Aim: to evaluate the use of different functional scales in detecting dementia in a population study. Methods: the study is part of the Helsinki Ageing Study. A random sample of 795 subjects aged 75 (n = 274), 80 (n = 266) and 85 years (n = 255) was taken. The prevalences of dementia (DSM-III-R criteria) in these age groups were 4.6, 13.1 and 26.7% respectively. The functional scale scores were known for 71% of the non-demented and 66% of the demented subjects. A structured questionnaire completed by a close informant included four functional scales: the index of activities of daily living (ADL), the modified Blessed dementia scale (DS), the instrumental activities of daily living scale (IADL) and the Functional Assessment Questionnaire (FAQ). Results: all the functional scales discriminated demented from non-demented subjects. Based on receiver operating characteristics analysis, the area under the curve (95% confidence interval) was 0.90 (0.80-0.94) for the ADL, 0.94 (0.87-0.97) for the DS, 0.95 (0.90-0.98) for the IADL and 0.96 (0.92-0.98) for the FAQ. The effects of age, sex and education in detecting dementia were minor or non-existent in the ADL, DS and FAQ scales, but age had an effect on the performance of the IADL scale. All the scales detected even mild dementia adequately. Conclusions: functional scales can be used in detecting dementia when functional assessment is already used for other purposes, such as among elderly primary care patients.