The concept of race has had a significant influence on research in human biology since the early 19th century. But race was given its meaning and social impact in the political sphere and subsequently intervened in science as a foreign concept, not grounded in the dominant empiricism of modern biology. The uses of race in science were therefore often disruptive and controversial; at times, science had to be retrofitted to accommodate race, and science in turn was often used to explain and justify race. This relationship was unstable in large part because race was about a phenomenon that could not be observed directly, being based on claims about the structure and function of genomic DNA. Over time, this relationship has been characterized by distinct phases, evolving from the inference of genetic effects based on the observed phenotype to the measurement of base-pair variation in DNA. Despite this fundamental advance in methodology, liabilities imposed by the dual political-empirical origins of race persist. On the one hand, an optimistic prediction can be made that just as geology made it possible to overturn the myth of the recent creation of the earth and evolution told us where the living world came from, molecular genetics will end the use of race in biology. At the same time, because race is fundamentally a political and not a scientific idea, it is possible that only a political intervention will relieve us of the burden of race.