Flood Myths Compared: The Famous Flood Tablet Is a Late Copy

What George Smith read in 1872

Smith was sorting fragments from Kouyunjik, the mound opposite Mosul where Layard had found part of the royal Assyrian library. In The Chaldean Account of Genesis (1876), Smith recalled the moment: "my eye caught the statement that the ship rested on the mountains of Nizir, followed by the account of the sending forth of the dove." He pieced the story together from parts of three copies.

On December 3, 1872, he presented it to the Society of Biblical Archæology. E. A. Wallis Budge of the British Museum, writing in 1920, called him the scholar "who was the first to translate" the Tablet XI text. The museum's Ashurbanipal Library Project lists Smith's trips to Nineveh in 1873, 1874 and 1876 as following "his dramatic discovery of the Flood Tablet." He went to look for its missing pieces.

The project gives Ashurbanipal's reign as 668 to about 630 BC. In his introduction to a 1920 Yale volume he published with Albert T. Clay, Morris Jastrow Jr. called the Nineveh form of the epic "a late copy of a much older original." The palaces where those tablets were dug up are the subject of the Library of Ashurbanipal entry.

Nine flood texts, side by side

Each row below is one text. The date of its earliest surviving copy is kept apart from the date a named source gives for the text, since the two can lie centuries apart. The copy column gives the first-ink date in the sources checked. It marks where the clay or manuscript evidence starts, not where a story began.

Table data: flood-myths-flood-texts.csv, with the source and link for every row.

The tables set the texts side by side and make no claim that any one of them comes from another. How the site dates a record is set out in Following the Paper Trail. Only four texts have both a dated copy and a date for the text in the sources checked: Gilgamesh Tablet XI, Apollodorus, Ovid and Gylfaginning. The chart below leaves a missing date blank rather than guess it.

Ziusudra, Atrahasis, Utnapishtim

The Sumerian Flood Story survives on a broken tablet from Nippur, CBS 10673 + CBS 10867. The Cuneiform Digital Library Initiative dates it to the Old Babylonian period, about 1900–1600 BC. Arno Poebel published it in 1914. In the English translation by the Electronic Text Corpus of Sumerian Literature, gaps of roughly 32 to 39 lines break the story.

What survives tells of Zi-ud-sura, a king and priest, who hears a warning at a wall and rides out seven days and seven nights of flood in a very large boat. Afterward the gods grant him eternal life in Dilmun.

Atrahasis is an Akkadian poem named for its hero. As English Wikipedia summarizes it, the oldest known copy is dated by its colophon, the scribe's note at the end of a tablet, to the reign of Ammi-saduqa of Babylon. Its article on that king gives the reign as 1646–1626 BC or 1638–1618 BC, depending on the chronology. In its summary of the poem, the god Enki gives the warning by speaking to the wall of the hero's reed hut.

That summary also notes that the Atrahasis tablets break off where Gilgamesh Tablet XI has its birds. Older books give older numbers for the same reign. In 1920 Jastrow placed a deluge fragment from Ammi-saduqa's eleventh year at "c. 1967 B.C." and noted that it names the hero Atra-hasis. Budge, who put it at "about B.C. 2000," added that its colophon marks it as the second tablet of a series that was not the Gilgamesh epic.

Tablet XI of the Epic of Gilgamesh has a survivor named Uta-napishtim, who tells Gilgamesh how the god Ea warned him through the wall of a reed house. Near the end, Ea says in Budge's 1920 translation, "I made Atra-hasis to see a vision," so at line 196 the hero carries the older poem's name.

Budge argued that the flood story "had not originally any connection" with Gilgamesh and was added by a late editor. English Wikipedia's summary, drawing on Jeffrey H. Tigay's book, lists lines such as "Wall, listen to me" in Atrahasis that were copied more or less directly into Tablet XI.

Budge's translation gives the birds at lines 147–155:

I brought out a dove and let her go free. The dove flew away and [then] came back; Because she had no place to alight on she came back. I brought out a swallow and let her go free. The swallow flew away and [then] came back; Because she had no place to alight on she came back. I brought out a raven and let her go free. The raven flew away, she saw the sinking waters. She ate, she pecked in the ground, she croaked, she came not back.

Genesis 6–9

God tells Noah, "Make thee an ark of gopher wood," and sets its size at "three hundred cubits" long, "fifty cubits" broad and "thirty cubits" high (Genesis 6:14–15, American Standard Version, 1901). It comes to rest "upon the mountains of Ararat" (8:4).

Noah sends out a raven, which "went forth to and fro" (8:7). Then he sends a dove three times. It comes back with nothing, then with "an olive-leaf plucked off," and the third time "returned not again unto him any more" (8:8–12). The flood ends with a covenant whose token is "my bow in the cloud" (9:13).

