We propose a novel account of scientific representation based on Springle (2021). On this account, representations function to provide agents with practical epistemic access to situations. Roughly, to successfully represent something is to have the kind of access to it that one has when one knows (epistemic) what to do (practical) with it, and one knows what to do with something when one is in a position to involve it in solving a practical problem. On this view, there are two major species of representation. "Embodied instructive representations" are fundamental (psychological) representations and function to provide direct practical epistemic access to facts. "Surrogative representation" are non-fundamental representations and function to provide indirect-practical epistemic access to facts by solving "proximity problems." We argue that scientific representations are a species of surrogative representation that functions to solve scientific varieties of proximity problems. We argue that this account is attractive because, unlike other accounts of scientific representation, it is essentially a generic account of representation. We also argue it captures what other accounts of scientific representation get right while avoiding their shortcomings.