Manuscript studies conducted online depend not only on facsimile digital images but on digital cataloguing and metadata. How can you find something you don’t already know? The “dark archives” in many cases are much of the digital archive, except for those portions one knows already by other means—print catalogues, previously published scholarship, personal acquaintance with specific libraries, etc. The challenges of standardized cataloguing and metadata reconciliation are well known in libraries, archives, and elsewhere. They have become more acute with digital catalogues and digital facsimiles. Ambitious projects of data aggregation and born-digital manuscript research will be defined by the quantity and quality of interoperable structured data. In this presentation, I illustrate these issues with a case study of more than four centuries of manuscript cataloguing at Parker Library, Corpus Christi College, Cambridge. I argue that in concert with the growth of digital images and data, manuscript studies need to integrate robust historiographies of cataloguing. Otherwise, digital resources become unreadable “dark archives.” I follow Tim Maly’s four-part typology of archival shades: known availables (the bright archive: information we know about and can access); known unavailables (dark to those who lack hardware, subscriptions, security clearance, etc.: information we know exists but can’t access); unknown availables (dark archives, information that exists but cannot be accessed for lack of indexing, retrieval technology, etc.); unknown unavailables (dark ontology, information you don’t even know you could know) (http://contentsmagazine.com/articles/dark-archives). As digital representations of medieval materials proliferate, they slip along this continuum from bright to dark. Parker Library provides for an exemplary illustration of this process because its manuscripts have been catalogued so many times—and because the manuscripts as well as the catalogues currently exist in digital formats. The catalogues range from Matthew Parker’s handwritten Register of 1574, through five print catalogues (1600-1912), to two editions of the digital platform Parker Library on the Web (2009, 2018). This succession of catalogues reveals the fallacy of technological progress: newer isn’t necessarily better. Metadata have been getting more fragmented, not more complete. Each catalogue, though, has contributed something distinctive even as it reflects vested interests in politics, classification, or technology. The most recent resource is the most mysterious unless compared to its predecessors. Medieval studies will probably always have a lot of “bad data.” And better data may never be created. In the meantime, digital resources can produce better “paradata”—that is, descriptions of how and why digital images and their metadata have been made so that users know what they are reading and what remains unread. The contours of the dark archive haveyet to be curated. Bibliography: https://aevum.space/sites/default/files/warren_dark_archives_references.pdf