A series of tests was conducted to provide information on the fracture tolerance ofthe bones of the face under impact conditions simulating those occurring in motor vehicle collisions. The results of that series were compared with the results of tests conducted under similar conditions on a Hybrid III dummy equipped with a frangible face.

The faces of eight cadavers were impacted at the subnasal maxilla and at the nasion with a horizontal steel bar of 25 mm diameter. The energy levels used were intended to cause fracture in about half of the impacts at each sire. Similar numbers of tests were performed at each of the two equivalent sites on the frangible face. The sequence of impact energies was intended to define the fracture threshold at each of the equivalent sites as precisely as possible. The data from both series of tests were analysed assuming that the probability of fracrure could be described by a Weibull distribution.

The cadaver tests for the maxilla resulted in only one significant fracture;​ no inference could be made about the probability of fracture at a given impact energy level within the range of the data. For the nasion, the cadaver data implied the existence of a plausible distribution of fracture probability as a function of impact energy. Using maximum force as the variate instead of impact energy produced essentially similar results. The data from the frangible face showed well defined thresholds at both impact sites. Although the cadaver data for the maxilla were somewhat inconclusive, it seems probable that the frangible face provides a quite conservative indication of the strength of the human face at the maxilla. For the nasion, the fracture threshold of the frangible face coincides approximately with the 40th percentile of the distribution of the cadaver tolerance data.