It sat in Egypt for fifteen centuries as the wrong book. Its 292 pages of vellum, 18 by 21 centimeters — about the size of a paperback — came from Saint Catherine's Monastery in the Sinai desert, and for most of the time anyone paid attention to it, what they saw was a Syriac copy of John Climacus's Ladder of Divine Ascent, a seventh-century monastic handbook. The book is called the Codex Climaci Rescriptus, and it now lives in Washington, at the Museum of the Bible. Rescriptus is the clue: written over. Underneath the Syriac, in a hand two centuries older and in a different language, there was another book.

Hipparchus of Nicaea worked in the second century BC. He is the astronomer who discovered, or at least first described, the precession of the equinoxes — the slow wobble that shifts the apparent position of the stars by about a degree per human lifetime. To notice that, you need a star catalogue of exceptional precision. Hipparchus built one. For most of the history of astronomy, that catalogue was treated as lost. Ptolemy, writing three hundred years later, produced the canonical one that would survive into the Middle Ages, and it was taken for granted that he had absorbed Hipparchus along with his own work. We carried Ptolemy forward. Hipparchus was a name.

Then, in 2022, a team using multispectral imaging found coordinates for the constellation Corona Borealis under the Syriac ink on a single folio of the codex. The coordinates were in equatorial form — the system modern astronomy uses — which had long been contested for Hipparchus. They were accurate, against the real sky, to within about one degree. That's tighter than Ptolemy. The erased text, copied around the fifth or sixth century AD, was a direct transmission of Hipparchus: the first anyone had read in close to two thousand years.

In January of this year, researchers took the manuscript to the Stanford Synchrotron Radiation Lightsource — a machine that bends electrons into curves so that they shed X-rays — and looked again. Multispectral imaging had worked on visible reflectance. The X-rays work on ink chemistry. The Syriac scribes used iron-based gall inks. The Greek scribes who carried Hipparchus forward had used an ink richer in calcium. Calcium fluoresces under certain X-ray energies; iron under others. By tuning the beam you can, in effect, turn one layer off and another on. What came up under the new pass was not only more coordinates. There were illustrations — shapes of constellations. Once the analysis is complete, the Codex Climaci Rescriptus is expected to be the most complete direct record of Hipparchus's observations in existence.

What they read, in translation, ran roughly: Corona Borealis, lying in the northern hemisphere — in length, 9°¼, from the first degree of Scorpius to 10°¼ within the same sign; in breadth, 6°¾, from 49° from the North Pole to 55°¾. The phrase that settled the long argument is "from the North Pole." Ptolemy, three centuries later, would measure star positions from the ecliptic — the Sun's path. Hipparchus, here, is measuring codeclination: degrees from the celestial pole. That is the equatorial frame, the frame modern astronomy still uses. A single preposition, preserved under an ink that didn't quite take, is worth more than the rest of the sentence: it tells us which sky Hipparchus was looking at, and how.

There is something strange about the survival mechanism here. The thing that saved the catalogue was its destruction. Parchment was expensive; recycling was common. A scribe in Syria in the seventh century needed a book. He scraped down what was already on the page — Greek he may not have read, numbers he had no use for — and wrote over it with pumice and a new pen. He did the job well enough that the overwrite has been durable, and not so well that the erasure was final. Under the hand-pressure that could reduce an ink to a shadow, calcium ions stayed in the vellum. The text became invisible. It did not become absent.

We lose books in various ways. Some burn. Some moulder. Some are thrown away. A few are saved by a kind of informational composting: overwritten by something else, carried forward as a second layer of the new book, reclassified as the wrong book, re-catalogued, re-shelved, re-photographed, re-imaged, re-tuned. A book can pass into the dark a long time and still be in the room. What is not yet readable is not the same as what is lost.

It seems like a reasonable week to start a commonplace book. I'm going to keep mine open on the table.

Sources: the multispectral imaging paper by Gysembergh, Williams & Zingg (2022); SLAC's January 2026 announcement of the synchrotron work; Scientific American for the ink chemistry detail; the Codex Climaci Rescriptus entry at Wikipedia for the manuscript's provenance.