Through the Layers, a Glimmer of Archimedes

By Lawrence Biemiller

Baltimore -- It's the most unassuming of treasures: a plain, careworn prayer book that dates to the Byzantine empire's waning days and has since survived fire, forgery, and fungi (just barely, in the case of the fungi). Its pages are mottled and moldy, its ornaments are few, its handwritten Greek is illegible in places, and it was made from parchment that had been used before and then erased. You can just barely make out faint lines of an older text, complete with drawings, that runs perpendicular to the prayers, but those faint lines are the real treasure -- seven ancient works by the great Greek mathematician Archimedes, including two that are known from no other source and a third preserved elsewhere only in a later translation.

Classicists and mathematicians have known about the prayer book since 1907, when the Danish scholar Johan Ludvig Heiberg published a transcription that he prepared from his photographs of the manuscript, which was then at a monastery near Constantinople called the Metochion. One of the two unique texts, the Method Concerning Mechanical Theorems, was quickly recognized as a work of the first importance. Sometime after 1911, however, the manuscript vanished, and it did not reappear until 1998, when an unidentified French family sold it at auction for $2-million to an unnamed buyer. Soon afterward, the Walters Art Gallery here persuaded the buyer to allow the manuscript to be properly conserved in the gallery's lab, and also to permit computer-aided digital imaging of its pages.

Scholars working on the project -- from the Walters as well as from the Johns Hopkins University and the Rochester Institute of Technology -- now say that Heiberg was able to transcribe only about 80 to 90 percent of the Archimedes text, and that the latest technology will help them recover significantly more, including the original drawings. (Heiberg, a philologist from the University of Copenhagen, disregarded the drawings in the manuscript and had modern illustrations prepared for his transcription.) The current project, which is being paid for entirely by the manuscript's anonymous owner, will take at least four years and will rely on digital cameras, high-tech microscopes, imaging software, and extensive scholarly analysis. It is expected to produce a substantial database of images that classicists and mathematicians will be able to work from, along with a re-creation, in book form, of the Archimedes volume as it might have originally appeared.

Archimedes was born in Syracuse, on the island of Sicily, sometime between 290 and 280 B.C. His life is the subject of a number of stories that scholars say are probably apocryphal -- for instance, he most likely didn't run naked down the street shouting "Eureka!" after figuring out in the bathtub how to measure the gold content of a crown without destroying it. He did, however, publish a number of mathematical manuscripts. The best known is On Floating Bodies, which defines buoyancy and goes on to accomplish mathematical feats that experts still describe as stunning. Archimedes died about 212 B.C. in the sacking of Syracuse by a Roman army.

Nine of his treatises have survived the centuries. Reviel Netz, an assistant professor of classics at Stanford University who is preparing a new English translation of Archimedes' works, says Archimedes was such a sophisticated mathematician that his writing would never have had a very large audience, and probably never existed in many copies. In early antiquity the copies would have been on papyrus scrolls; sometime after the invention of the book, the surviving texts were gathered between covers and codified.

The nine surviving works come from three sources, Mr. Netz says. One is the prayer book, known as a euchologion, and another is a single Latin translation, made in the 13th century, of a Greek text that was subsequently lost. The third source, also now lost, was copied several times in the 15th century before disappearing in the 16th. Four copies of it survive.

William Noel, who is the curator of manuscripts at the Walters and is spearheading the Archimedes project, says the seven texts that now appear as faint lines of underwriting in the prayer book were probably copied from an earlier compendium in the 10th century, during a Byzantine renaissance. But in the 12th century, when the empire was besieged and learning had grown scarce, the book was apparently judged useless. It was taken apart, and the ink was soaked or scraped from the parchment. The sheets were cut in two, prayers were copied onto them, and a new volume was assembled. The book is a palimpsest, to use the formal term for something written on pages previously erased.

"For 700 years this thing had a very rich life as a prayer book," Mr. Noel says. The palimpsest is believed to have been at one time in the library of the Greek Orthodox monastery at Mar Saba, in the desert of Judea, and later to have been part of the library of the Greek Patriarch in Jerusalem. Heiberg found the manuscript at the Metochion intact and in fairly good shape, although scorched around the edges. He photographed the sheets he identified as belonging to the Archimedes text, then returned to Copenhagen to pore over the photos with a magnifying glass. "Heiberg was a genius," says Mr. Noel, but even so he appears to have misidentified some pages, and there were lines he couldn't read because the book was still bound.

Abigail Quandt, the Walters's senior conservator of manuscripts and rare books, currently has custody of the palimpsest. Its pages are blotched with pale purple mold stains, and in places they are riddled with fine holes. Sometime after Heiberg photographed it, "the book was stored in very damp conditions," says Ms. Quandt, who is documenting the condition of the book prior to disassembling it for imaging. "Mold caused great damage to the parchment and ink. That's the really sad thing -- the worst damage was in modern times." How the damage occurred is still unclear.

According to the French family that sold the book in 1998, a family member bought the volume in the 1920's. But Ms. Quandt says four forged illuminations painted into the book were made after 1929 -- they appear to have been traced directly from a book of plates published that year. The full-page paintings, of Matthew, Mark, Luke, and John, might have been added because someone wanted to pass the volume off as the Gospels, rather than a more ordinary prayer book.

Equally intriguing is the history of the Archimedes texts, says Mr. Netz, of Stanford. Early mathematicians frequently made notes in the margins of manuscripts to explain points in the text, he says, and, because the notes were helpful, subsequent copyists often incorporated them into the body of the work. Heiberg enclosed in brackets passages he thought were not Archimedes' own, based at least in part on his estimation of the mathematician's character, but Mr. Netz says that by today's standards Heiberg was too quick with some of his judgments and that he may have bracketed too much.

At this late date, there's no hope of seeing Archimedes' works "unmediated," as Mr. Netz puts it. But he and other scholars think the palimpsest's versions of the treatises -- even though they were copied down 12 centuries after the mathematician died -- may come reasonably close to what Archimedes intended. Thus the interest in the drawings Heiberg ignored, which are clearly intended to be illustrative rather than precise. The drawings are also interesting to historians of science because they illustrate how knowledge was transmitted in earlier centuries.

Last year competing teams from Johns Hopkins and R.I.T. tried out various imaging techniques on five sheets of the palimpsest. By working with a variety of filters and with sophisticated software, the two teams were able to produce images far clearer than the originals. Because the two teams have different strengths, gallery officials have asked them to work together on the project, and imaging is expected to begin in earnest in April."We will not get a perfect recovery -- nobody expects that," says William A. Christens-Barry, a physicist who led the Hopkins team and who is a research associate professor in urology and oncology. But he says the new images will pick up many lines Heiberg missed. "Some pages will be easy," he says. "Some will be very, very hard."

Mr. Noel notes that, while classicists and mathematicians are eager to see the Archimedes text, other scholars are interested in the euchologion as a religious work. "A 12th-century prayer book we treat with the greatest respect. It's important in its own right, and it's integral to the history of the manuscript." He adds that whoever cut up the original text almost certainly spared it a worse fate. "As a useless manuscript it could have been used for any number of things," says Mr. Noel. "He probably saved Archimedes by putting him in a Christian disguise."

Copyright © 2001 by The Chronicle of Higher Education. Published January 26, 2001.