The aim of the study was to examine the association between height and cancer mortality in a socially homogeneous group of subjects. The study was based on a cohort of students, 8397 men and 2329 women, aged 16-30 y, who attended the University of Glasgow between 1948 and 1968. Mean follow-up time was 40 y. Height was measured at a medical examination performed at the student health service. The outcome measures used in the study were all-cause mortality and mortality from: all cancers, smoking and non-smoking related cancers and cancers related to sex hormones. No substantial or statistically significant associations were seen between height and all-cause or all-cancer mortality in either sex. Neither were any significant associations found between height and any of the sub-types of cancer studied (ie those related to smoking, those not related to smoking, and those related to sex hormones). Previous observations which have shown positive associations between height and cancer mortality have generally been based on populations with diverse social origins, among whom the variation in height will reflect variation in health and nutrition in childhood. The relatively low level of such variation in the present study may account for the negative findings. Public Health (2000) 114, 451-455