Viking Blog / History & Finds
The Knowledge Archive

The Codex Regius: 45 leaves on which the Poetic Edda depends

History & Finds The Codex Regius: 45 leaves on which the Poetic Edda depends

A small vellum book, written by a single unknown hand in Iceland in the late 13th century, with no title, no author's name and a hole of eight leaves in the middle of the text: that is the Codex Regius, shelfmark GKS 2365 4to. Nearly everything now called the “Poetic Edda” stands in this one manuscript. This article is about the book itself – how it is put together, what it lacks, how it reached a bishop in 1643, a king in 1662 and Iceland again in 1971, and what can and cannot be concluded from a single manuscript.

Detail of a darkened parchment page with dense Icelandic script and a hole in the vellum
Detail of folio 41r of the Codex Regius (GKS 2365 4to) with the beginning of Atlamál. The scribe wrote around the hole in the vellum. Image: Árni Magnússon Institute, Reykjavík. Public domain, via Wikimedia Commons.

One book, two names, one shelfmark

“Codex Regius” is Latin for “King's Book”. In Icelandic it is Konungsbók eddukvæða, the King's Book of the eddic poems. The name describes a former owner, not the contents: for a good three hundred years the manuscript lay in the Royal Library in Copenhagen. Its shelfmark dates from that time. GKS stands for Gammel kongelig samling, the “Old Royal Collection”, 2365 is the running number, and “4to” marks the quarto format. Editions and commentaries usually abbreviate the manuscript as R, less often as K.

What the Edda is as a whole, and how the Poetic Edda differs from Snorri's Edda, is covered in our overview. The subject here is the object itself: its quires, its scribe, its owners and its travels.

What the manuscript says about itself

The Árni Magnússon Institute in Reykjavík, which keeps the manuscript today, describes it briefly: a vellum book written in the late 13th century by an unknown scribe, 45 leaves thick. According to the institute, nothing is known of its origin or history until Bishop Brynjólfur Sveinsson wrote his monogram and the year 1643 in it. Between the writing and that entry lie more than three and a half centuries without a single report.

The date rests on the script, not on any note in the book. The Icelandic philologist Einar G. Pétursson, in his encyclopedia entry, dates the manuscript palaeographically – by the forms of its letters – to about 1270–1280; Terry Gunnell writes “c. 1270”, the English Wikipedia “the 1270s”. Pétursson also records that the book was written by one hand that is not known from any other manuscript. The scribe gives no name, place or date.

Nor does the book have a title. Guðbrandur Vigfússon, who described the manuscript for his edition of 1883, noted that there is no superscription in the original hand, “Edda” or otherwise. In the Middle Ages the name “Edda” belonged to Snorri's handbook; it was transferred to the collection of poems only in the 17th century.

The format is modest. Pétursson speaks of ninety small quarto pages, which matches 45 leaves. From its small size and its economical use of space Gunnell infers that contemporaries probably did not value the book as highly as, say, Möðruvallabók or Flateyjarbók, two large and lavishly planned compilations. That is an inference from how the book was made, not a recorded opinion. I found no exact measurements in the sources consulted and give none.

The scribe did not take the poems down by ear. Pétursson calls the Codex Regius a copy of an older manuscript that is now lost. The chief evidence is a second book, the fragment AM 748 I a 4to (more on it below). It shares several poems with the Codex Regius and is so close to it that both Pétursson and the institute assume a common written exemplar.

This matters for the age of the poems. The 1270s are only the date of this one copy. Behind it lies at least one written stage, and behind that, on the usual assumption, an oral one. How long that oral phase lasted the manuscript does not say. On the dating of individual poems see How Christian is the Edda?.

The order of the book: gods first, then heroes

Gunnell and Pétursson count 29 poetic pieces, ten mythological and nineteen heroic. The figure varies in the literature, depending on how one classes prose passages, fragments and the two poems at the seam.

First comes Völuspá, the prophecy of the seeress, which runs from creation to destruction and a new earth. It is followed by Hávamál, then Vafþrúðnismál and Grímnismál. These first four pieces turn on Odin and on knowledge. Next come Skírnismál, Hárbarðsljóð, Hymiskviða, Lokasenna and Þrymskviða – one poem about Freyr and then four in which Thor appears.

At the seam stand Völundarkviða, the lay of the smith Völund, and Alvíssmál, Thor's question game with a dwarf. The institute explains their position by saying that their characters exist between worlds: Völund is neither a god nor an ordinary hero, and the dwarf Alvíss belongs neither to the Æsir nor to humankind. That is a modern explanation; the compiler gives no reasons.

The heroic part opens with the three Helgi poems (see Helgi and Sigrún). A short prose piece on the death of Sinfjötli leads over to the Sigurd material: Grípisspá, Reginsmál, Fáfnismál, Sigrdrífumál. Then comes the gap, after which the text resumes with the end of a Sigurd poem, followed by the poems about Gudrun, Brynhild and Atli; Guðrúnarhvöt and Hamðismál, on Gudrun's sons and the Gothic king Jörmunrekr, close the book. The full sequence with all texts is in this site's source archive.

Two things stand out in this arrangement. First, the heroic part tells a continuous family history; the groups, as the institute puts it, are linked by kinship and marriage. Second, between the poems, and sometimes inside them, there are prose sentences that fill gaps in the action, introduce characters or lead on to the next piece. The Codex Regius is an edited collection, not a mere songbook. Whether the editor was the scribe of the surviving book or already the maker of its exemplar cannot be decided.

