This idea that harmony between the races is impossible--let's call it the "National Review fallacy"--rests largely on confusion about a form of natural selection known as "kin selection." The issues are a bit arcane, but I'll try to provide a rough sketch of some of them.
"Kin selection" accounts for the evolution of altruistic impulses toward close relatives. The textbook example of kin selection is a newly minted gene that inclines a ground squirrel to stand up and give an alarm call upon seeing a predator. At first glance, this gene would seem to have no chance of proliferating via natural selection, since it attracts the predator's attention and thus endangers the organism in which it resides. But remember: The gene will also reside, on average, in half of that organism's siblings--and their survival prospects are enhanced by the gene's effect (i.e., by the warning call). So, even if this "warning call" gene occasionally causes the death of its possessor, the gene itself may still flourish by natural selection, as long as more than two siblings are saved for every one ground squirrel that is lost. (If this Cliff Notes version of kin selection doesn't seem to make sense, then please go and read the excerpt from the chapter titled "Families" on the Web site for my book, The Moral Animal, then come back, and keep reading.) In our species, the result of this evolutionary process seems to be a kin-directed altruism that is roughly proportional to the closeness of relatives. Most people would be more inclined to risk their lives for a sibling than for a cousin, and for a cousin than for the average Joe. (This assumes, among other things, that these people have been reared in close enough proximity to these relatives to develop the emotional bonds that mediate kin-selected altruism.)
Here is where confusion enters the minds of people eager to believe that whites and blacks are innately hostile toward one another. They try to extend the logic of kin selection beyond the scope of the family and carry it all the way up to the level of whole races. They are assuming, in other words, that there is a universal law dictating that altruism between individuals be proportional to their degree of genetic relatedness--and that natural discord among people thus will be proportional to their genetic difference.
There are at least two major problems with this logic. The first is a fairly technical (though consequential) analytical flaw, first identified in another context by Richard Dawkins and labeled "Washburn's fallacy." (See his "Twelve Misunderstandings of Kin Selection." Zeitschr. Tierpsychol, no. 51 : 184-200). I won't even try to explain the fallacy here, except to say that a) It consists of assuming that kin selection would make altruism proportional to overall genetic relatedness--that is, the percentage of all your genes that you have in common with another organism; and b) This assumption has been memorably characterized as implying that humans should, in theory, be "nicer to mosquitos than to marigolds." That characterization was made by Martin Daly, Catherine Salmon, and Margo Wilson. For their explicit application of Dawkins' analysis to the National Review fallacy (they don't call it that, of course), see their chapter in the forthcoming textbook Evolutionary Social Psychology, edited by Douglas Kenrick and Jeffrey Simpson, and published by Lawrence Erlbaum Associates.
The second problem with the idea of some iron law correlating altruism with genetic relatedness--and thus correlating natural discord with genetic difference--is at least slightly more accessible. Kin selection isn't some inexorable force of evolution. It's just a theoretical possibility, one that will only be realized if the circumstances of evolution are conducive to its realization. In the case of altruism directed toward close relatives, we know that circumstances were indeed so conducive: Throughout human evolution, people were reared a) near close relatives; and b) near people who weren't close relatives. Thus there was lots of opportunity for the flourishing of genes that led humans to discriminate between the two, favoring the former at the expense of the latter. But in the case of comparable discrimination between members of one's own race and members of other races, there was no significant opportunity for the evolution of such a trait. Because during human evolution (that is, during that short span of human evolution that took place after distinct races began forming), there was roughly zero contact among different races; people in Africa didn't vacation on the Riviera back then. Saying that white people evolved an innate aversion to blacks, or blacks an innate aversion to whites, is like saying people evolved an innate aversion to some poison plant that grows only on Mars; the opportunity simply wasn't there.
None of this is to suggest that human nature doesn't vastly complicate race relations. People are obviously inclined to derogate groups whose interests seem to clash with those of their own group, and to identify those groups by whatever means are available. Skin color can be an unfortunately handy means of doing the identifying. What's more, kin selection itself may complicate race relations in various subtle ways. For example: Nepotism, one legacy of kin selection, is often de facto racial discrimination, since your close relatives are usually members of your race. When a white boss promotes his niece, he is discriminating against some whites (the ones who aren't in his family), but against all blacks.
All told, the obstacles to intergroup harmony posed by human nature are big enough that there is little exaggeration in saying that xenophobia is a part of human nature, at least in this sense: Uncritical hostility toward an identifiable group of people--identifiable by language, dress, color, whatever--is an inherent capacity, activated under certain predictable circumstances. But that is very different from saying we are designed to automatically dislike people with particular skin colors, and that racial harmony therefore is impossible--which is what the National Review article said.
Full-disclosure paragraph: The article in which the "National Review fallacy" appeared was a review of my book, The Moral Animal. One of the review's major complaints was about my alleged failure to realize that Darwinism is a thoroughgoing vindication of the reviewer's various political beliefs (e.g., the impossibility of racial harmony). No doubt some of my animus toward the article is related to these comments about my book. Still, I'm not inventing the idea that the "National Review fallacy" is indeed a fallacy. The same opinion is held by, for example, George Williams, one of the greatest evolutionary biologists of this century and arguably the chief architect of evolutionary psychology. By the way, his 1966 book Adaptation and Natural Selection (Princeton University Press), which laid the theoretical foundation for the modern Darwinian study of social behavior in animals, still is in print.