Growth hormone (GH) concentrations (in ng/ml) were determined by radioimmunoassay, in plasma obtained at about 3-hr intervals during a 24-hr sampling span, from 42 boys and 12 girls of short stature (2-4 standard deviations below their peer group mean), and 13 boys and 9 girls of standard stature. Subjects had 11.20 +/- 0.37 years of age at the time of study, and were living on a diurnal waking (approximately 07:30 to approximately 22:30), nocturnal resting routine during sampling. Analysis of these data by single and population-mean cosinor methods as well as by analysis of variance revealed circadian and ultradian prominent components characterizing most groups. Accordingly, a multiple component analysis was undertaken for data of each group separately, as well as for all subjects. A comparison of circadian parameters indicates similar characteristics between short and standard children, whether one compares boys [P = 0.674, 0.371 and 0.749 for comparison of rhythm adjusted means (M), amplitudes (A) and acrophases (phi), respectively], girls (P = 0.993, 0.914 and 0.397), or all children (P = 0.859, 0.712 and 0.865). Differences are found, however, in circasemidian characteristics as well as in the prominent 8-hr ultradian component documented for the short but not for the standard children. These ultradian components should be taken into consideration in the design and later evaluation of a time-specified treatment of children of short stature.