In commenting on another poster's misunderstanding of Regression to the Mean, I recalled a question on that subject that I have had before.
My understanding is that Dalton expected that regression to continue towards the mean in later generations. Do modern biologists still teach that?
The reason for regression in the first generation is easy to explain. Assume that you have a herd of cattle of one breed. You separate them into two populations, Population A having shorter legs than the median, and Population B having longer legs than the median. (Those with legs a precisely the median length, you sell off.) Call that generation 0 and the populations A0 and B0. Breed the populations seperately.
Population B1 will have shorter legs than Population B0:
Some of the length results from environmental circumstances which will not be repeated in the next generation.
Some of the length results from dominant alleles in several loci. When bull with Aa and Bb genes covers a cow with Aa and Bb genes, their descendants might have AA and bb genes, which will result in shorter legs.
Some of the length results from recessive alleles which are doubled, and thus expressed. When a bull with CC and dd genes covers a cow with cc and DD genes, all their descendants will have Cc and Dd genes. If c and d lengthen legs, then the descendants will have shorter legs.
It is hard to see why population B2 should have shorter legs than population B1. The percentages of alleles will be the same, and the distribution in B1 is as random as the distribution in B2.
Is there something here I'm not seeing? Has biology moved on from Dalton while I wasn't looking?