In 1960, at a conference on permanent/durable book paper sponsored by the ALA and the Virginia State Library, Robert E. Kingery (Chief of the Preparations Division of the New York Public Library) proposed, as part of the solution to the brittle book problem, "the cooperative storage of microfilms with member libraries obtaining electrostatic reproductions when needed."
In 1964, Gordon Williams reported to the Association of Research Libraries the results and conclusions of his investigation into the same problem. Since widespread deterioration of microfilm from microblemishes had just been discovered, and since he appreciated the value of originals over copies, he recommended a combination of deacidification and cold storage, with copies being made for use on demand, for every deteriorating record. A federally supported central agency would be responsible, and would keep a master list of copied (i.e., microfilmed) books. Masters would be stored centrally.
In 1976, when Mr. Williams attended the planning conference for a national preservation program in Washington, DC, at the Library of Congress, he recommended photocopying instead of microfilming to provide copies for readers, but he still recommended a central storage facility for the books. He knew it was a workable idea: after all, he was the Director of the Center for Research Libraries, which is a central storage facility for little-used books.
In an April 1990 report of a book condition survey undertaken for the Washington [DC] Research Library Consortium, the recommendations include not only the centralized storage of microfilm master negatives, but a shared preservation microfilming service and a shared storage facility for library materials.
In the July 1990 Newsletter of the Commission on Preservation and Access, there is a one-page special report, "The Concept of a Central Collection of Preservation Microfilm," which evaluates the accessibility of microfilms stored in 13 institutions, which together are expected to hold about half a million microfilmed titles by 1992. Very few could provide statistics of use, the report says, but the overall impression is one of low usage, which can probably be explained by the "wide range of costs, [lack of] timeliness in responding to requests, lack of machine-readable bibliographic records, and random selection practices." Other problem with both access and with measuring or keeping track of access are mentioned. In order to control costs and speed document delivery, a centralized distribution service managed by a third party is recommended, but no details are provided.