The Homophone Meaning Generation Test (HMGT; Warrington, 2000) is a new measure of verbal fluency that has been demonstrated to be sensitive to the presence of anterior lesions. In the present study we used the HMGT healthy standardization sample (N = 170) and demonstrate that scores on the HMGT do not differ significantly from a normal distribution and that the test has adequate reliability (α = .82). A table for obtaining confidence limits on an individual's score is presented. A regression equation for the estimation of premorbid HMGT performance was constructed using the National Adult Reading Test as the predictor variable. In a sample of 36 cases with anterior lesions estimated premorbid scores were significantly higher than obtained scores (p < .001). Premorbid ability acted to suppress group differences on the HMGT; the partial correlation between neurological status (healthy vs. anterior lesion) and HMGT performance controlling for premorbid ability (.53) was significantly higher than the raw correlation (.44). In addition, hierarchical discriminant function analysis demonstrated that the inclusion of premorbid ability improved classification over that achieved by HMGT scores alone. These results support both the underlying rationale and the clinical utility of controlling for premorbid performance when interpreting verbal fluency scores. (JINS, 2002, 8, 547–554.)