Given that race is largely a social construct rather than a biological fact, it is difficult to claim any relationship between "race" and intelligence. For example, in the United States we've inherited racial definitions that classify anyone with "one drop" of black blood as "black"—or less extremely, a legacy that defined has people with, say, one black grandparent as "black," even though they are, strictly speaking, 75 percent white. To intelligence test such a person as "black" and draw conclusions based on that premise simply doesn't make logical sense. Furthermore, DNA testing has shown that the human race is highly similar in terms of DNA, and that two people deemed "white"—say, an Italian and a Swede—can have more genetic variation than a black African and a white Frenchman. Finally, the genetic variants that can be found between blacks and whites generally cluster around phenotypical physical features, such as skin pigmentation and hair texture, that have no bearing on intelligence.
Furthermore, it should be pointed out that much intelligence testing is also socially constructed, privileging the kind of cognitive (as opposed to emotional, artistic, kinesthetic, or social) training that higher-caste groups have greater access to in Western culture.
With race as a suspect category, it is difficult to defend studies that purport to connect race and intelligence. If we could find statistically significant (i.e., large enough) groups of people of different races with verifiably very distinct genetic differences and manage to test their intelligence in culturally non-biased ways, we might learn something, but generally such studies have been sloppily done and motivated by political agendas that need to be carefully questioned, given the ugly legacy of racism in our culture and other cultures.