An answer to your prayers that raises some searching questions

Life before Googling - how did we cope? These days, we take it for granted that the internet enables us to search the world's largest library from the comfort of our desks. Google, the most popular search engine, handles more than 200 million inquiries a day, many of them from academics who need quick access to an overview of an unfamiliar topic, a potted biography or even just a correct spelling.

A few years ago, it was fashionable to sniff at the web as a reliable source of information, as if ordinary libraries did not contain shelves full of unreliable and out-of-date material. After the rapid rise in the quality and quantity of general information on the web, most researchers now begin there when they are looking for answers to general academic queries. What, then, is the role of the staple of the reference library - the general reference book? Has it been made redundant by the web?

That question will be asked, implicitly or explicitly, by every potential purchaser of such books, including the Oxford Companion to the History of Modern Science . This mighty tome covers its subject from the Renaissance to the present day, and seeks, in the words of its unsigned preface, to be "modern in coverage, proper in demeanour, generous in breadth and cosmopolitan in scope".

These lofty aspirations, together with the breadth of its theme, make it easy to criticise. So let me say at the outset that this Companion is a superb piece of work, brimming with scholarly but easy-to-read gems. It must have cost its leading contributors a lot of work and - if this is typical of such reference texts - they will not have been well remunerated, nor will they gain much credit or publicity for the hours they invested in the project. They deserve the thanks and congratulations of everyone who benefits from their labours.

But how many scholars will use it, I wonder, compared with the number who access similar material on the web? Although I cannot reliably answer that question, I can comment on the efficacy of using this Companion compared with Googling. Over the past three months, I have been making the comparison.

The first point to be made in the Companion 's favour is that the standard of its articles is pleasingly high. Most are well written and informative, many are mini-masterpieces, and only a minority are turkeys. I came across very few misstatements, no serious factual errors and only a tiny number of typographical slips. Bearing in mind the amount of undigested and unreliable tosh on the web, there is no doubt that scholars stand a much better chance of being given an accurate brief on a topic here than by playing pot-luck with Google. This is the Companion 's cardinal virtue.

The author of many of the most rewarding pieces is, predictably, its editor-in-chief, John Heilbron, a distinguished historian of science, formerly professor of history and vice-chancellor of the University of California, Berkeley, now fellow of Worcester College, Oxford. Almost every one of his 33 entries is a model of how to summarise a field with authority, elegance and concision. Some of his entries even contain something missing from most of the others - humour. I especially liked his comment that Lord Rutherford is buried in Westminster Abbey, and in the periodic table of elements at number 104.

Heilbron, having recruited a crack team of consultants and advisers, has done a fine job of commissioning distinguished academics to write the entries. Each entry is signed and comes with a short and helpful list of supplementary reading. Among the contributors who have done exceptional work here are Daniel Kevles (whose entries on Aids, biotechnology and the Sputnik are exquisitely crafted), Silvan Schweber (first-rate pieces on particle physics) and Kathryn Olesko (excellent on Nazi science and the historiography of science).

Among the entries are 100 biographical sketches, most of them about a leading scientist, though some cover a group. It is not clear that these sketches are a good use of space, as all of the figures profiled are covered well elsewhere, notably in the compendious Dictionary of Scientific Biography . But it must be said that the Companion 's potted biographies are of a strikingly high standard, exemplified by Arturo Russo's delightfully written life of Enrico Fermi, one of the 20th century's most underrated scientists and an extremely rare example of a physicist who could hack it with the best experimenters and theoreticians.

Although the topics in the Companion have been generally well chosen, some are too broad. In some cases, this means they defy an authoritative summary (apparatus, anatomy, Africa); others we do not expect to find covered at all. Who would expect this book to feature four pages on Latin America and a piece of almost the same length on Europe and Russia? Given the limitations intrinsic to the searchability of the hardback, I fear many fine essays here will be read only as rewards for serendipitous browsing.

Such browsers will not be disappointed - the editing is consistently good.

Sensibly, the target audience appears to be the intelligent non-specialist who wants to mug up on an unfamiliar subject. Only occasionally do the different levels of related pieces jar - in the separate entries for Baconianism and Cartesianism, for example. Students seeking a quick insight into the contrast between these different approaches to science will be disappointed: they will find here learned and thoughtful pieces but will not be able to see the essential point that these two "isms" respectively embody the bottom-up and top-down approaches to scientific research, a vital point to grasp when trying to understand how science is done. A student could learn more about this after a minute's skilful Googling.

Novice surfers on the web soon learn that the most important advantage search engines have over hunting through books is the web's sheer quantity of searchable stored information. Although Google and its competitors sample only a small fraction of the information on the web, and even then display the results of a search in the form of an inordinately long undifferentiated list, this is preferable to being unable to find material at all. Although Heilbron and his colleagues present us with a remarkably wide range of material, some important contemporary topics - such as apoptosis (programmed cell death), volcanology, femtochemistry and string theory - are omitted, while others - such as the Big Bang and the role of science in modern drama - are treated only skimpily. Perhaps Oxford University Press should have employed a few more practising scientists to ensure that the content was truly up to date.

Another key aspect of the Companion 's accessibility concerns its portability. Its somewhat intimidating heft means that most people would normally expect it to use it near a desk; no one would want to carry it around. Here, Googling is once again asserting its authority with ever-greater force, as wireless connections increasingly become commonplace in the home and the workplace. In universities, it is now commonplace to see people searching for information on the web in libraries because it is quicker than trying to find it in the books on the shelves a few paces away.

Expensive reference books such as this one tend to be bought mainly by libraries. Few individuals will want to pay £80 for their own copy, although it is pretty good value as academic books go. The key question is: should libraries buy it? Even with budgets as tight as they are today, I would recommend every history of science library to acquire a copy, even if it made a little dent in their latest digitisation project.

Overall, I have to say that during the past three months of research, I have spent much more time Googling than referring to this Companion .

Although the book has some clear advantages over its electronic rival, my comparison has all but persuaded me that there is no future in publishing works of this generality only in book form. Indeed, my guess is that this may be the last general reference work in the history of modern science to be published in book form alone. If I am right, Heilbron and his collaborators should be remembered for providing the genre with a brilliant swansong. And I hope that, after the book has made its publishers a tidy profit, they will make its content available on the web. A book of gems like this is heaven-sent for discerning Googlers.

Graham Farmelo is senior research fellow at the Science Museum, London.