Objectives Craniofacial variation in past and present Amerindians has been attributed to the effect of multiple founder events, or to one major migration followed by in situ differentiation and possibly recurrent contacts among Circum‐Arctic groups. Our study aims to: (i) detect morphological differences that may indicate several migrations; (ii) test for the presence of genetic isolation; and (iii) test the correlation between shape data and competing settlement hypotheses by taking into account geography, chronology, climate effects, the presence of genetic isolation and recurrent gene flow. Methods We analyzed a large sample of three‐dimensional (3D) cranial surface scans (803 specimens) including past and modern groups from America and Australasia. Shape variation was investigated using geometric morphometrics. Differential external gene flow was evaluated by applying genetic concepts to morphometric data (Relethford‐Blangero approach). Settlement hypotheses were tested using a matrix correlation approach (Mantel tests). Results Our results highlight the strong dichotomy between Circum‐Arctic and continental Amerindians as well as the impact of climate adaptation, and possibly recurrent gene flow in the Circum‐Arctic area. There is also evidence for the impact of genetic isolation on phenetic variation in Baja California. Several settlement hypotheses are correlated with our data. Conclusions The three approaches used in this study highlight the importance of local processes especially in Baja California, and caution against the use of overly simplistic models when searching for the number of migration events. The results stress the complexity of the settlement of the Americas as well as the mosaic nature of the processes involved in this process. Am. J. Hum. Biol. 28:646–661, 2016. © 2016 Wiley Periodicals, Inc.