Optical surface scanning technologies produce dense three-dimensional (3D) data sets, which allow detailed analysis of surface morphology. This paper describes a method of analysing change in facial shape independently of change in size. The 3D data from three male subjects from the age of 6-21 years were recorded using an optical surface scanner. A series of 22 conventional landmarks were located with the aid of horizontal and vertical profiles across the face, and were analysed using geometric morphometrics. The 3D landmark co-ordinates were scaled and aligned using Generalized Procrustes Analysis (GPA) and analysed by Principal Component Analysis (PCA) to determine the shape change over the growth period for each individual. The results show that the centroid size reaches a steady value at different times for each of the subjects. When analysing shape versus age, highly significant correlations were found with principal component 1 (PC-1), but not with other principal components. PC-1 encompassed 40 per cent of the total variance for each subject. The movement of facial landmarks with time that is represented by PC-1 in each of the individuals is described. The use of these techniques has enabled the individual characteristics of facial growth to be identified and also has revealed the subtle changes in shape that continue after change in size has ceased.