In a prospective study concerning the pathogenesis of impaired glucose tolerance and Type 2 (non-insulin-dependent) diabetes mellitus, 346 subjects with no clinical history of diabetes were given a standard 75 g oral glucose tolerance test. The expected positive associations between 120-min plasma glucose concentration and age and body mass index were observed in both sexes and between 120-min plasma glucose and waist/hip ratio in male subjects. An unexpected negative correlation was found between 120-min plasma glucose and height in both sexes (r = -0.23, (95% confidence interval, -0.38 - -0.07) p less than 0.007 for male subjects and r = -0.24, (-0.37 - -0.11) p less than 0.006 for female subjects). These negative associations with height remained significant after controlling for age and body mass index in male subjects but not in female subjects. In the latter a highly significant negative relationship of height with age was recorded (r = -0.33, (-0.45 - -0.20) p less than 0.0001). Comparison between individuals with impaired glucose tolerance and control subjects matched for sex, age and body mass index showed that subjects with impaired glucose tolerance are significantly shorter. Mean (+/- SEM) height in the male subjects with impaired glucose tolerance (n = 29) was 173.4 +/- 1.1 cm vs 176.9 +/- 1.3 cm in control subjects, p = 0.02. In the female subjects (n = 39) mean (+/- SEM) height was 159.4 +/- 1.0 cm vs 162.4 +/- 1.0 cm in control subjects, p = 0.02. The negative relationship between height and glucose tolerance is a new epidemiological observation which has not been previously reported.(ABSTRACT TRUNCATED AT 250 WORDS)