The “information explosion” sparked by digital technology has fostered an increasing awareness of the sheer mass of information available today in a variety of media, from traditional formats such as paper to the more recent film, optical, and magnetic formats.1 Institutions charged with collecting, storing, preserving, and making accessible recorded information are struggling to keep pace with the growth of information production, even though their brief is to collect only a portion of what is published and an even smaller portion of what is produced and disseminated in unpublished form. With so much information coming at us, how do we distinguish between what is of long-term value and what is ephemeral? And of that ephemera, what should be selected to preserve to ensure for the future a rich record of the present?

Libraries have struggled with these questions for a long time. Even aggressive collecting will not meet all present and future information needs. In 1875, the Librarian of Congress noted that the quarters the library occupied in the U.S. Capitol were inadequate for the burgeoning collections and, in a thinly veiled appeal to Congress for a new building, described the problems the library had with newspapers and journals.

Though carefully preserved and promptly bound for preservation, there is no longer the possibility of even receiving half the issues of these representative journals, so important in our current history and politics; and the time will soon come when a legislator in search of a fact, a date, a political article, or a table of statistics known to be in a certain newspaper at a certain date, will find it only at the bottom of a lofty pile of journals, all of which must be displaced before it can be reached. Besides the issues of the daily press, the periodicals which are taken under the copyright law or by subscription, embracing most of the monthly and quarterly magazines and reviews, accumulate with such rapidity that no device yet invented will long avail to produce them when wanted. (Cox 2000)

Information is of little long-term use if it is not preserved and made accessible through indexing or cataloging. At the same time, much of what has been collected and made accessible by libraries and archives admittedly receives little or no use. This leads some to argue that investments in preserving so much material that has so little demonstrated use are a waste of resources. This point of view is nothing new. In the 1890s, around the time when the Library of Congress was moving its collections into a new, purpose-built building, the consolidation of the Lenox and Astor Libraries and Tilden Foundation in New York to create the New York Public Library led to the collections being closely scrutinized. This move prompted the following letter to the New York Herald from a concerned citizen.

To the Editor of the Herald:

Before the Astor Library moves its quarters, it would do well to get rid of some scores of thousands of its untouched volumes in order to make room for readable books. When, after asking in vain for some work that is the talk of the town, I look up and around at the rows of shelves packed with “things in books’ clothing.” I find myself echoing the poet’s wish that

“. . . From the dead
Old Omar would pop forth his head,
And make a bonfire of them all.”

Like Charles Lamb, I can read anything that I call a book, but there is much in the Astor Library which, in my humble opinion, should be categoried as waste paper.

Clara Marshall
New York, April 6, 1895

Today, little has changed. There is ongoing tension between how much information is produced and how much can be acquired, preserved, and made accessible in meaningful ways. There is tension between those who think we should collect as broadly as possible to expand our research base and those who think too much information can impede one’s ability to find meaningful information. And there is always the question of who should be bearing the burden of whatever preservation society deems necessary.

Currently, the Library of Congress has more than 17 million books and 95 million manuscripts, films, photographs, maps, sound recordings, and other non-book items. The New York Public Library has 42 million items, including more than 14 million books. Despite these impressive numbers, these libraries hold only a portion of what is published, or of what is created but not published. The Library of Congress, for example, which receives items that are registered for U.S. copyright deposit as well as foreign materials, reports that approximately 31,000 items arrive at its doorstep each working day. Of these, about 7,000 are selected to become part of its permanent collections (LC 2001). If libraries, archives, or historical societies do not collect instances of recorded information, then the chance of their survival is slim. Loss is inevitable.

Although information overload is not a new problem, the introduction of digital technology onto campuses and into research libraries has fundamentally altered the information landscape and created problems for scholars and students that have potentially serious ramifications. The creation and dissemination of digital resources are creating new models of service and access, such as licensing rather than owning essential intellectual assets. The mutability of digital documents is redefining what constitutes a text. For example, are back issues of a journal that are in digital form simply a bunch of articles or a rich database? Moreover, the conversion of texts into searchable texts is resulting in increased interdisciplinary research, as researchers in one field serendipitously find resources that had heretofore been confined (in print) to another field.

Accompanying the trend toward networked information are others that at first seem paradoxical. For example, at the very time that more material is made available online and retrievable any time, anywhere, there is increased attention among scholars to original, unreformatted materials and an increased appreciation for the material aspects of these sources (Tanselle 1998). There are eruptions of public outcry when material losses in libraries and archives are discovered. Scholars demand increasing attention to an ever-expanding range of candidates for preservation, but library budgets simply cannot support those demands. Preservation has thus become an unfunded mandate, the more pernicious for often being implicit. Academic institutions have learned the huge costs of penny-wise facilities management and deferred maintenance. It is reasonable to fear that libraries are incurring future costs by deferring preservation.

