‘Racial’ or ‘ethnic’ drugs, a product of the new genetics and its mapping of genomes and populations, are now being developed and manufactured on a large scale. This article focuses on the conceptualization and identification of genetic signatures at the population level, many of which, I argue, evoke the ancient and quite common folk idiom of bodily inscription—in particular, fingerprints and birthmarks. The current geneticization of ‘colour’ and the biosociality engendered by it, I suggest, invite critical rethinking of the concept of insular populations and the distinction between bodily surface and deep structures (phenotype and genotype). While the new genetics has shifted the conceptual ground for discussions of human variation, moving away from phenotypic traits such as markings of the skin, drawing attention to what some molecular biologists refer to as the ‘universe within’, the notion of racial difference is repeatedly reinvented along familiar lines, under the banner of populations studies. I argue that, although human variation is both a legitimate and important subject in its own right and some approaches to variation do a better job than others, researchers need to be attentive to their assumptions about sampling. Circularity and subjectivity seem to be inevitable parts of the exploration of human diversity, and sampling cannot take place without a subjective, pragmatic judgment about how to proceed.