Classifications of European populations were produced based on 59 gene frequencies and 10 cranial measurements, but recorded for different population samples. A map-quadrat approach circumvented the problem of noncorrespondence of sampling localities. Clustering by standard numerical taxonomic procedures shows that these data are represented poorly as hierarchic classifications, but are associated with geographic regions. The agreement between classifications based on gene frequencies and those based on cranial variables is substantial (88%). Ordinations show genetic and morphometric continuity among these populations, although there is good regional separation on the whole. The relationships between phenetics, geography and language affiliation were examined by quadratic assignment methods. Both gene frequency and craniometric distances show significant correlations with geographic distances. The expected significant correlation between geographic distances and language affiliation is demonstrated. The gene frequency distances also show significant correlations with language distances. None of the partial correlations are significant. These analyses demonstrate a significant association of phenetics with language and geography, despite the lack of clear cut taxonomic structure in quadratted population data. For gene frequencies, these associations are more pronounced than for cranial measurements.