Some authors defending the “hereditarian” hypothesis with respect to differences in average IQ scores between populations have argued that the sorts of environmental variation hypothesized by some researchers rejecting the hereditarian position should leave discoverable statistical traces, namely changes in the overall variance of scores or in variance–covariance matrices relating scores to other variables. In this paper, I argue that the claims regarding the discoverability of such statistical signals are broadly mistaken—there is no good reason to suspect that the hypothesized (...) environmental causes would leave detectable traces of the sorts suggested. As there remains no way to gather evidence that would permit the direct refutation of the environmental hypotheses, and no direct evidence for the hereditarian position, it remains the case, I argue, that the hereditarian position is unsupported by current evidence. (shrink)