Thirty short, slowly growing children with normal plasma growth hormone response to standard provocation tests were randomly assigned to a group (n = 20) undergoing therapy with methionyl growth hormone, 2 IU/m2 subcutaneously once daily, (group 1) or a control group (n = 10, group 2). The mean (+/- SD) height velocity increment in group 1 was 3.0 +/- 1.9 cm/yr in the first year, compared with -0.2 +/- 0.7 cm/yr in group 2. Of the 18 children who completed the first year of treatment, 12 had a height velocity increment of more than 2 cm/yr and 11 of them continued treatment for a second year (group 1A). The remaining six children also reached height velocities greater than the mean for bone age, but because of a low height velocity increment they were termed nonresponders and their growth hormone dosage was increased to 4 IU/m2/day (group 1B). Of the 10 children in the control group, seven received authentic biosynthetic growth hormone in the second year of the study (group 2); the remaining three received no therapy (group 3). The mean height velocities (measured in centimeters per year) before and during the first and second years of therapy were 3.6, 7.6, and 6.1 in group 1A; 5.7, 6.9, and 7.3 in group 1B; 4.2, 4.0, and 6.7 in group 2; and 5.0, 4.9, and 5.2 in group 3. The effect of doubling the dosage was a further increase of 1.9 cm/yr. Bone age advance paralleled growth acceleration, resulting in an unchanged height standard deviation score for bone age and ambiguous results on final height prediction. Growth hormone therapy in such short children appears to be safe and efficacious in increasing growth velocity for 2 years, but its efficacy in terms of increasing final height is uncertain.