There is, however, a crucial difference between deferring maintenance costs and deferring expenses for preservation. When we do not allocate sufficient funds for preservation, we face the probability that significant resources will be lost forever. Library collections are among the most valued of a research institution’s intellectual and cultural assets-assets that form a crucial part of what might be called “public goods.” Unfortunately, with library collections, as with other public goods, many of those who make claim to their use are not prepared to take responsibility for their well-being.

This report addresses the value of library collections, especially those in physical formats, to scholarship and teaching. It discusses the costs of the benefits these collections bring and the roles of each member of the various communities who have a claim on their use and a responsibility for their well-being. The report focuses on preservation-what it takes to ensure the present and future usability of collections. It is grounded in the recognition that without preservation today, there will be no access tomorrow.

Preservation is a critical part of good stewardship of our intellectual and cultural heritage. Its chief challenges at the turn of the twenty-first century are fourfold:

Quantity. Because of the relentless growth of research libraries and their collections, an immense number of research items demand resources to remain accessible. In 1999, the 121 member libraries of the Association of Research Libraries (ARL) reported owning a total of 462,965,204 volumes (ARL 2000). The three greatest periods of growth for research libraries occurred after World War I, after World War II, and during the 1990s. A look at 12 representative public university libraries during this period reveals a typical growth pattern.2 In 1907, these libraries held an average of 107,425 volumes. By 1961, that number had grown to 1,772,831, and by 1995, the average number of volumes per library was 5,334,620 (Molyneux 1998).

Stability of Media. Library collections exist in a variety of physical formats, all of which are vulnerable to some degree. As the rate of information production has increased, storage media have become more compact and efficient. However, miniaturization comes at the expense of stability and longevity. With the exception of preservation-quality microfilm, the new media of the twentieth century are more fragile than those of the nineteenth, including the infamous wood-pulp paper that has been known to deteriorate into flakes over time (Conway 1996). The media invented in the last 150 years to capture light and sound are generally extremely fragile, dependent on machines for playback, and subject to rapid technological obsolescence. For example, the wax cylinders on which are inscribed the earliest known recorded voices of Native Americans are susceptible to mold, heat, scratching, skin oils, and other physical trauma. Moreover, they are dependent on playback equipment for which no replacement parts are manufactured and that few technicians are able to repair. Nonetheless, the information on them is invaluable and should be saved for future generations.

Economics. Since 1993, preservation budgets in ARL libraries have remained flat. The number of staff assigned to preservation is at a 10-year low (Reed-Scott 1999). At the same time, the demand for access to original materials has grown, especially for access to those materials in special formats that often are at greatest risk from physical handling or environmental stress. Meanwhile, although the technology for reformatting for access has greatly improved, the funding for preservation continues to decrease. More money now goes to digital reformatting of items to provide access than to microfilming to preserve the low-use brittle books that are rotting on shelves.

Unknown and Unfixed Values of Artifacts. The most difficult challenge for libraries is deciding how to set priorities for preservation. As long as the claim on preservation resources exceeds the available funds, it will be necessary to select which materials will get treatment and which will not. The choices are made in the context of changing perceptions of value and the fluid dynamics of intellectual inquiry. The recent elevation of nineteenth-century popular imprints and ephemera to a status of high research value poses an exemplary challenge. Providing access to those items, which are often at high risk from embrittlement and routine physical handling, has put great strain on library resources. Knowing that the intellectual interests and research methodologies of scholars will change over time, research institutions have collected “just in case” there is a demand in the future, rather than “just in time” for current demand. Research libraries and archives are full of items that have not been consulted in decades, if ever, and for which future demand is unpredictable, yet which make their claim for preservation attention.

How do a library and its home institution make sound fiscal and intellectual decisions about what to preserve, when, for whom, and at what price? Despite the enormous collections of printed materials that have been amassed, entire categories of primary sources have disappeared before collecting institutions and their users understood their value. A notorious example of such neglect is the fact that 80 percent of all silent films made in the United States are gone without a trace. Fifty percent of films made in the nitrate era (that is, before 1950) have also perished. Among those extant, a significant portion are not well preserved. Given that the materials that have vanished were not well documented at the time of their creation, the full extent of this loss will never be known.

1.1 The Charge

To seek several perspectives on the importance of the artifact to academic inquiry and teaching and to propose strategies for addressing the problems just defined, the Council on Library and Information Resources (CLIR) created the Task Force on the Artifact in Library Collections in 1999. The task force was charged with answering a challenging question: What will constitute good stewardship of our intellectual and cultural heritage in the first decades of the twenty-first century?

Scholars and librarians have common interests in the collection-building process. Although scholars are routinely, and often intimately, engaged with librarians in the acquisitions and use phases of collection building, they are usually exposed only to specific aspects of preservation activities, such as testifying to the value of a collection proposed for preservation microfilming. Rarely do scholars have the opportunity to view the preservation function as a whole.

