Assessments of skeletal maturity are usually made from handwrist radiographs, using either the method of Greulich-Pyle (GP) or Tanner-Whitehouse (TW). Recently the FELS method has been developed, and it represents a potentially valuable approach to skeletal age assessment. The present study evaluates the accuracy and precision of FELS skeletal age assessments compared with ratings by the GP and TW methods in a group of Italian children and adolescents. The hand-wrist radiographs of subjects (171 males and 156 females 1 to 17 years) referred to the "Istituto di Puericultura e Medicina Neonatale" of the University of Genoa in Italy between 1985 and 1990 were assessed according to each method. Two independent observers rated the radiographs and one observer reassessed them after 6 months or more. GP estimates rather closely match chronological age; TW ratings tend to overestimate chronological age by 7-9 months around puberty, mainly in boys; and the FELS method tends to overestimate chronological age by amounts that increase with chronological age. The degree of precision of skeletal age assessments is within the usually accepted limits. Higher levels of repeatability and reproducibility are apparent for the FELS estimates than for GP and TW assessments. Thus, skeletal maturity is likely to be adequately interpreted by the FELS method as well as by the more commonly used GP and TW systems in Italian children and adolescents. © 1992 Wiley-Liss, Inc.