Shelved for Good: The Quiet Dismantling of Britain's Specialist Science Libraries

In the autumn of 2022, the library of a prominent London medical school sent a letter to its academic staff. The tone was administrative, the language measured. It informed readers that the physical collection — accumulated over more than a century — would be "rationalised" in line with the institution's digital transition strategy. Bound volumes of nineteenth-century clinical journals, annotated monographs, and grey literature that had never been digitised would be offered first to partner institutions, then to specialist dealers, and finally, if unclaimed, disposed of. The word "disposed" appeared twice. The word "irreplaceable" did not appear at all.

This episode is not unusual. Across Britain, specialist science libraries — those attached to university departments, research institutes, learned societies, and NHS trusts — are being closed, consolidated, or gutted of their physical holdings with a frequency that has begun to alarm archivists, research librarians, and historians of science alike. What is unfolding is not merely a logistical adjustment to the digital age. It is, many argue, the systematic dismantling of an infrastructure upon which serious scientific inquiry has long depended.

The Arithmetic of Closure

The pressures driving library closures are well understood, even if their consequences are not. British universities have faced sustained financial strain since the 2008 financial crisis, intensified by the shift to tuition-fee-dependent funding models and, more recently, by inflation eroding the real value of student fee income. Library budgets, which rarely attract the political visibility of teaching posts or research grants, have been among the first casualties of institutional belt-tightening.

Between 2010 and 2023, the number of professional librarian posts in UK higher education fell by approximately 30 per cent, according to figures compiled by the Chartered Institute of Library and Information Professionals. Specialist subject librarians — those with postgraduate training in both librarianship and a scientific discipline — have been disproportionately affected. Many institutions have replaced them with generalist information advisers, or eliminated dedicated liaison roles entirely in favour of centralised digital support services.

The physical collections themselves have fared no better. The logic of digitisation, compelling in its simplicity, holds that if a text exists in electronic form, its print counterpart is redundant. Subscription to large journal aggregators such as JSTOR or Elsevier's ScienceDirect appears, on a spreadsheet, to render entire periodical runs superfluous. Space freed from shelving can be repurposed as collaborative study zones, hot-desking areas, or — as has occurred at several post-1992 universities — commercial lettings.

What this arithmetic consistently fails to account for is the substantial proportion of scientific literature that has never been, and may never be, digitised.

The Dark Matter of Scientific Knowledge

Estimates vary, but scholars working in the history and philosophy of science suggest that between 40 and 60 per cent of pre-1990 scientific literature remains unavailable in any reliable digital format. This includes not only obscure Victorian monographs but entire runs of specialist society publications, conference proceedings, technical reports, and government-commissioned research that was printed in limited quantities and distributed narrowly. For fields with long temporal baselines — ecology, geology, epidemiology, pharmacology — this material is not merely historically interesting. It is scientifically active.

Dr Fiona Calloway, a research librarian at a Russell Group university who asked that her institution not be named, describes the practical implications with some urgency. "When a researcher is working on, say, the long-term toxicology of an industrial compound, they may need access to studies published between 1955 and 1975 that simply do not exist online," she explains. "Those studies live in physical collections. If those collections are gone, the researcher either doesn't know the prior work exists, or they spend months trying to locate a surviving copy elsewhere. In many cases, they simply proceed without it."

The consequences for reproducibility and systematic review are significant. A literature review that cannot access the full historical record is, by definition, incomplete. Yet the structural incentives of contemporary research — short grant cycles, publication metrics, institutional audit frameworks — rarely reward the kind of deep archival labour that locating such material demands.

The Expertise That Cannot Be Indexed

Perhaps more difficult to quantify, and therefore more difficult to defend in institutional budget discussions, is the loss of specialist librarian expertise itself. The trained science librarian occupies an unusual intellectual position: fluent in the organisational logic of a discipline, familiar with its major debates and methodological traditions, and skilled at navigating the uneven, sometimes chaotic landscape of scientific publication across different eras and formats.

This expertise is not replicated by discovery platforms or AI-assisted search tools, however sophisticated. Commercial database algorithms are optimised for relevance within indexed material; they cannot surface what has not been indexed, nor can they advise a researcher on the intellectual provenance of a contested dataset, the reliability of a particular journal in a given decade, or the existence of an unpublished technical report held in a filing cabinet three floors below.

"A good science librarian is part archivist, part methodologist, part intellectual historian," says Professor James Hartley of the University of Manchester's School of Environment, Education and Development, who has written on research infrastructure. "They are one of the few people in an institution who has a genuinely panoramic view of the literature. When that role disappears, researchers don't always know what they're missing, because the person who would have told them is no longer there."

Institutions Pushing Back

Not every institution has acquiesced. The Radcliffe Science Library in Oxford, part of the Bodleian Libraries system, has maintained its specialist collections and subject librarian model, supported in part by the university's endowment and its explicit commitment to research depth over access convenience. The library's science librarians continue to offer one-to-one research consultations, manage complex inter-library loan requests for obscure historical material, and contribute to the intellectual life of their departments in ways that extend well beyond catalogue maintenance.

Similarly, the libraries of several learned societies — among them the Royal Society, the Linnean Society, and the Geological Society of London — have resisted calls to digitise-and-disperse, recognising that their physical holdings constitute a form of disciplinary heritage with both scholarly and cultural significance.

Smaller institutions have found more creative solutions. A handful of post-1992 universities have entered into formal collection-sharing agreements with neighbouring institutions, preserving specialist holdings through distributed stewardship rather than individual ownership. Others have sought Heritage Lottery Fund support to digitise high-priority materials in-house, though such projects are resource-intensive and cover only a fraction of what is at risk.

A Policy Vacuum

What is conspicuously absent from this landscape is any coherent national policy framework. Unlike the legal deposit system, which mandates the preservation of published works at six designated libraries, there is no equivalent mechanism governing the specialist collections of universities or research institutions. When a university decides to disperse its chemistry library, it does so without regulatory constraint and, frequently, without any requirement to notify other institutions that might wish to acquire the material.

UKRI, the principal public funder of British research, has to date treated library infrastructure as an institutional rather than systemic concern. Research councils fund projects; they do not, as a rule, fund the archival conditions that make projects possible. The result is a structural blind spot in British science policy — one that will become increasingly visible as researchers in the coming decades attempt to build upon foundations that have been quietly removed.

The letter sent to that London medical school's academic staff concluded with an assurance that "access to scientific literature will be fully maintained through digital channels." It is a sentence that will read, in time, as either a promise or an epitaph — depending on how much of what was shelved is ever found again.