Identification of craniofacial landmarks, particularly condylar anatomy, on the lateral cephalometric radiograph is erratic. The accuracy of recognition is critical for proper diagnosis of malocclusion and for assessing growth and orthodontic treatment. The aim of this study was to evaluate the identification of condylion and other cephalometric landmarks commonly used, or thought to be easily identifiable. A lateral cephalograph was taken on each of 34 adult subjects. Five examiners, three orthodontists, a dental radiologist and a second-year orthodontic resident rated the condyle, along with sella (S), nasion (Na), point A (A), infradentale (I), pogonion (Pog) and menton (Me) as identifiable, non-identifiable and interpreted. In addition, distances between several of these landmarks were measured. The left condyle, subject to less magnification than the right condyle because it is closer to the film, was more identifiable than the right condyle, which had the highest rating as non-identifiable. Among other landmarks, nasion and point A were the least identifiable, Pog and Me the most. Correlation coefficients for measurements between identifiable landmarks (I-Me) were greater than coefficients for distances involving less identifiable landmarks (S-A). Interestingly, linear measurements were less variable than the identification of corresponding landmarks. These findings corroborate the previous conclusions that clearly identifiable (and the least amount of interpreted) landmarks should be used for proper evaluation of dentofacial relationships. They also suggest, on average, precision in landmark identification is more critical for research purposes than in routine clinical cephalomteric measurements, which serve only as a guide to diagnosis.