In this article, the author suggests some principles to be considered when attempting to assess learning in history courses. He has found it most effective to pursue these goals within the framework of the Decoding the Disciplines process. This approach, developed in the Indiana University Freshman Learning Project, suggests that faculty seeking to understand the learning processes in their courses can productively begin by defining "bottlenecks," i.e. places where large numbers of students have difficulty mastering some concept or action that is essential to success. Then the investigator can begin the intellectually demanding process of defining the steps or operations students would need to overcome the bottleneck. Generally, this requires a painstaking deconstruction of the processes professionals in the field employ automatically, and, like the exploration of other largely unconscious phenomena, it may require the assistance of others who are less involved with the material. Once the task at hand has been broken down into its component parts, each of these can be modeled for students, they can be given opportunities for practice and feedback, and the mastery of each operation can be assessed individually. The author traces the application of this process to two interrelated bottlenecks frequently encountered in history courses: (1) students' inability to find appropriate evidence to support an interpretation; and (2) their difficulty in making the connections between the evidence and the interpretation clear to their readers. These skills are absolutely essential to any history course that goes beyond simple memorization of facts, and yet they are not part of the skill set of many current college students. (Contains 2 footnotes.)