That scientists employ metaphors is not disputed. Metaphors serve as heuristic tools for suggesting new hypotheses, new areas of research, and new research strategies. They also function as rhetorical devices for the communication of scientific ideas. What has been contested is the role of metaphors in the more "cognitive" areas of science, those dealing with descriptions of data and providing and evaluating explanations. In this article, I argue for the centrality of metaphors in scientific explanations. I draw on some ideas from M. Black (1955, 1993), M. Hesse (1966), and T. S. Kuhn (1973) to sketch out a model of scientific explanation as metaphorical redescription. I close by arguing that D. Dennett's recent algorithmic interpretation of Darwin's theory of evolution by natural selection is an exemplification of the model here sketched. (Michael Bradie)