"The truth is rather different. There is far more genetic variation within racial groups (around 85%) than there is between racial groups (just 15%). Indeed, surface appearance is often a highly misleading way of assessing the genetic distance between populations."

This doesn't prove what he thinks it proves.

This is the exact same argument: There is more variation within male heights than there is between the male & female average height. This is true - the difference between the average male height and female height is only a few inches, and it is quite a bit smaller than an interval which covers 90% of men. But, this has nothing to do with his point. The height difference between males and females is still relevant to us, even though the in-group variation for males & females is also high.

tl;dr: male height variation is much bigger than the gap between male & female heights, yet we still think the male-female height gap is a real & meaningful phenomenon.

[1] http://en.wikipedia.org/wiki/Human_genetic_diversity:_Lewont...