Mechanistic explanations are a common form of explanation in such sciences as neuroscience and psychology. In giving these explanations, researchers attempt to explain a capacity of a mechanism in terms of the capacities of its components. The most popular account of mechanistic explanations currently offered is by Craver (2007), who characterizes them in terms of the “ontic” conception of explanation. We are told that an ontic explanation involves showing how the features of a mechanism fit into the objective, causal structure of the world. In this talk I want to raise some concerns about the view that mechanistic explanations should be seen as a form of ontic explanations. I will suggest that there is a perfectly reasonable way to describe these explanations within the alternative “epistemic” tradition of explanation that Craver rejects. I consider what this alternative might look like with some examples and describe what the consequences are for his account.