English Wikipedia's article on the Genesis flood narrative, read on 2026-10-18, summarizes the prevailing scholarly view that the story was adapted from older Mesopotamian flood traditions, especially the Epic of Gilgamesh. The same article describes Genesis 6–9 as two sources woven together, one Priestly and one not, and notes scholars who question that division.

Its article on the Book of Genesis reports a 21st-century consensus that the book was composed after the Babylonian captivity, possibly in the fifth century BC. It also reports that Ronald Hendel and Aaron Hornkohl propose a date before 550 BC, and that Russell Gmirkin argues for the late 270s BC.

Deucalion and Pyrrha

Ovid's Metamorphoses, a Latin poem of AD 8, has Jupiter flood the world while Neptune sends the rivers over the land. Deucalion and Pyrrha reach Parnassus "in a small and fragile boat." Jupiter ends the flood when he sees the two of them left (Book 1, Brookes More's 1922 translation).

The goddess Themis tells them to throw "the bones of your great mother" behind them. Deucalion reads the riddle as the stones of the Earth. The stones they throw become people.

The Greek Library attributed to Apollodorus tells the story another way. English Wikipedia dates this handbook of myths to the first or second century AD and traces its surviving manuscripts to a 14th-century one in Paris. In this version Prometheus advises his son Deucalion to build a chest. Deucalion floats in it "over the sea for nine days and as many nights" before landing on Parnassus (1.7.2, J. G. Frazer's 1921 translation).

After a sacrifice to Zeus, the stones that Deucalion and Pyrrha throw over their heads become men and women. Neither the Latin nor the Greek version sends out birds.

The same Library records a second flood sent as a punishment. Cassiepea, wife of the Ethiopian king Cepheus, boasted that she was more beautiful than the Nereids. Then "Poseidon, sharing their wrath, sent a flood and a monster to invade the land" (2.4.3). Her story is the subject of the Cassiopeia entry, where Poseidon's flood stands beside the flood stories in these tables as a parallel, not a source.

Manu's fish

The Satapatha-Brahmana is a Sanskrit ritual text. As English Wikipedia summarizes, A. B. Keith placed it in the eighth to sixth centuries BC and Asko Parpola around 700 BC, while its translator Julius Eggeling put its final written form at 300 BC. Its flood story sits inside a passage about an offering (I.8.1, Eggeling's 1882 translation, pp. 216–219).

A small fish comes into Manu's hands as he washes. It promises to save him from a flood if he raises it, first in a jar, then in a pit, then in the sea. Manu builds a ship, ties it to the fish's horn and reaches the northern mountain. The flood sweeps every other creature away, and a woman produced from his offering becomes the means of his descendants.

The text does not name its fish. The drawing below, dated about 1700–1725 or later, shows Matsya, the fish form of Vishnu.

Yu's channels, Ymir's blood

The Chinese Mengzi collects sayings of the philosopher Mencius and is thought to have been compiled around 300 BC. Its flood is one that nobody sails out. In the time of Yao, it says, "水逆行，泛濫於中國," "the waters flowed backward and spread over the central states" (Translation: Where Tales Began).

Then "禹掘地而注之海," "Yu dug the ground and led the water to the sea" (Translation: Where Tales Began). After that, people could live on level ground again (3B:9). The zh.wikisource text quoted here does not name the edition it follows.

Gylfaginning is part of the Prose Edda. Anderson's 1879 introduction names Snorri Sturluson, who died in 1241, and his nephew Olaf Thordsson, who died in 1259, as the men usually credited with the work. Its flood is made of blood. When the sons of Bor kill the giant Ymir (Ymer in Rasmus B. Anderson's translation), "there flowed so much blood from his wounds that they drowned therein the whole race of frost giants; excepting one, who escaped with his household" (section 7, in Anderson's chapter "The Creation of the World").

That survivor is Bergelmer: "He and his wife went on board his ark and saved themselves in it." New races of frost giants come from the pair. Anderson dates two parchment manuscripts of the Prose Edda to about 1300. Following Rudolf Keyser, he credits Gylfaginning to Snorri and dates that part to about 1230.

Why this table has no arrows

Among the sources checked, one pair of texts in the table has a published case for direct copying. It is Atrahasis and Gilgamesh Tablet XI, with their shared lines and Budge's 1920 view that the flood was added to the Gilgamesh series. For Genesis, the summaries above report a prevailing view of adaptation from Mesopotamian traditions, with the date still argued.

For the Greek, Roman, Indian, Chinese and Icelandic texts, the sources checked record similar episodes in different languages and centuries but no line of copying between them. The chart below places each text by region and period and draws no line between any two.

This entry treats the flood as an episode found in many texts, not as one story told many times. How folklorists number whole tales, and why a shared number does not prove a shared ancestor, is the subject of the Aarne–Thompson–Uther Index entry.