This article uses fine-grained data to demonstrate that, since 1996, the State of Texas has undergone a process of ‘secular geographical polarization’ – a continuous divergence in the geographical bases of its political parties. It is suggested that this process exemplifies a new era of partisan politics in the American South. Analyses of spatial regression models show that the geographical polarization can be partially explained by a tighter link between demographic characteristics and aggregate voting patterns, but that growth in spatial clustering cannot be attributed entirely to demographics. The possibility that spatially-bounded social contexts are affecting partisan change is thus explored. Finally, the article's findings are considered in light of the growing debate about geographical polarization in the American electorate.