The serrated, or denticulated, ziphodont teeth of theropod dinosaurs display variability in their extent of denticulation. The functional model proposed here tests the hypothesis that denticles will not exist in areas that do not frequently contact the substrate. This area, defined as the “dead-space,” is determined by the direction the tooth moves through the fleshy substrate. The extent of denticulation, as well as the dead-space dimensions, is measured from photographs of 235 isolated and in situ theropod teeth, to determine a meaningful relationship between the two variables. Both Euclidean and geometric morphometric methods are employed, and the data are expressed in bivariate and ordination plots. The model predicts the direction of tooth movement through the curvature of the tip/apex. Tooth position and taxon are considered. The results show that the mesial margin is usually partially denticulated, while the distal margin is usually totally denticulated. Curved teeth have large dead-spaces, and tend to be less denticulated mesially. Straighter teeth are more extensively denticulated, to the point where they became symmetrical. The mesial denticulation is determined by the dead-space, and dictated by the substrate contact. The dead-space almost always predicted less extensive denticulation; a consequence of the model's limitations. Tooth curvature increases with a more distal position, due to rotation based on the proximity to the hinge. Denticulation indicates that theropods used a distally oriented puncture to modify the substrate, similar to modern analogues. Although there is little taxonomic variation, Troodontidae show unique and extreme degrees of mesial denticulation. Anat Rec, 292:1297–1314, 2009. © 2009 Wiley-Liss, Inc.