Type I procollagen concentrations were measured by radioimmunoassay in sera from 14 growth hormone-deficient children before and during 12 months of treatment with human growth hormone. Basal procollagen levels were lower than those of control children and comparable to those of normal adults. With treatment, the mean procollagen level increased into the range of the control children and was significantly greater than the baseline level at 1, 2, 3*, and 12 months (p<0.01; *p<0.05). Although there was no significant statistical correlation between the growth velocity during treatment and the serum procollagen level, there was a suggestion that a high basal procollagen may be predictive of a less than optimal response to human growth hormone.