Each participating library generally decides which of its materials will be placed in the repository; however, the repositories, to varying degrees, have a hand in managing the intake of materials. All the repositories studied impose some restrictions on eligibility of materials for storage. Some guide the selection of materials for storage in a structured way, with the goal of reducing duplicative holdings and, in some cases, strategically building shared collections.

All repositories require that materials accepted be under at least minimal intellectual control. Materials must generally be included in an online catalog or integrated library system, and ideally this catalog is shared by the repository's member libraries. The sharing of a catalog or integrated library system permits the libraries to centralize certain processing operations such as assigning location and holdings information and bar coding.

The repositories impose few absolute embargoes; however, as a rule, they do prohibit storage of highly flammable materials such as nitrate film negatives, deteriorating or volatile materials, materials infested with mold or vermin, and materials that might be hazardous to the other collections.⁹ CONStor and the California repositories do not accept materials that are so damaged or deteriorated as to inhibit routine handling and delivery.

Other repositories exclude certain items because of their physical properties. To achieve maximum density and allow flexibility for placement of materials, most facilities accommodate materials that fall within a limited range of formats. For example, books and boxed archives must be stored vertically in containers of a narrow range of uniform sizes and shapes. The California regional library facilities do not allow realia, except for items that are integral to a particular book or archival collection. The SRLF also discourages deposit of materials in obsolete formats, such as pneumatic tapes and 5-1/4" floppy disks, which will deteriorate and hence probably never see use; this encourages libraries to explore conversion of those materials to usable formats.

Other pragmatic factors come into play when determining which items are appropriate for storage. The two main criteria are use and format. Low-use and "no-use" materials are favored candidates for transfer to repositories, on the assumption that their absence will pose minimal inconvenience to users. Faculty acceptance of the placement of research materials in off-site repositories is the primary factor in library decisions on deposits. The resident academic faculty, particularly those in the humanities and social sciences, are the primary users of the lesser-used materials that are frequently placed in repositories. For many faculty members, removal of materials from the campus library means the loss of immediate access to materials they need for research and teaching. Even though many university libraries no longer permit browsing in campus stacks, the impact of relocation is usually a serious consideration.

For this reason, many consortia build faculty consultation and involvement into decisions concerning materials to be deposited in the repository. The Five Colleges of Massachusetts, the California facilities, Duke, and CONStor all have formal mechanisms to involve faculty members in these decisions. Some libraries, for example, Duke, SWORD, and Columbia University, have devoted considerable effort to promoting faculty acceptance of repository storage by making them aware of the advantages of the facility and by building into workflows and services additional conveniences whenever possible. Despite such efforts, faculty reluctance to accept remote storage often acts as a brake on the rate at which some of the repositories are populated.

Selection of low-use materials for storage usually involves targeting certain large categories of materials, such as archives and other special-collections materials, older imprints, government documents, volumes of science journals that are no longer current, and foreign language materials. Materials that are easily handled in microform, such as archives, or materials in electronic form, such as JSTOR and Elsevier Science journals, are also prime candidates for selection. Such materials can be identified and expeditiously segregated from the rest of the collection.

Implementation and refinement of collection-management functions such as circulation tracking and control in integrated library systems enable libraries to identify low-use materials efficiently. Automated inventory control and retrieval systems make it possible to monitor use levels of materials that have been transferred to the storage facility. Nonetheless, identifying and isolating such materials can be extremely labor-intensive, and the cost of doing so is a major obstacle to making optimum use of the repositories.

A more viable option may be the "prospective" segregation of certain categories of materials. Under this system, materials are designated in advance and transferred to repository storage as the owning library receives them. Because these materials tend to be more frequently used, however, few libraries designate newly acquired materials per se for off-site or repository storage.

Repositories also provide an economical and practical means of storing problematic bodies of material, such as materials that are unavailable for use because of access restrictions imposed by donors or collections that have not yet been physically processed or prepared for use. These materials are often already in storage for one reason or another. ReCAP, for instance, holds some sealed papersā€”university archives and government archivesā€”that are closed to use for a specified number of years. These might also include special-collections materials that have only collection-level control and are not yet indexed.

Finally, deposit at the facilities provides relief for an on-campus space squeeze. Among the first materials that Duke relocated from on-campus libraries to its LSC were those that were displaced by construction and renovation projects.

3.1 Programmatic Selection Efforts

Some repositories actively manage the intake of materials to achieve goals that go beyond merely providing a place for low-use and no-use materials. One such goal is to reduce redundancy or duplication in their collections. The repositories take various approaches to this task. Some simply discourage libraries from including duplicates in their collection deposits. The California repositories, for instance, prohibit the placement of multiple copies of titles at the facility, although it is up to the library to check potential deposits against the facility's extant holdings before depositing.

CONStor not only discourages duplication among holdings stored at the facility but also helps eliminate duplicate materials before they are accepted. CONStor checks materials selected for storage by a participating library against its CONStor deposits through a central processing operation that serves all the participating libraries. The best copy is retained and placed in storage. It remains the property of the depositor. The inferior copy is returned to the owning library. Items selected for placement in the CONStor repository are publicized to the other institutions and to home campus faculty through Web pages and listservs.

