In this paper, we contend that what to teach about scientific reasoning has been bedeviled by a lack of clarity about the construct. Drawing on the insights emerging from a cognitive history of science, we argue for a conception of scientific reasoning based on six "styles of scientific reasoning." Each "style" requires its own specific ontological and procedural entities, and invokes its own epistemic values and constructs. Consequently, learning science requires the development of not just content knowledge but, in addition, procedural knowledge, and epistemic knowledge. Previous attempts to develop a coherent account of scientific reasoning have neglected the significance of either procedural knowledge, epistemic knowledge, or both. In contrast, "styles of reasoning" "do" recognize the need for all three elements of domain-specific knowledge, and the complexity and situated nature of scientific practice. Most importantly, "styles of reasoning" offer science education a means of valorizing the intellectual and cultural contribution that the sciences have made to contemporary thought, an argument that is sorely missing from common rationales for science education. Second, the construct of "styles of reasoning" offers a more coherent conceptual schema for the construct of scientific reasoning--one of the major goals of any education in the sciences.