Maximum gape is important to the ecology and evolution of many vertebrates, particularly gape-limited predators, because it can restrict the sizes and shapes of prey that can be eaten. Although many cranial elements probably contribute to gape, it is typically estimated from jaw length or jaw width, or occasionally from a combination of these two measures. We measured maximum gape directly for 18 individuals of the western diamond-backed rattlesnake, Crotalus atrox. We measured each individual's body length, several external cranial dimensions, several cranial osteological dimensions from cleaned skeletons, and we calculated gape index values from two published gape indices (GI). Cranial bone lengths and gape circumference showed negative allometry with snout-vent length (SVL), indicating that small individuals have relatively larger heads and gapes than their larger conspecifics. We then used Akaike's Information Criterion to determine which external and osteological measurements were the best predictors of gape. Body size (SVL) was the best predictor of maximum gape overall; however, when SVL was excluded from the analysis, quadrate (QL) and mandible lengths (MdLs) were the best predictors of maximum gape using both external and osteological measurements. Quadrate length probably contributes directly to gape; however, the importance of MdL to gape is less clear and may be due largely to its allometric relationships with head length and SVL. The two published GI did not prove to be better indicators of actual gape than the jaw and QLs in this study, and the gape values they produced differed significantly from our empirically determined gapes. For these reasons, we urge caution with the use and interpretation of computed GI in future studies. The extensive variation in quadrate and mandible morphology among lineages suggest that these bones are more important to variation in gape among species and lineages than within a single species.