Thirty-three males who had received oxandrolone therapy for short stature were evaluated for gynecomastia which had its onset during or following completion of therapy. When compared with a group of treated boys who did not develop gynecomastia, mean duration of treatment was similar (3.3 v. 3.4 yr), and appeared to have no relationship to the appearance of gynecomastia. When compared with a group of normal untreated males with transient gynecomastia, the subjects' mean age (16.1 v. 13.6 yr), mean bone age (14.7 v. 12.8 yr) and mean pubertal stage (4.2 v. 2.9) at onset were significantly greater. Ten of the 33 had transient gynecomastia and 23 had persistent gynecomastia which required mastectomy. Mean age, bone age, pubertal stage and duration of treatment were not different between these two groups. Mean breast size was significantly different (1.6 v. 6.4 cm) and was similar to that observed in two untreated groups with transient (1.7 cm) or persistent (6.9 cm) gynecomastia, with the exception that the untreated subjects tended to be younger and in an earlier pubertal stage at onset. We conclude that oxandrolone may delay appearance of gynecomastia and that if it occurs, it is more likely to be larger and persistent than typical adolescent gynecomastia. Patients receiving oxandrolone for growth stimulation should be advised of this possibility.