[clarkm]

> The reason is that there's a limit to how big the variance can get- it doesn't grow forever, it grows until limited by other factors (i.e., there are only so many intelligence-promoting alleles that you can possibly have).

What makes you think that? You could say the same thing about seed size in plants, but comparing modern corn kernels to wild type seeds provides a clear counterexample.

You should look at the type of stuff they've done in plants and animals. Here's a paper on the selective breeding of fruit flies for flight speed. The results are astounding:

"The mean apparent flying speed of both lines increased from approximately 2 to 170 cm/sec and continued to respond at diminishing rates, without reaching a plateau, for 100 generations. Competitive fitness tests in generations 50 and 85 showed minimal or no fitness loss in selected lines compared to controls."

http://www.ncbi.nlm.nih.gov/pubmed/8878686