A recent debate over the causal foundations of evolutionary theory pits those who believe that natural selection causally explains long-term, adaptive population change against those who do not. In this paper, I argue that this debate – far from being an invention of several articles in 2002 – dates from our very first engagements with evolution as a quantified, statistical science. Further, when we analyze that history, we see that a pivotal figure in the early use of statistical methodology in evolutionary theory, W.F.R. Weldon, changes his mind about precisely the central claim at issue. I close by drawing some morals which I think the case can offer for the contemporary debate going forward.