Two methods for estimating premorbid IQ were employed in a sample of 199 dementia patients and 26 control subjects: (1) the National Adult Reading Test (NART), a present ability measure, and (2) an age, sex, race, education and occupation regression formula-a demographically based estimate. The dementia sample consisted of probable Alzheimer's disease, multi-inf arct dementia and a mixture of the two. Controls consisted of the spouses of the patient sample. The patient sample was divided into three levels of dementia severity equated for age and level of education. The NART estimates in the mild and moderate/severe dementia groups differed significantly from those for the very mildly demented patients and controls. The results suggest that the applicability of the NART in estimating premorbid IQ in dementia may be limited.