As part of an examination of the population structure of short-beaked common dolphins (Delphinus delphis) in the North Atlantic, I tested if there were systematic differences in cranial morphology, in relation to geographic location, for common dolphins both within the western North Atlantic (wNA; n = 141) and between the wNA and eastern North Atlantic (eNA; n = 106). Cranial specimens from the wNA were obtained between Nova Scotia, Canada, and Florida. Those from the eNA came from the Irish Sea and the coasts of Ireland and the United Kingdom. A Wilks' λ canonical discriminant analysis (CDA) was performed on the within-groups covariance matrix to test whether significant differences in group centroids (multivariate means) existed between putative population units separately for males and females. In addition, the CDA was used to reclassify each dolphin into a geographic group based on the discriminant function. The CDA of 35 cranial variables found no evidence (males: Wilks' λ = 0.603, P = 0.286; females: Wilks' λ = 0.145, P = 0.08) of population structure below the species level within the wNA. Thus, the 1-population model for this region was supported. CDAs revealed significant differences between the eNA and wNA for both males and females (males: Wilks' λ = 0.371, P < 0.0001; females: Wilks' λ = 0.260, P < 0.0001). Cross-validated reclassification rates for males were 78.8% (eNA) and 87.6% (wNA) and for females were 90.6% (eNA) and 81.4% (wNA). Measurements associated with the rostrum were important discriminating variables that might reflect differences in feeding habits between these areas.