This dissertation is an investigation of post-canine tooth root morphology in a global sample of modern humans. Tooth roots are variable in number, shape and orientation, and internal canal form and number do not necessarily covary with external morphology. However, this variation is poorly understood in anthropological and biological contexts. This is, in part, due to the inaccessibility of tooth roots for metric and morphological assessment. Early studies relied on x-rays, which are problematic when visualizing root structures, which are often curved or layered one on top of another. Computed tomography (CT) allows for clear visualization of tooth roots, and has revealed a previously unknown, complex combination of external and internal morphologies. Using CT scans from a global sample of humans (n = 945) a novel phenotype system is developed comprised of five elements: Root presence/absence (E1), canal root presence/absence (E2), canal location (E3), external root morphology (E4), and canal morphology and configuration (E5). Together, these five elements capture the external and internal morphology of the tooth root complex and are used to carry out four objectives: (1) to test and describe patterns of variation and divergence between root and canal number in individual teeth and between populations; (2) to develop a predictive model of tooth root morphology based on canal count and configuration; (3) to identify and define the total tooth root phenotypic set of the human sample; (4) to investigate if and how the total phenotypic set can delineate and define geographic and population structure in our sample. Novel statistical approaches are developed and used to ascertain complex patterning. Results indicate that there are clear differences between patterns of root to canal number both within and between teeth of the maxilla and mandible, and that these patterns are different between populations; that root canal number and orientation are powerful predictors of external root morphology; that the combined phenotype elements capture variation within and between populations; and that the combined phenotype elements can accurately identify and delineate population substructures. These findings are discussed in terms of evolutionary and developmental biology and biomechanics, and population structure and diversity.