Do you have some references -- preferably URLs -- that show these differences in sub-areas?
Here's a chart illustrating the point that I was making:
The chart shows that there should be plenty of women able to handle undergraduate math or science scores (requiring, say, two SD above average ability). It also would show (if the scale were much larger) that there's a far smaller proportion of women than men able to handle PhD level math or science (requiring > 3 SD above average ability).
Note how the male Normal Curve is flatter and wider than the female curve -- i.e. there are more men at the extreme ends, but close to the same numbers around the "average".
More here:
http://www.lagriffedulion.f2s.com/math.htm