A growing body of research has examined the experiential grounding of scientific thought and the role of experiential intuitive knowledge in science learning. Meanwhile, research in cognitive linguistics has identified many conceptual metaphors (CMs), metaphorical mappings between abstract concepts and experiential source domains, implicit in everyday and scientific language. However, the contributions of CMs to scientific understanding and reasoning are still not clear. This study explores the roles that CMs play in scientific problem-solving through a detailed analysis of two physical chemistry PhD students solving problems on entropy. We report evidence in support of three claims: a range of CMs are used in problem-solving enabling flexible, experiential construals of abstract scientific concepts; CMs are coordinated with one another and other resources supporting the alignment of qualitative and quantitative reasoning; use of CMs grounds abstract reasoning in a “narrative” discourse incorporating conceptions of paths, agents, and movement. We conclude that CMs should be added to the set of intuitive resources others have suggested contribute to expertise in science. This proposal is consistent with two assumptions: that cognition is embodied and that internal cognitive structures and processes interact with semiotic systems. The implications of the findings for learning and instruction are discussed.