We tested a revised version of the Hierarchic Dementia Scale (HDS), proposed by Cole and Dastoor (1980), in order to improve its clinical usefulness and to enrich our knowledge about ageing. The scale was built with 20 subtests which covered the entire range of cognitive and motor functions. Each subtest was hierarchically organized so that success in a item implied success in inferior items. This hierarchical principle was time-saving and was validated by Cole and Dastoor. 149 control subjects performed this test. They were equally divided in 4 age-groups (55-64, 65-74, 75-84, 85-97) and 2 educational levels. None of these subjects had previous history of somatic or neuropsychiatric disease. They were completely self-sufficient in daily life. A large part of the controls failed in the most difficult items of some subtests: Learning, Calculation, Mental Control, Drawing, Recall, Similarities, Constructional Praxis. For these subtests, significantly different mean-scores were observed between age-groups and educational levels. However, the influence of each factor was variable from one subtest to another. Moreover, subgroups seem to exist in our population according to specific difficulties in some of these subtests. This study calls for caution in the interpretation of results in demented patients. Comparisons with other psychometric tools remain to be performed. This scale seems to be more useful for the quantification and follow-up of cognitive deficits than for the early diagnosis of dementia. In addition, this scale, which briefly explores many aspects of cognitive functions, seems especially useful to approach the heterogeneity of DAT.