Assessment of the results of treatment of tall girls with estrogens by comparing actual mature height with predicted mature height is unreliable, unless the accuracy of prediction has been tested in untreated tall girls. We report such an investigation in 14 tall girls who were treated with estrogens and in 14 untreated tall girls. Of the three prediction methods, those of Bayley and Tanner were reliable in our hands, whereas the method of Roche was less adequate. The mean reduction of height by treatment was about 7 cm using Tanner's method of prediction, 10 cm when Bayley's method was used. The standard deviation of the difference between actual and predicted height as found in untreated girls makes it hazardous to assess the amount of height reduction in the individual treated case with any certainty.