Successive height predictions were made by several methods for a group of healthy Finnish children (30 boys and 30 girls), examined annually at ages of 7 to 17 years (1st series) and for 7 boys aged 14 to 19 years with familial delayed growth and puberty (2nd series). The methods used were those of Bayley & Pinneau (BP), Walker (W), Tanner et al. (T) and RWT, and two simple principles: the relative height method (RH) which assumes constancy of height S.D.S. throughout growth, and the index of potential height (IPH) method which assumes constancy of height S.D.S. for bone age (BA). The predictions with RH, W and IPH were inaccurate. BP, T and RWT were for the 1st series as accurate as for the basic series of the respective methods, and none was superior to the others. The BA of average Finns was delayed as compared with the standards of Greulich-Pyle Atlas. When corrections were made for this delay, the IPH method gave predictions comparable in accuracy to BP, T or RWT. In the 2nd series prediction was more accurate with the corrected IPH, BP and RWT methods than with those using BA according to TW2RUS.