Libraries have traditionally built local collections to serve local patrons and have acquired material to meet the needs of their communities. For academic libraries, that has meant a focus on resources for teaching and research. Preserving the collections has been synonymous with preserving the information contained in them. Because libraries acquire physical objects for use and reuse (often under the “first-sale” doctrine of the United States Copyright Code), this has meant that they own physical property rights to the items on their shelves. (They rarely own the intellectual property rights.) They thus are not only preserving information but also preserving institutional property. While this may appear self-evident, it becomes an important economic consideration when libraries begin collecting materials too fragile to save for long, such as videotapes, or materials that have no physical form, such as electronic databases. It also matters greatly when libraries are urged to preserve resources that are of national importance but that receive little local use, as was the case in the brittle-books microfilming program of the 1980s and 1990s. When do national priorities trump local ones, especially when the funding is locally based? And how does funding affect preservation selection priorities?
Building Collections: Libraries and Archives
Research libraries, no matter how large, collect only a small portion of all the information created and disseminated at any given time. In the nineteenth and early twentieth centuries, most academic libraries were shaped directly by the research needs of their faculties. This resulted in some very rich veins of bibliographical ore, but it also generally produced holdings that were deep, but not broad, in coverage.
With the growth of all sectors of higher education after World War II, libraries and their collections boomed. Collecting became somewhat regularized, professionalized, and far more ambitious. Many libraries at large, research-oriented institutions made concerted efforts to collect in virtually all areas that their academic departments covered. Research collections were critical in recruiting graduate students as well as faculty, and universities scaled up their collecting activities to provide on-site access to their users.
Libraries selectively acquire currently produced items that will become the primary documents of tomorrow. To measure how the growth in book publishing has made selection more difficult, one has only to note the substantial increase in the number of books published in the United States since 1880.
Historical Statistics of the United States: Colonial Times to 1970 (U.S. Department of Commerce 1989) and the Statistical Abstract of the United States (U.S. Census Bureau 2001) give the following annual output figures for the first year of each decade between 1880 and 1990:
These figures represent only titles produced in the United States. Research libraries, however, typically acquire titles from countries around the world. In 19921993, the United Kingdom published 86,573 new titles; by 19951996, the figure had increased by 24 percent, to 107,263. During the same period, Japan increased its book production 58 percent (from 35,496 to 56,221); Russia, 26 percent (from 28,716 to 36,237); and Italy, 20 percent (from 29,351 to 35,236) (Association of Research Libraries 1999b, 10). A recent analysis of library purchases shows that acquisition of foreign materials is holding steady, but the portion of imprints collected of the total published is declining (Association of Research Libraries 1999b, 10). The worldwide output is roughly 600,000, of which the average Association of Research Libraries (ARL) library purchases about 50,000.
Most materials in these collections, as well as those in all academic libraries, are text based. The collections are built along disciplinary lines and are created to serve research and teaching needs. It has fallen to specialized repositories to collect the plethora of information of research value that libraries do not routinely acquire. Such groups as historical societies and special libraries (e.g., the American Antiquarian Society, the Newberry Library) tend to collect in specific topics (e.g., American history, regional history, travelers’ accounts) and formats (e.g., manuscripts, maps, photographs, musical scores), and to serve as critical supplements to the research libraries on campus. (There are quite a few specialized libraries on campuses as well.) In contrast to research libraries, many of the specialized repositories collect chiefly primary sources. They are distinguished from the book collections of libraries because they are usually unpublished (e.g., manuscripts) or rare (e.g., broadsides); furthermore, only the secondary literature that is germane to those collecting strengths is acquired. Their collecting policies are shaped by research needs, to the exclusion of curricular ones.
Despite the recent appropriation of “archive” as a verb to mean “to store” or “to preserve,” the traditional meaning of archives as a noun is narrower. Archives are institutions that collect records created in the course of an activity-for example, business records or government records. The collection strategies of archives are different from those of libraries. The National Archives and Records Administration, for example, is charged to collect, organize, and make available the records of the federal government. It collects information created by the government in the course of its business. This information includes not only paper and electronic records but also films, photographs, posters, and other visual and sound materials. Archives tend to merge the function of record management and true archiving for long-term access. They may acquire large caches of records that they will retain for stated periods of time for legal reasons. However, the average archives will accession for permanent retention only 1 to 5 percent of those records. During a process known as appraisal, they determine which files have historical value, and they keep only those materials.
