THE ARTICLE · 7 MIN
The Library of Alexandria is often said to have died in one great fire, set by one villain. Depending on who tells the story, the villain is Julius Caesar, a Roman emperor, a Christian mob or a Muslim caliph. The ancient sources do not support any single version. Here is what each one says.
One caution first. The historian Roger Bagnall, in a study of the evidence, writes of “our nearly total ignorance about it”, and notes: “There is no ancient account of the foundation of the Library.” Everything below should be read in that light.
48 BC: Caesar’s fire
Partly true In 48 BC, during the fighting in Alexandria, Julius Caesar set fire to ships in the harbour. Books did burn — but the writers who describe it disagree about how many, where, and even who started it.
- Caesar himself (Civil War III.111) writes that he burned the ships: “omnesque eas naves et reliquas, quae erant in navalibus, incendit” — he set fire to all those ships and the others in the dockyards (our translation). He says nothing about books. An author describing his own war has reasons to leave things out, so his silence is a hint, not proof.
- Seneca (about a century later): “Forty thousand books were burned at Alexandria” (On Peace of Mind 9.5, trans. Aubrey Stewart). He does not say the whole library burned.
- Plutarch (around AD 100): the fire “spread from the dockyards and destroyed the great library” (Life of Caesar 49.6, trans. Bernadotte Perrin). His translator added a note: “The destruction of the Library can have been only partial.”
- Aulus Gellius (2nd century AD) says the Ptolemies collected “nearly seven hundred thousand volumes”, burned “not intentionally or by anyone’s order, but accidentally by the auxiliary soldiers” (Attic Nights 7.17.3, trans. John C. Rolfe).
- Cassius Dio (early 3rd century AD) lists “the docks and the storehouses of grain among other buildings” as burned, “and also the library”, whose volumes, “it is said, were of the greatest number and excellence” (Roman History 42.38.2, trans. Earnest Cary).
- Orosius (early 5th century AD) says the fire burned 400,000 books “proximis forte aedibus condita” — that happened to be stored in nearby buildings (6.15.31, our translation). He adds that he had seen book cupboards in temples, which, he says, were emptied when they were plundered “by our own people in our own times” (6.15.32, our translation).
So the accounts pull in different directions. Plutarch names the great library, and Gellius says the Ptolemies’ collection was all burned. Dio’s Greek speaks of storehouses “of grain and of books” (Cary’s translation says “the library”) and gives the books’ number and quality only as hearsay; Seneca gives only a number; and Orosius places the burned books in buildings near the harbour.
Did the Library survive Caesar?
Partly true Two pieces of evidence show that the Museum and a substantial collection of books continued in Roman Alexandria. Neither names the palace Library itself:
- Strabo, the geographer who visited Alexandria about two decades later, describes the Museum as part of the royal palaces, with “a public walk, an Exedra with seats”, and an official “who formerly was appointed by the kings, but is now appointed by Caesar” (Geography 17.1.8, trans. H. L. Jones). He does not mention the Library.
- The emperor Domitian (ruled AD 81–96) restored libraries that had been destroyed by fire, Suetonius says, partly by “sending scribes to Alexandria to transcribe and correct them” (Suetonius, Domitian 20, trans. J. C. Rolfe).
Bagnall concludes there was “still some substantial library in Roman Alexandria”, and points out that belief in a Caesarian destruction “requires the uneconomical assumption that the Library was destroyed in the fire but the Museum was not.”
The 270s: the palace quarter destroyed
In the early 270s (272 or 273, depending on the scholar), fighting under the emperor Aurelian wrecked much of the palace district, where the Library had stood. The historian Ammianus Marcellinus says the city’s “walls were destroyed and she lost the greater part of the district called Bruchion” (22.16.15, trans. Rolfe).
Bagnall notes signs of a consensus forming that major destruction of the palace library most likely came in 273, but adds: “there is no direct evidence concerning the Library in the ancient sources for this”.
AD 391: the Serapeum
The Serapeum, a great temple of the god Serapis, was said to hold a second, “daughter” library. In 391 Christians destroyed the temple. Socrates Scholasticus records that “At the solicitation of Theophilus bishop of Alexandria the emperor issued an order at this time for the demolition of the heathen temples”, and “Then he destroyed the Serapeum” (Ecclesiastical History 5.16, trans. A. C. Zenos).
Disputed The three accounts of 391 that we read — Rufinus, Socrates, and the pagan writer Eunapius — do not mention books. Eunapius, who was hostile to the Christians, says “they demolished the temple of Serapis and made war against the temple offerings”, but says nothing about a library. Orosius, a Christian writing in the early 5th century, is the exception, though he names neither the Serapeum nor a year: he says the book cupboards he had seen in temples were emptied “by our own people in our own times”. Scholars read this differently. Bagnall leaves open whether a library survived until the Serapeum fell. Dirk Rohmann, in the journal Klio (2022), argues that Orosius “is our most important witness that the Christian mob destroyed not only the temple building, but also the library”. The evidence is thin enough that neither reading is proven.
