There are many problems in measuring "biological" age rather than chronological age, not least the necessity of demonstrating that this measure has validity in terms of better predicting mortality or morbidity. This review considers nine problems in measuring biological age and attempts to provide a full or partial solution to these difficulties. The problems presented include the fact that different organs age at different rates; biological age depends on the parameters studied; 'regression dilution bias' may produce misleading results; the validity of the measurement has to be established; the age range for measures has to be agreed; persons affected by accident or disease may or may not be excluded; accidental deaths may have to be excluded from a study of validation; individuals age at different rates at different times, and the fact that longitudinal data are necessary. The solutions suggested include both the presentation of aging profiles including several systems and the provision of summary statistics; adjusting for 'regression dilution bias', and calculating rate of aging from longitudinal studies over different age ranges and in the two sexes separately. It is argued that in this field of research it is well worth confronting the difficulties and extending our body of knowledge by further careful studies.