Repositories that specialize in non-print materials are often called archives; however, film and sound archives do not have records in the technical sense because they are not documents created in the course of business. Their collections are scaled in an archival way: they are large and often described at the collection level. These special-format archives-some affiliated with universities, some with museums, some independent-are also a crucial part of the information landscape. They must play leading roles in any strategies the task force proposes to ensure the preservation of and access to artifactual collections.
Shared Access and Shared Collections
Until recently, the only way in which a library could make a publication available to its users was to own it. If demand were great, a library would have to own more than one copy. With the advent of photocopying and faxing, together with robust interlibrary loan and document-delivery systems, libraries can provide access to an item without purchasing it. Nevertheless, libraries still have large collections that replicate the collections of other libraries, in part because of the constraints of dealing with physically fixed information.
One of the chief advantages of digital technology is its ability to overleap constraints of time and place and to deliver information at any time to any computer that is connected to the Internet. Now that libraries no longer face the physical and temporal constraints that existed before the introduction of electronic networks, the models of access are changing significantly. Libraries understand that students and scholars prize instant access. Students seek out assigned readings a few hours before they are expected to have read them. Faculty members track down bibliographical references under tight deadlines from publishers. As service organizations, libraries assign the physical locations of their collections on the basis of known patterns of use. The problems associated with keeping collections readily accessible have been highlighted in recent years, as more and more libraries are building offsite storage facilities for low-use materials. As economically compelling as offsite storage is, it succeeds only when faculty members are willing to recognize the time constraints built into the system (in many cases, a 24-hour wait between requesting the material and receiving it). But for resources such as journal articles, which can be delivered to the desktop at 3 a.m., how valuable to the scholar is the physical proximity of the hard copy? How much should a library spend to make both easily available? Faced with choosing between the two, how should the library decide which takes priority?
The new delivery technologies-from photocopy to interlibrary loan to digital access-are also having an effect on the substance of collection building. Do all libraries need to collect essentially the same materials, if they can be networked, either virtually or through interlibrary loan? Research library collections have been undergoing a certain homogenization, partly because of the spiraling costs of journals and partly because of the advent of buying consortia. Libraries have begun to admit that they can no longer aspire to collect comprehensively, both because of the escalating costs of journals and monographs and because of the increased quantities of materials. Between 1986 and 1996, the price of the average journal rose 147 percent, and that of a book by 63 percent. The number of interlibrary loans handled during the same period increased 102 percent. The access cost increased to an average of $30 per transaction. During the same period, the consumer price index increased by 41 percent. Citing these statistics, the Committee on Institutional Cooperation said in a 1998 press release, “We and our colleagues on the faculty must acknowledge that research universities can no longer afford to build comprehensive collections.” [Emphasis in original]. The committee called for new efforts of cooperative collecting (Big 12 1998).
The subject of cooperative collection development-the sharing of responsibility for important but low-use materials, many of them in foreign languages-is not new. When library collections first started growing rapidly after World War II, there was an attempt to share collecting responsibilities. When this attempt gave out, it was followed by a plan to build a national periodicals center in the 1970s. In the 1980s, the Research Libraries Group (RLG) developed a method, based on the Library of Congress classification scheme, that its members could use to inventory collections and determine areas of strength. The so-called Conspectus was an instrument that libraries used for a period of time to learn what others had collected, in part to reduce duplication of effort in certain specialized fields. Individual libraries declared their responsibility for collecting and preserving literature in areas they chose. During the heyday of filming brittle books, this inventory system was the basis for many decisions about what each institution would film. The Conspectus fell into desuetude in the 1990s because, among other reasons, libraries were unable to keep acquisition information current.
