We have carried out a study to determine the normal values of the 'block design test' in a sample group of 58 healthy elderly persons, a test which is usually used in the clinical evaluation of constructive apraxia. We therefore describe a grading scale which has not been previously published, in order to check the possible usefulness of this drawing as a rapid test of cognitive evaluation. A study was made of the possible interaction of the test with sample parameters, age, sex, educational level and also the inter-examiner reliability. We have observed a significant interaction between the variables of sex and educational level. Also the inter-examiner reliability is not very good. In view of our results, we consider that the 'block design test' is not very useful in the evaluation of constructive praxis in healthy elderly persons, and therefore it does not seem to us to be a good test for the rapid evaluation of cognitive function in this group.