Objective: This study attempted to estimate the relationship between independent assessments of "chin prominence" and "facial attractiveness" on the same set of standardized facial photographs. Methods: Paired sets of lateral cephalograms and standardized facial photographs generated at the same visit were available for a random sample of 48 treated Class I and Class II patients drawn retrospectively from the practice of a single experienced clinician. Each member of two independent groups of trained observers independently scored the photographs of each patient for "facial attractiveness". The mean patient scores of the two observer groups were then correlated with each other and with previously reported "chin prominence" data from the same records (JDR 82 Spec Issue B, 2003, Abst. 3022) Results: The Pearson correlation "r" between the estimates of the two groups of observers for "facial attractiveness" was 0.79, implying that each group’s estimate accounted for approximately 60% of the variability in the other. Each group’s estimate of "facial attractiveness" was then correlated with earlier data on "chin prominence" from the same photographs. Here the associations were substantially lower, the analogous "r" values being 0.53 and 0.42. We next correlated each observer group’s "facial attractiveness" scores with each of 32 conventional cephalometric and demographic variables that had previously been evaluated for their ability to predict "chin prominence". In general, the cephalometric variables correlated much less strongly with photographic estimates of "facial attractiveness" than they did with photographic estimates of "chin prominence". Conclusion: "Chin prominence", measured on either photographs or lateral cephalograms, did not of itself account for more than a third of the variance associated with estimates of "facial attractiveness" on photographs.