Understanding the purpose of sex remains one of the most important unresolved problems in evolutionary biology. The di¢ culty is not that there are too few theories of sex, the di¢ culty is that there are too many and none stand out. To distinguish between theories we suggest the following question: Why are there no triparental species in which an o§spring is composed of the genetic material of three individuals? A successful theory should confer an advantage to biparental sex over asexual reproduction without conferring an even greater advantage to triparental sex. We pose our question in the context of two leading theories of sex, the (deterministic) mutational hypothesis that sex reduces the rate at which harmful mutations accumulate, and the red queen hypothesis that sex reduces the impact of parasitic attack by increasing genotypic variability. We show that the mutational hypothesis fails to provide an answer to the question because it implies that triparental sex dominates biparental sex, so the latter should never be observed. In contrast, we show that the red queen hypothesis is able to explain biparental sex without conferring an even greater advantage to triparental sex.