The second one seems perfectly consistent with the hypothesis of bigger differences at the extreme right of the curve.
The reference on test design is interesting. I already knew that the tests have long been "re-balanced" to make them more “gender balanced” and “race neutral”. It appears that leaving out questions requiring spatial reasoning is one of the ways to accomplish that.
The re-balancing of the tests makes it almost impossible to have a meaningful discussion about I.Q.
Just to be clear, I wasn't arguing that it wasn't bigger at the extremes. I was arguing that even relatively close to the median, there are certain abilities which are notably divergent between typical male and female populations - and occupations that allow little room for divergent strategies will tend to be dominated by one group or another.
An interesting minor reference in one of the links is how females who are exceptional in verbal ability - even among females - appear to be able to perhaps be able to use that area of ability to compensate in the SAT-M, whereas males generally don't.