Y'know, I think it would be very interesting to answer the question of whether the observed gender gap in ability in the hard sciences is due to genetics or some other factor.
However it's a damn-near-impossible to do decent research on this question. There are only two possible answers, A and B, but if your results show A then you'll be lauded and become the subject of approving articles like this one, and if your results show B then you'll be condemned and possibly hounded out of your job like Larry Summers. When there's such an incentive to get the socially-approved answer rather than the scientifically correct answer, lousy research tends to proliferate.
In my above comment I worked out that there MUST be genetic differences.
Even if you start from a point of complete genetic equality in aptitude, all it takes is any kind of cultural or social bias to create a selection pressure which will lead to genetic difference.
Given the historically different roles of men and women, it would be vanishingly unlikely for there not to be genetic differences.
If being very good at a task leads to more reproductive success, and only one gender performs that task, then only one gender receives the benefits of that selection pressure (to the extent of sex-chromosome-specific loci).
On the flip side, there is evidence that being TOO much of an outlier is negative. The smartest people tend to be more socially marginalized, both voluntarily and involuntarily. So it's possible that the top people are essentially evolutionary mistakes (as they are LESS likely to reproduce), and as evolution seems to roll the dice more with males than females, more males will turn out like this.
In a post-Darwinian society, this all goes out the window. It's just a historical relic of not being able to tinker with our genome directly, and having to rely on sexual reproduction.
However it's a damn-near-impossible to do decent research on this question. There are only two possible answers, A and B, but if your results show A then you'll be lauded and become the subject of approving articles like this one, and if your results show B then you'll be condemned and possibly hounded out of your job like Larry Summers. When there's such an incentive to get the socially-approved answer rather than the scientifically correct answer, lousy research tends to proliferate.