These and other attempts to coordinate acquisitions at the national level have failed. There are, however, examples of successful regional models, one of which is the Triangle Research Libraries Network in North Carolina. Other examples include the Five-College Depository Library in Massachusetts (see Section 4) and the Center for Research Libraries.
User studies have repeatedly shown that many volumes in research libraries have extremely low use and or no use at all. A 1979 study of one medium-size research library showed that “any given book purchased had only slightly better than one chance in two of ever being borrowed. As books on the shelves aged and did not circulate, the chances of their ever circulating diminished to as low as one in fifty. Journal use, in general, was also discovered to be low” (Branin, Groen, and Thorin 2000). Given this, how could cooperative collection efforts fail?
Among the reasons generally cited for the failure of such efforts are the pressure for each library to serve local needs and its reluctance to cede responsibility in any one area to another library. Moreover, until recently, there were not enough commonly available bibliographical records to make such cooperation workable. Even today, several ARL libraries have not converted significant portions of their card catalogs into online records.
The tensions between local and national priorities for acquisition are mirrored in selecting for preservation. How should libraries manage the conflict between the present needs of researchers and the needs of future, or even physically distant, users? The challenge is to find ways to address the needs of local service today with the needs of the future and national priorities in mind. What should be the selection criteria?
Best Practices for Preservation Treatments
The fundamental principles of preservation apply to all formats, from print to digital, but the techniques and costs of the various treatments differ dramatically. The purpose of library preservation is to ensure the present and future use of information in whatever form it has been recorded. Library conservation is not like museum conservation, which aims to make an object fit for essentially passive use, such as exhibition. Library materials can be heavily used and must withstand the risk of misuse.
The great bane of libraries and archives is that there are no media on which information has been recorded that do not decay. As the carrier decays, the information goes with it. The optimal treatment is to preserve both the information and the carrier. In an academic library, for example, the ideal way to save a text has been to save the book or journal or whatever it has been printed to. However, given the fragility of media such as wood-pulp paper, not to mention magnetic tape or onion-skin paper, this is not always possible.
A review of the diverse formats and media on which library collections are recorded led the task force to accept the following premises:
- All physical formats, from paper to magnetic tape, will decay over time.
- Physical handling compromises all physical formats.
- All copying from one physical medium to another, such as from book to photocopy or from LP to tape, results in some loss of information and will usually compromise the physical integrity of the original as well.
These are the facts that lead to a tension between preservation and access. In the world of physical objects, one is usually bought at the price of the other. The art of preservation is to minimize the risk of loss while continuing to keep collections usable for researchers. This entails the following:
- identifying the risks
- deciding what measures should be taken to mitigate those risks
- deciding which measures have priority
Little about preservation, other than its theory, is simple. Identifying risks can be tricky, and too often the risks become known only by the effects of decay, when it is too late to prevent loss. For example, we did not know much about the deleterious effects of light and humidity on wood-pulp paper until tests performed in the 1970s proved the connection. By that time, library collections were already full of books that had been printed on wood-pulp paper instead of the more durable, but also more expensive, rag paper. By the same token, we do not know the effects of car pollution on paper. Early test results indicate that car exhaust (as opposed to factory exhaust) is deleterious both to rag paper and wood-pulp paper (Commission on Preservation and Access 1997). If intervention is required to save an item or a collection of items, the timing of that action is also affected by factors that are hard to determine. As a rule of thumb, the newer the medium, the shorter its life span. Most new media, however, have not been tested well enough to make it possible to predict rates of loss; only when deterioration sets in do we begin to understand the life span. Who would have thought that the so-called safety film onto which nitrate film was transferred would itself deteriorate and fade in a few decades? And what does this mean for setting preservation priorities? Do we concentrate on stabilizing information on newer and more fragile media, such as reel-to-reel tape? Does that mean libraries should do so in lieu of deacidifying older books printed on acidic paper?