While the effort to eliminate duplication stems in part from the desire to make the most economical use of space, such efforts also may stem from a desire to control redundancy in or rationalize management of the holdings of participating libraries. Such is the case, for example, in the Five Colleges of Massachusetts, CONStor, and California regional facilities. Rationalization may involve coordinating collecting responsibilities, negotiating collectively electronic-journal licensing, and assembling shared collections of record. The repository may fit into this scheme by serving as the locus, separate from any of the individual libraries, for assembling, as it does at CONStor and Five Colleges, shared last-copy and copy-of-record collections.

This is possible only in systems or consortia where the repository program is closely linked with the collection-development and preservation programs of the participating libraries. Such a connection became possible at Massachusetts's Five Colleges and Ohio's CONStor because both repository efforts came about as part of broader joint collection management projects funded largely by The Andrew W. Mellon Foundation. This circumstance prompted the two consortia to shape their repositories as part of their larger collection-development and preservation strategies. The progress that some consortia have made in linking selection of materials for the repository to overall collection-development and preservation aims also has to do with the organizational structure underlying the consortium efforts. This is treated in section 4 of this report.

3.2 Withdrawal from Storage

The repository collections are considered relatively stable bodies of material. Repositories encourage the idea that materials moved to the facilities are intended for permanent storage. This notion is consistent with the principle that minimum maintenance and traffic promote cost-effective operations. Because a given library's materials are normally interfiled with those of other libraries at the repository, removal of large bodies or categories of materials that might be scattered about the repository is labor-intensive and costly.

Policies governing the removal of materials from storage by depositing libraries vary. Some repositories maintain "one-way door" policies; however, in a concession to real-world conditions, they will, under certain circumstances, permit materials to be removed and reintegrated into the original library's campus collections. For instance, an estimated 2,000 items are removed from the NRLF each year. Such transfers occur for a variety of reasons. In most cases, the reason is a substantial and constant rise in requests for the materials, or at least a spike in use during a brief period. Faculty members' requests for the return of materials are also honored. In other instances, renovations, expansions, and new construction give libraries more room to shelve materials on-site.

3.3 Implications for Collective Management

The common facilities, collection-management policies, and regimes described in the previous sections were designed to enable libraries to realize economies in the care and administration of their low-use collections. The repositories, in subjecting the collections to many of the same procedures and conditions of service, achieve a high degree of coordination among the depositing libraries, managing the collections as a single entity with respect to access and control.

When certain collection-management functions are merged or performed centrally under the auspices of a consortium, the library relinquishes a measure of control over the collections, even though it may retain ownership of them.

Encouraged by the Office of the President at the UC, librarians have begun to discuss the concept of "shared collections." The proposed UC definition of shared collections does not address the issue of ownership, but allows campuses to decide whether to deposit an item into the shared collection. For the California libraries, such a limited sharing of collection materials "prospectively" might be enabled around electronic journals, where joint licensing, a form of resource sharing, has yielded economic and logistical benefits for the system. The California Digital Library, under the auspices of the University of California Office of the President, negotiated a university-wide contract for the digital database of Elsevier and Association for Computing Machinery titles that includes a limited number of print copies of each title. The print issues are to be part of a shared corpus of materials managed under the UC library system and stored at the SRLF from time of receipt.¹⁰

The Five Colleges of Massachusetts have actually merged ownership in some repository collection materials. Materials from the four private colleges in the consortium (Amherst, Mt. Holyoke, Hampshire, and Smith) that are placed at the repository become the property of Five Colleges, Inc., with one exception: Amherst maintains a separate collection of its own materials at the facility. In addition, materials owned by the University of Massachusetts Amherst, the fifth member of the consortium, remain under that university's ownership even after deposit at the facility, as required by state law.

3.4 Collection Ownership

There are important differences between common management of collections and shared ownership. For example, depositing libraries retain the right to withdraw their materials from the repository under the former arrangement, even though this may be difficult to accomplish on a large scale. (None of the repositories surveyed had yet received a request for wholesale removal of a depositor's collections.) Ownership of collectively managed materials nonetheless continues to be a volatile issue, particularly for large libraries whose stature in the community of American research libraries is closely linked to the number of volumes they own. But within the context of the repositories, the practical distinction between shared management of a body of materials and actual ownership can become difficult to make.

Ownership aside, the experience of the Five Colleges and others suggests that the retention of redundant materials by individual libraries can be far less costly when there is a shared "active" copy that is cooperatively managed for long-term retention and accessibility. The Five Colleges consortium identifies and retains the best copy of titles from among the member libraries' holdings; the second and third copies are returned to owning libraries for disposition as they see fit. In this way, the repository becomes the locus for de-duplication of shared holdings. The inactive copies can be disposed of or kept in less-expensive "dark" storage.

In this respect, the issue of ownership is something of a red herring. Control, rather than ownership, is the factor that affects economics of these ventures. It is in the cooperative management of the materials that economies and rationalization of resources are realized. If materials are managed cooperatively, that is, subject to uniform policies, services, and rights, then many of the inefficiencies and redundancies that otherwise accompany single party ownership can be avoided.