The possibility of measuring human activity patterns is discussed in terms of data obtained by attitude assessment, activity questionnaires, personal monitoring devices and fitness assessment. Specific problems are noted in applying these several techniques to elderly populations. Attitude assessments indicate the intention to exercise, and practical barriers may limit the translation of intention into overt behaviour. Questionnaires are open to exaggeration, but nevertheless remain the most practical method of evaluating large populations. Personal monitors can only be applied to small samples, but may nevertheless be helpful in improving questionnaire design. Fitness tests raise concerns of safety, and seem likely to bias population sampling towards the active elderly; scores are also influenced more by inherited factors than by current activity. Activity measurements seem unlikely to yield more than a crude estimate of total dietary needs in a population, but they may nevertheless help to identify a subset of old people where detailed nutritional examination is particularly cost-effective. The accurate determination of activity patterns will also have value as a means of encouraging the elderly to adopt an active lifestyle, maximizing their physical potential.