Recently, in response to something I wrote, a kinist presented a "scientific" argument against miscegenation. The same individual also made a negative, passing reference to Richard Lewontin.
Since Lewontin is still alive, and has an email address on his academic webpage, I decided to quote the "scientific" argument to him in case he'd like to comment on it. He was kind enough to respond (see below).
Dear Steve Hays,
Two comments on what you sent me:
(1) The idea of "genetic mismatches" has been made up out of the person’s head. There is absolutely no evidence of such "mismatches" with the implication that if I have [a] mixture of "racial" characters I am somehow debilitated. Moreover, the offspring of a mating between a West African and European does not have the skin color of one parent and the body shape of another. Indeed body shapes vary immensely both within and between groups. These characters, to the extent that they are inherited, are the consequence of a very large number of genes and if the parents were homozygotes for allelic variants of these genes, the offspring would be a heterozygote at all the gene loci involved. The offspring of this mating does not have the skin color of one parent and the body shape of another. It has a new, intermediate skin color unlike either parent and a new body shape.
 As to the argument about clustering in n-dimensional space, the statement that there are clusters is quite true (although of course the clustering has been much broken down with the history of matings between members of the original clusters). The clustering, as your racist implies, is hierarchical, with subclusters within clusters. If you want to call a cluster a "race" then you will do so, but where do you stop? Members of families will cluster within nation and tribe clusters, within geographical clusters, etc. More important, the proportion of all genes whose variation is reflected in such clustering is very small. Most gene variation (85%) occurs within local populations and another 6% between populations that would be assigned to the same race. I don’t have to be a geneticist to know that skin color variation has its variation mostly between classical "races". On the other hand the frequencies of different major blood group genes (ABO, MN, etc) do not contribute to this clustering but the variation is spread over all the so-called races. The question that is REALLY at issue is whether there are genes "for" things like intelligence, morality, various normal variations in psychology, etc. etc. Nobody has found these genes. If these genes are finally found, we could then ask whether there are differences in frequencies of different variants among geographical groups and whether there is a clustering in the gene frequency space corresponding to the clustering based on the few genetic characters that we know that do cluster geographically, like skin. If we can judge by the genes we do know the chance is 9:1 that there will be no significant racial clustering for any particular character. But occasionally we may find some interesting genetic case that does cluster. Nobody knows the answer to that question. We are certainly NOT in a position to say that there are such important cases or what they will be. When your racist person can cite such a case then he (she) will be entitled to talk about it.