Nubian biological evolution has come under much scrutiny in preceding years with strong opinions dividing scholars' beliefs as to how it occurred. A new approach is necessary to elucidate subtleties of their population structure in order to shed light on the complex subject. This study employs a spatial–temporal model in an effort to test if in situ biological evolution was the mechanism for biological change. Biological distance was calculated from the phenotype using 20 cranial discrete traits observed in Nubian samples from the sites of Semna South, Kerma, and the islands of Hesa and Biga. The periods represented at Semna South are Meroitic, X-Group and Christian, whereas the Kerma site yielded skeletal material from the Kerma period, and the islands of Hesa and Biga date to the Christian period. Mahalanobis D2 with a tetrachoric matrix was used to calculate biological distances among the samples, and three-way Mantel tests were applied to the distance, spatial and temporal matrices. Time was not significantly correlated with biological distance. However, an inverse relationship of time and biological distance is expected under the spatial–temporal model. A lack of significant geographic correlations, as found here, is unusual in most populations, but given the spatial construct of the sites along the Nile, this relationship is not completely unexpected. The lack of significant correlation among time, space and biological distance does not support the in situ hypothesis. Copyright © 2011 John Wiley & Sons, Ltd.