Alternatives to the partial Mantel test in the study of environmental factors shaping human morphological variation

Abstract

The importance of ecological factors in shaping morphological differences among human populations is highly debated despite numerous studies that have examined worldwide patterns of morphological variation (Roberts, 1953; Beals et al., 1984; Katzmarzyk and Leonard, 1998; Relethford, 2004a; Roseman, 2004; Harvati and Weaver, 2006; Betti et al., 2010). Some craniofacial studies have pointed out that the main forces behind cranial shape variation are random factors such as genetic drift and mutation (Relethford, 2004a; Manica et al., 2007; Betti et al., 2010). Conversely, other studies have suggested that craniofacial form and size variation, as well as dental size and postcranial morphological variables, are associated with environmental variation (Beals et al., 1984; Katzmarzyk and Leonard, 1998; Roseman, 2004; Harvati and Weaver, 2006; Perez and Monteiro, 2009; Bernal et al., 2010). The above studies generally assess the importance of ecological factors using correlative analyses of morphological (e.g., cranial shape, body size) and environmental (e.g., minimum temperature) variables. However, one of themajor problemswith this approach is that spatial structure may be present in the morphological variable (Legendre and Legendre, 1998). First, morphological variation may be related to other environmental or non-environmental variables that are spatially structured, generating spatial dependence among populations (Legendre and Legendre, 1998). For instance, repeated migration from one particular locale into another results in spatial structureda well-defined clinedleading to high morphological similarity between groups in close geographic proximity as well as