The detection and estimation of the effects of paternal and maternal age and birth rank on mutation rate are considered. Smith's (1972) method and discriminant function techniques are compared using data on Apert's acrocephalosyndactyly and achondroplasia.