Specialized primary source holdings, not only manuscripts and books but also audio and moving images, are difficult to discover, often requiring users to navigate multiple search tools. These discovery challenges arguably lead to underutilization of specialized primary source holdings in the higher education curriculum. Faculty often include collections in their syllabi only if they have a direct relationship with an archivist or know of specific relevant collections. Similarly, archivists have the most success matching collections to courses when they have built individual relationships with professors, becoming familiar with course content. Particularly at a time when academic libraries are under increasing pressure to link their holdings to student outcomes, a new discovery paradigm to augment personal relationships is needed. This article suggests a conceptual model that would provide a mix of traditional methods and new data mining tools to increase access points to curricular content. The article consists of two parts: a review of existing methods, both human and computer, for connecting curriculum to library resources and a pilot of a software curriculum-to-collection crosswalk that matches course content to specialized primary source holdings via subject. The crosswalk creates recommendations of specialized primary source holdings relevant to specific courses for use by special collections librarians and archivists in working with faculty and students.