I agree with the heritability argument for explaining (partially) cross-sectional data, but of course longitudinal changes over short periods--such as the huge changes in marital behavior, delinquency, adult crime, dropout from the labor force, etc., over the last half-century, are pretty hard to explain by appealing to changes in the genetic makeup of the population during the same period. Gotta keep both parts of the equation in mind.

Sounds like that one needs a lot more study. For instance, might the operative genetic component not be more related to some other physical similarity (such as, for instance, age at onset of menses--which, in turn, I believe, is related to body weight)? The size of the study also seems extremely small. Remember, some significant proportion of "first intercourse(s)" will likely be between siblings, half-siblings, and cousins--simply, in many cases, as a result of proxmity and access. If these are included, it uneccesarily adds to difficulty in isolating actual causation, as well as increasing the likelihood that some other aspect of relatededness is the responsible (or more responsible) one.