The aims of this study were to compare the reproducibility of three radiographic methods of assessing the quality of alveolar bone grafts, namely the Bergland, Kindelan and Chelsea Scales, and evaluate their application in the mixed and permanent dentitions. Additionally the use of occlusal versus periapical radiographs was assessed. Three examiners applied each scale on two occasions to the radiographs of 48 cleft lip and palate patients who had received alveolar bone grafts in 59 sites (11 had bilateral clefts). The agreement between repeated assessments by the same observer at different time points was measured by the kappa statistic, for each of the three assessors and each of the types of radiographic scale in turn. None of the three scales was found to be more reproducible than the others (kappa statistics for intraobserver variation ranged from 0.61 to 0.70). The agreement between observers was also similar across the three radiographic scales (multiple kappa statistics for inter-observer variation ranged from 0.45 to 0.51). Likewise, neither occlusal nor periapical radiographs were found to enable greater reproducibility of assessment. Surprisingly there was a tendency to greater reproducibility in the mixed than in the permanent dentition, which suggests the outcome of alveolar bone grafting may be assessed at an earlier stage than currently adopted. The outcome of alveolar bone grafting in this group of patients was generally successful.