Mandibular growth was estimated by four cephalometric measurements on three lateral cephalograms of 21 subjects. The estimates made by three standard cephalometric approaches were compared with that made by a "scientific" method. The scientific method was based on the change in position of the cephalometric landmark condylion on cephalograms orientated on two metal implants inserted in the mandible. The other three investigated "standard" cephalometric methods were each based on estimating the difference in length between two cephalometric landmarks: (1) pogonion-condylion, (2) pogonion-articulare, and (3) maximum mandibular length. In general, the individual estimates of the amount of mandibular growth by the scientific method and that estimated by each of the standard cephalometric methods were not proportional. Some factors affected the estimates of mandibular growth. For example, (1) the growth direction of the condyle, (2) the change in position of pogonion on the mandible during the growth, and (3) the apposition of bone on the chin in some cases. The amount of mandibular growth, which is estimated by the standard cephalometric methods, and growth velocity curves, which is based on such estimates, are therefore not valid. Accordingly, conventional cephalometric analyses are not as reliable at traditionally envisaged.