Other external factors enter into the calculation of risk: for example, funding and fashion, or supply and demand. Leaving aside money for the moment, demand for library materials that are robust enough to withstand use means that preservation departments must first serve the needs of their local clientele. Much of the time and effort of preservation departments go to keeping the collections in circulation. Materials that are more frequently used receive more attention than do low-use items. Materials that are not frequently used, no matter how short the expected life span of the medium, receive short shrift. The Brittle Books Program, which started in the late 1980s, is the exception that proves the rule. When the larger research community become aware of the limited life span of many nineteenth- and twentieth-century imprints, the U.S. Congress allocated funds to the National Endowment for the Humanities to preserve the information in these resources that are, by definition, low-use. This has unfortunately not been true of feature and documentary films of the first half of the twentieth century, whose value for research was recognized too late for rescue. This fate has almost befallen comic books from the 1930s, and it will soon befall oral histories and ethnographic recordings on cassette tapes.
Deciding what measures to take begins with assessing the need, looking at the costs of treatment options, and looking for value for money. The most cost-effective means by far of ensuring the fitness of collections is to prevent damage. Preventive care is the heart and soul of most library preservation programs. The most cost-effective treatments are those that can be applied globally; one example is ensuring proper storage conditions. The decay that comes from the “inherent vice” of all physical media can be retarded through proper environmental conditions both in storage and in use. (The environmental needs [e.g., ideal temperature and humidity] of paper-based materials differ from those of film-based materials, and optimal storage requires separate facilities for each material.) Other global preservation treatments include the following:
- emergency-preparedness and disaster-recovery plans
- research and testing of materials (such as bar-code stickers that do not cause damage)
- development of standards and specifications for treatments
- education of staff and users about proper handling techniques
Collection-level treatments that act primarily to stabilize materials include the following:
- proper housing (e.g., using protective coverings, Mylar sleeves, and acid-neutral folders, or boxing a book that is fragile)
- binding loose materials
Item-level treatments are more expensive than are collection-level treatments because they require not only treatment of individual items but additional, often times skilled, labor to assess items and select them from within large holdings. Item-level treatments include the following:
- preservation photocopying
- creating photographic surrogates
- remastering sound recordings and moving images
If preservation fails, then conservation is the next step.* Item-level treatments include the following:
- book conservation, which may include repairing both the interior and exterior, removing the old boards, reattaching pages, reassembling and resewing the text block, making a new binding, and creating a customized box to hold all the source materials;
- paper conservation, which may include washing, mending, leaf-casting, backing, deacidifying, and reattaching fragments; and
- photo conservation, which may entail stabilizing edges, applying emulsion, creating copy prints for access, humidifying, drying, flattening, and mending and matting vintage prints.
The purpose of conservation is to return an item to use, even if that use is very limited. Preservation experts commonly say that preservation without access is not a good use of resources. A rare book, such as the Bay Psalm Book, may be so deteriorated that it cannot be opened without causing further damage. The goal of intervention in such a case is to render the book usable-that is, capable of being opened, even partially, under certain circumstances. In nearly all cases, the goal is not to restore an item to its “original state.” That would generally entail recreating the book using modern materials with the same look and feel and technique. Such an approach compromises its authenticity. Age and deterioration are facts of life, and few preservation experts strive to return an object to its original form.
In part because funding for research and development efforts in library preservation has traditionally been so modest, treatments have been deployed in the past without proper testing, and these treatments themselves have turned out to be deleterious. Lamination is one example; the use of adhesive tape is another. Hence, contemporary experts honor two principles: first, do no harm; and second, do nothing irreversible. This means carefully documenting each treatment and testing materials to ensure that nothing damages original items.
In all formats, the guiding principle of selection for preservation treatment is to make an item fit for purpose. The act of choosing what gets treatment and deciding why is based on a constantly changing evaluation of the value, use, and condition of an item, and all of these factors are constrained by limited resources and inadequacies of technology. While the library and its personnel can be relied on to provide the expertise to assess condition, decide on treatments, track use, and so forth, the question of the changing valuation of the significance of artifacts for research and teaching is a matter that needs to be informed by scholars and teachers.
* “Preservation” is the generic term for all types of treatments, both preventive and corrective, that serve to stabilize items. “Conservation” usually refers to an item-level treatment that involves active intervention.