The Caliph Omar story
Myth A famous story has the caliph Umar ordering the books burned after the Arab conquest of Egypt in 642. According to the historian Bernard Lewis:
- The earliest source “is a Baghdadi physician called ‘Abd al-Latif, who was in Egypt in 1203”, who mentions it only in passing.
- The full story appears in Ibn al-Qifti, who “wrote a history of learned men in about 1227”, and was later repeated by Barhebraeus, “the principal source used by Western historians”.
- It “first became known to Western scholarship in 1663”.
Even the earliest of these was written more than 550 years after the event. Historians have long doubted it. Edward Gibbon wrote: “For my own part, I am strongly tempted to deny both the fact and the consequences.” Alfred Butler, in 1902, noted that the story “makes its first appearance more than five hundred years after the event to which it relates”, and that a scholar who appears in it, “Philoponus, was dead long before the Saracens invaded Egypt”. Lewis also reports an objection others have raised: the story has the books heating the city’s four thousand bathhouses for almost six months, and to do that “a library of at least 14 million books would have been required.” Bagnall’s verdict on such stories: they “originate centuries after the fact and are surely fiction.”
How big was it? Nobody knows
Unverified The famous figures do not agree, and some come in different versions in different manuscripts:
| Figure | Source | What it counts |
|---|---|---|
| 40,000 | Seneca | books burned in Caesar’s war |
| 400,000 (some manuscripts: 40,000) | Orosius | books burned in buildings near the harbour |
| nearly 700,000 (some manuscripts: 70,000) | Aulus Gellius | books collected by the Ptolemies |
Bagnall’s conclusion is blunt: “the ancient figures for the size of the Library or the number of volumes lost in the Alexandrine War do not deserve any credence.” The actual numbers, he estimates, were “probably lower, perhaps by as much as one order of” magnitude.
What probably happened
Bagnall argues against looking for one culprit: “It is hard to give up villains, but it looks as if we must abandon the search for some individual or small group to blame.” Papyrus rolls decay, and without constant recopying a collection wears out; in his words, “Even without hostile action, then, the Library, or Libraries, of Alexandria would not have survived antiquity.”
How much each event contributed — Caesar’s fire, the destruction of the 270s, the loss of the Serapeum, and centuries of neglect — is still argued. What the scholars we read agree on is narrower: no single event destroyed all of Alexandria’s books.
Our reading
A library that died in one fire makes a better story, and each version comes with a villain that suits the teller. The evidence points somewhere less dramatic and more useful: collections survive only while someone keeps paying to copy, repair and use them.
Sources
- Roger S. Bagnall, “Alexandria: Library of Dreams”, Proceedings of the American Philosophical Society 146:4 (2002).
- Caesar, De Bello Civili III.111 (Latin text, The Latin Library).
- Seneca, De Tranquillitate Animi 9.5, trans. Aubrey Stewart (1900).
- Plutarch, Life of Caesar 49.6, trans. Bernadotte Perrin, Loeb (1919).
- Aulus Gellius, Attic Nights 7.17.3, trans. John C. Rolfe, Loeb (1927).
- Cassius Dio, Roman History 42.38.2, trans. Earnest Cary, Loeb (1916).
- Orosius, Historiae adversus paganos 6.15.31–32 (Latin text, The Latin Library).
- Strabo, Geography 17.1.8, trans. H. L. Jones, Loeb (1932).
- Suetonius, Domitian 20, trans. J. C. Rolfe, Loeb (1914).
- Ammianus Marcellinus 22.16, trans. John C. Rolfe, Loeb (1940).
- Socrates Scholasticus, Ecclesiastical History 5.16, trans. A. C. Zenos.
- Eunapius, Lives of the Philosophers and Sophists, trans. Wilmer Cave Wright, Loeb (1921).
- Rufinus, Church History 11.23, trans. Philip R. Amidon (1997).
- Bernard Lewis, “The Vanished Library”, letter, The New York Review of Books (27 September 1990).
- Dirk Rohmann, “The Destruction of the Serapeum of Alexandria, Its Library, and the Immediate Reactions”, Klio 104.1 (2022).
- Edward Gibbon, The History of the Decline and Fall of the Roman Empire, ch. 51.
- Alfred J. Butler, The Arab Conquest of Egypt (Clarendon Press, 1902), ch. XXV.
Checked September 2026.
Related: Medieval people did not think the Earth was flat · The iron maiden was never a medieval torture device
- history
- myths
- ancient history
- libraries