In this paper I consider the view that scientific change is the result of a selection process which has the same structure as that which drives natural selection. I argue that there are important differences between organic evolution and scientific growth. First, natural selection is much more constrained than scientific change; for example it is hard to populations of organisms to escape local maxima. Science progresses; it may not even make sense to say that biological evolution is progressive. Second, natural (...) selection depends for its power on the specifics of its domain, so I doubt that there is much point in seeing a selective regime in science as an instance of a more general family of selective regimes. Third, the replicator/interactor distinction fits scientific change much less well than biological evolution. But a family of selective theories of science can be identified ranging from the very ambitious to the very modest. Though the very ambitious programs of evolutionary epistemology are in trouble, there is space for one which is not a trivial redescription of what everyone already knows, but which is sensitive to the peculiarities of its domain. That selective theory explains important aspects of the community organization of science, an organization which is central to scientific progress. (shrink)