with subtest weaknesses. Examination of the external correlates of profile membership revealed differences among profile groups for age, marital status, education, and occupation. Our findings suggest that variability in and across the 9 profile types in this sample reflects increased contributions of unique abilities in comparison to the influence of the underlying primary and secondary WAIS–R dimensions of cognitive ability. (JINS, 1997, 3, 120–127.)