By Henry Petroski

Knopf. 279 pp. $26

A few years back, Henry Petroski -- a professor of engineering at Duke -- brought out a much admired account of the gradual development of the pencil. In retrospect, that book seems to have inaugurated our current vogue for sprightly accounts of various obsessive historical enterprises -- the discovery of longitude, the bloody trade in nutmeg, the mathematical implications of pi or zero, even the nefarious uses of the footnote and the making of the Oxford English Dictionary. Now, Petroski himself has returned, this time with an intensive look at a possibly even more unlikely subject: the bookshelf.

Make no mistake. The Book on the Bookshelf is neither a history of reading (see Alberto Manguel's splendid work of that title) nor a brief account of book-collecting (consult Nicholas Basbanes's capacious A Gentle Madness for that). No, this is a tightly focused overview of how, over the past two millennia, scrolls and bound volumes have been housed, arranged and cared for.

Normally, even the most well-disposed reader could be excused for assuming that such a topic might be just a teeny bit limited in appeal. Yet just as Nicholson Baker can pronounce a moving elegy to the card catalogue, so Petroski manages to support his personal feelings about book storage with telling historical anecdotes -- and to offset his engineer's fascination with load, sag, lighting and shelf construction with quite lyrical passages from that obsessive biblio-visionary, Melvil Dewey (he of the Decimal System). As our genial and learned guide notes, what makes the history of technology so fascinating is that "it not only teaches us about the way things used to be done, it also gives us perspective on how things are done today -- and how they most likely will be done in the future." As a result, The Book on the Bookshelf can be relished even by those who don't aspire to a degree in librarianship.

You do, I suppose, have to care about books. Certainly, anybody who buys novels and biographies and Washington memoirs and slender collections of poetry and kids classics soon recognizes that most inexorable and inflexible of truths: There are more books in the world than there are shelves to put them on. You can build bookcase after bookcase, you can periodically cull your library, you can adopt any system you like to keep the bibliomania under control, and none of it will make a bit of difference. The laws of physics are, after all, immutable: All book collections inevitably expand to overwhelm the space devoted to them.

It is just this fact that has largely powered the evolution of the bookshelf. For instance, during the late Middle Ages illuminated manuscripts, because of their rarity and value, were actually chained in place to sloped writing desks. But once Gutenberg started printing from moveable type, books suddenly became much more plentiful. (It's been estimated that more than 100,000 different titles were published during the 16th century.) As a result, "If a room that was designated as a library came to be filled with lecterns," and more space was needed, "the only alternatives were to expand the lecterns into another room or to modify them by building upward. It was the latter solution that was commonly adopted." Attached bookcases consequently sprung ceilingward from the writing desks; Petroski reproduces several photographs of old leather-bound volumes tightly packed on their shelves with rows of chains hanging down in front of them. However, as books grew even cheaper, there was soon little need to keep them manacled in place, so a space-saving librarian could eliminate the lectern part of this storage system, leaving what are essentially free-standing bookcases.

This is grossly simplified history, of course, and fails to convey some of the charm of Petroski's narrative (albeit one interrupted by occasional arias celebrating the engineering intricacies behind, say, stack construction). Did you know that books were once stored spine inward? For centuries bindings bore no titles and there was little need to shelve a volume in what seems the logical way to us; besides, chains were best attached to the front cover next to the foredge, the edge opposite the spine. With some natural affection, Petroski lingers over the architectural growth of the carrel, or library research cubicle, which developed by enclosing sections of the cloistered walks in medieval abbeys. He discusses, too, with the enthusiasm of an Abbot Suger, the paramount importance of natural light in the layout of any study or reading room. In one chapter he patiently analyzes the placement of the books in three Durer woodcuts of Saint Jerome. Elsewhere he even shows off fanciful book wheels and lazy Susans intended to aid the industrious monk or alchemist who wished to consult several tomes at once. Because of the book burnings that accompanied the destruction of Catholic religious houses in Henry VIII's England (and elsewhere), the pressure of overcrowded book rooms was sadly alleviated for most of a century. "The Protestant Reformation of the sixteenth century created a great disruption," Petroski concludes with wry understatement, "in the development of libraries and their furniture." In one particularly appealing excursus, our biblio-historian describes Samuel Pepys's library -- limited to 3,000 titles, arranged by size, on double-tiered shelves, in 12 book cases, with exact call numbers indicating each title's location.

The second half of The Book on the Bookshelf focuses largely on modern library stacks. We learn that tall narrow windows indicate the location of a building's main shelf holdings. Petroski traces the construction of the old British Museum Reading Room, built in a courtyard under a dome bigger than St. Peter's in Rome. In passing, he reminds us that a character in E.M. Forster's novel Howards End suffers a heart attack, overturns a bookcase and is smothered to death by the falling volumes: How many bookmen and women have worried about -- or tacitly yearned for -- such a death? Melvil Dewey, it turns out, was quite a visionary, even managing to foresee the moveable compact shelving found in modern research libraries: "A series of 100 double faces [double-sided bookshelves] could be swung together, with one 75 cm aisle for the series. On an average it would be necessary to move 25 cases to transfer the aisle to the face needed . . . "

In his final pages, Petroski reflects on compact discs, the Internet and various approaches to the digital storage of information. Wisely, he doesn't hazard anything beyond a wan guess or two about their impact on libraries during the next millennium. Instead, he ends his appealing survey with an appendix largely devoted to the advantages and disadvantages to 25 different systems for arranging one's books -- by author's last name, by size, by ISBN number, by publisher, by read/unread titles, by order of acquisition, etc.

Alas, Petroski for all his wisdom in these matters does overlook one widespread method of book arrangement -- what one might dub the domestic jumble, when restless book accumulation slowly sweeps aside vain efforts at logic and design. As Wallace Stevens once observed, "A great disorder is an order. . ." A truth that some of us know all too well.

Michael Dirda's e-mail address is