Not about intelligence specifically, but I believe this was the first (well-known) paper making the claim: http://www.philbio.org/wp-content/uploads/2010/11/Lewontin-The-Apportionment-of-Human-Diversity.pdf

The point is that even if the heritable component of (say) intelligence among white people formed a bell curve, and the heritable component of intelligence among black people formed a bell curve, a priori you'd expect the two curves to be pretty much the same.

(Lewontin's other conclusion, that "race" is "biologically meaningless", is separate and doesn't work because what small racial differences there are are statistically clustered: http://onlinelibrary.wiley.com/doi/10.1002/bies.10315/abstract;jsessionid=831B49767DB713DADCD9A1199D7ADC49.d02t02)