No good formal arguments exist for a central question in biology: Why, in species that have sexual reproduction, are there usually only “males” and “females”? We present a nonlinear optimization model that supports the conclusion that having only two sexes maximizes long-run viability.