Recent vigorous debates in the scholarly community and in the public arena about the value of saving books and newspapers as artifacts, in addition to preserving their content through reformatting and extending access through digitization, have raised questions about the involvement of scholars in these decisions. Do scholars know what happens to original materials after preservation treatment? Do they participate in decision making about the disposition of original materials after preservation treatment or reformatting? Are provosts and other academic officers fully informed about the economic factors that may constrain a library’s ability to keep its collections fit for use? The task force was asked to engage in a systematic review of the scholarly community’s interests in this and to explore the broader question of how to develop a vision for building collections in a new information environment comprising both physical and digital materials.

The task force was asked to articulate a framework for making or evaluating institutional policies for the retention of published materials and archival or unpublished materials in their original form. The crucial questions associated with this task were as follows:

  • What qualities of an original are useful or necessary to retain in their original form? Under what circumstances are original materials required for research?
  • When is it sufficient and appropriate to capture intellectual content through reformatting and not necessarily to retain the original?
  • Which preservation options provide the most appropriate and cost-effective means of preserving the original?
  • From both custodial and scholarly perspectives, what are the advantages and disadvantages of these various preservation options?

Given that resources for preservation are finite and duplication of efforts can be costly, the task force was also asked to consider the advisability and feasibility of creating one or more national repositories into which one or more copies of all materials published in the United States would be deposited and permanently retained.

The task force was asked to interpret its charge broadly. It was asked to give primary consideration to print formats but also to consider the burgeoning numbers of non-print and electronic research sources that increasingly demand attention from preservation specialists. In fact, shortly after the task force convened, its members identified the problems facing non-print materials as being as important, certainly larger in scope, and probably more urgent than those facing print materials. The task force therefore gave almost equal consideration to three types of collection materials: print, analog audiovisual, and digital. Although technical constraints and considerations of preservation actions cannot be ignored, members were urged to focus on their implications for the research process, rather than on the technologies themselves. Preservation and access are technology-dependent, and best practices for ensuring long-term access to information will change over time.

1.2 The Work of the Task Force

In taking up its charge, the task force confirmed that the pressure on research libraries to keep up with the past while preparing for the future is of vital interest to scholars, governing boards, academic officers, and funders, as well as to librarians and archivists. Members agreed on several premises that would guide the task force’s work. Although these premises can be simply stated, each speaks to the uncertainty and dynamism that characterize the environment in which research libraries operate. The premises are as follows:

  • Information technology will continue to change rapidly.
  • Best practices for preservation and access will change.
  • Digital resources will increase significantly.
  • Scholarly research trends will change; in anticipation of such changes, scholars will continue to demand that collections be as inclusive as possible.
  • Intellectual property-rights management will evolve and must be respected.
  • Financial and human resources will not keep pace with demand.
  • Financial and human resources should be allocated in the most cost-effective manner to achieve an acceptable trade-off between quality of resource and expenditure of time and money.

Task force members began with a survey of current library practices in preservation and in collection development. They gathered information about what research libraries do and do not collect, how libraries will ensure access to those collections over time, and how they set priorities for investments in their collections, especially for preservation. (These findings are summarized in Appendix I.) Task force members then identified the particular needs of library materials in three areas: print, analog audiovisual, and digital. They determined that the ways in which these media are collected and preserved-and even which institutions do or do not engage in certain activities-are distinct issues that warrant separate consideration (see Section 3, States of the Artifact, 1800­2000). Throughout their inquiry, task force members returned to the simple but almost overwhelming fact that the web of resources that are useful for scholarship continues to grow at a rapid pace. More information is being produced and disseminated than before, and ever-widening orbits of intellectual inquiry call for ready access to these materials. While many of these sources, from business records to digital art, fall outside the traditional purview of library collecting, they are part of the larger information environment that influence public expectations of what libraries can and should be doing.

The task force relied on the testimony of experts from the library and archival fields for information ranging from the technical to the financial to the theoretical. On the basis of these findings and in consideration of the economic realities of funding preservation and access, the task force proposed definitions of responsible stewardship of research and cultural materials, and articulated the roles that each party in the research community and beyond plays in that stewardship (see Section 5, Summary and Recommendations).


1One study reports that the world produces a startling 2 billion gigabytes of new information a year, or roughly 250 megabytes for every man, woman, and child on earth (Lyman and Varian 2000).

2The libraries are University of California, Berkeley; University of Illinois, Urbana-Champaign; Indiana University; University of Iowa; University of Kansas; University of Michigan; University of Minnesota; University of Missouri; University of Nebraska; Ohio State University; University of Washington, Seattle; and the University of Wisconsin, Madison.