Hard tissue palatal dimensions and shape in 29 boys and 32 girls (mean age 14 years) and 30 men and 35 women (mean age 20 years), all with a complete permanent dentition including the second molars, and free from respiratory problems, were analyzed. Palatal landmarks were digitized with a computerized 3D instrument, and their coordinates were used to derive a mathematical model of palatal form. Palatal shape (size-independent) was assessed by a four-order polynomial in the sagittal and frontal plane projections. Palatal dimensions in the frontal and sagittal planes were computed, and compared between ages and sexes by two-way factorial analyses of variance. The effect of age on palatal shape was different in the two sexes: while in females the curves in both the frontal and sagittal planes were practically superimposed, in males the youngest subjects had somewhat 'lower' curves than the oldest subjects. Sex had a significant (p<0.05) effect for all variables. Age was significant only for palatal slope in the sagittal plane (p<0.01), and for raphe inclination in the horizontal plane (p<0.05). For all variables but palatal length in the sagittal plane, a significant (p<0.01) sexxage interaction was found: while in males all linear variables were larger in adults than in adolescents, the reverse was found in females. Overall, while in adolescent females palatal form appeared to have already attained adult characteristics, in males some modifications between adolescence and young adulthood were measured. These modifications should be taken into consideration as a possible factor of relapse after orthodontic treatment.