Abstract
After briefly describing James Lovelock's Gaia hypothesis, I will argue that Gaia does not reproduce, or rather that it has what Peter Godfrey-Smith would call 'too many parents' to undergo natural selection according to Lewontin's Recipe. So the hypothesis did not and still does not make sense to most Darwinians. If (i)Lewontin's Recipe were extended to include differential persistence as well as differential reproduction, (ii) part/whole relationships were accepted as a form of 'reproduction', or (iii) the 'gene's-eye view' of Richard Dawkins as further extended by David Hull and us were adopted, then the Gaia hypothesis, and lesser claims about some multispecies communities, holobionts and ecosystems, would make sense. This last is what the It's the song not the singer(s) theory aimed to do.This article is part of the discussion meeting issue 'Chance and purpose in the evolution of biospheres'.