Typological classification of human zygomaticomaxillary suture (ZMS) shape is often used in forensic assessment of ancestry, following earlier studies reporting higher frequencies of angled sutures among American Indians and higher frequencies of curved sutures among Caucasians. In this paper we present a new method of digital morphometrics to quantify and compare ZMS shape in 60 American Indian and 60 European crania. Suture outlines were recorded as three-dimensional (3D) contours on digital models of adult male and female crania created with a portable 3D laser scanner. Each contour was represented by about four hundred point coordinates, which were transformed via Fourier analysis into amplitude coefficients suitable for use in linear discriminant analysis. Discriminant functions were created that accurately predicted group membership for 83% of the crania in the sample, after leave-one-out cross-validation. The results were compared with traditional typological classifications based on visual evaluation of ZMS shape, and the contour-based method was found to be more effective than the typological approach. However, the distribution of ZMS types within the two sample groups did not conform to previously reported patterns. This discrepancy indicates that ZMS shape may reflect not only genetic factors, but also environmental factors such as diet and stress. In addition, some evidence for sexual dimorphism in the zygomaticomaxillary complex was observed. Based on these findings, we recommend caution when using ZMS shape analysis in forensic ancestry determination.