Many of the poems are dialogues: Odin and a giant quiz each other, Thor and a ferryman trade insults, Loki attacks the gods one after another. According to the institute's description, marks in the margins suggest that some of these poems were written down in the way medieval scribes set out texts with distributed roles. Gunnell holds that the poems were meant to be heard and seen in performance. What is attested is the scribal practice; what a recital looked like in the 13th century, or earlier, no source records.

The first stanza of the book

The book begins without a preface. The first stanza of Völuspá is also its entrance, a request for a hearing:

“Hearing I ask from the holy races,
From Heimdall's sons, both high and low;
Thou wilt, Valfather, that well I relate
Old tales I remember of men long ago.”Völuspá, st. 1, translated by Henry Adams Bellows (1923); Old Norse text and Simrock's German version in the source archive

In Guðni Jónsson's edition the Old Norse stanza opens “Hljóðs bið ek allar helgar kindir”. The word hljóð means “hearing, silence”; the speaker is asking an audience to be quiet. The whole poem, stanza by stanza, is in our bilingual reading edition.

The great lacuna

In the middle of the Sigurd section a whole quire is missing. A quire, or gathering, is a bundle of folded sheets from which a book is sewn; here it comprised eight leaves, that is sixteen pages. In 1883 Vigfússon described five complete gatherings and a final one from which three blank leaves had been cut away; between the fourth and the fifth at least one more is missing. Five gatherings of eight leaves plus a final one of five give exactly the 45 surviving leaves; that arithmetic is mine, but it agrees with every statement I have seen.

The text breaks off in Sigrdrífumál, the poem in which Sigurd wakes the sleeping valkyrie and receives runes and counsel from her. According to the source archive, the Codex Regius ends after the first line of stanza 29; the rest of the poem is supplied from later paper manuscripts. The institute explains it thus: the end of Sigrdrífumál probably stood at the front of the lost gathering and survives in copies. Whatever else stood on those eight leaves is lost.

After the gap the manuscript starts in the middle of a Sigurd poem of which only the end is there. It is therefore called Brot af Sigurðarkviðu, “Fragment of a Sigurd lay”. According to Vigfússon, Bishop Brynjólfur marked the place with a Greek note at the head of the gathering meaning roughly “a speech without a beginning”. So the gap already existed in 1643; the institute says the same.

The content of what is lost is known in outline from Völsunga saga. Its author turned the heroic poems into prose while the missing pieces were still known: Sigurd's meeting with Brynhild, the drink of forgetfulness, the wooing on Gunnar's behalf, the ride through the fire, the quarrel of the queens. He quotes a few stanzas verbatim. Our articles on the saga and on Sigurd go into this. The saga replaces the content, not the wording. Whether one long lay or several short ones stood on the eight leaves cannot be determined; estimates of the number of stanzas are calculations from the space available.

1643: a bishop writes his mark in the book

The tangible history of the manuscript begins with Brynjólfur Sveinsson, bishop of Skálholt in the south of Iceland. In 1643 he entered his monogram and the year. From whom he got the book is, according to Pétursson, not known.

The book contains poems from which Snorri Sturluson quotes stanzas in Gylfaginning; the institute names Völuspá, Vafþrúðnismál and Grímnismál. From this the bishop drew a wrong conclusion.

The “Sæmundar Edda” mistake

Among Icelandic scholars of the 17th century the idea circulated that there must have been an older Edda from which Snorri had drawn, written by Sæmundr Sigfússon, called fróði, “the Learned”, a priest of the 11th and early 12th centuries. According to Pétursson, Jón Guðmundsson was the first to mention such an Edda by Sæmundr, in 1623 – twenty years before the Codex Regius surfaced. When Brynjólfur obtained the manuscript of poems, he took it for that supposed work. Hence “Sæmundar Edda” and “Elder Edda”.

Pétursson's verdict is unambiguous: Sæmundr had nothing to do with either compiling or writing it. The manuscript does not name him, it bears no title in the first hand, and by its script it was produced well over a century after his death. “Edda” as a name for the collection of poems is a 17th-century label, and “Elder Edda” reverses the chronological relation of the two books: Snorri's work is older than the manuscript of the poems, even if many of the poems are probably older than Snorri.

1662: a gift for Frederik III

In 1662 Brynjólfur sent the book, together with other valuable vellums, to King Frederik III of Denmark. Iceland was then part of the Danish realm. In Copenhagen the volume entered the Royal Library and later received the number 2365 among the quarto manuscripts. In 1883 Vigfússon describes an 18th-century binding with the arms of Christian VII; the edges of the vellum, he says, were damaged in the binding, and the original cover was probably lost. Its medieval binding is therefore unknown.

The other Codex Regius: GKS 2367 4to

According to the English Wikipedia, a second manuscript reached the king with the same consignment of 1662, and it too is now called “Codex Regius”: GKS 2367 4to. It contains not the poems but Snorri's Prose Edda with Gylfaginning and Skáldskaparmál, dates from the early 14th century and returned to Iceland in 1985. It is likewise kept at the Árni Magnússon Institute. Wikipedia gives its extent as 55 vellum “pages”; whether leaves or pages are meant is not clear from the passage.

Two shelfmarks that differ in a single digit, two books with “Edda” in their modern titles, both “King's Book”: confusion is easy. The number settles it. 2365 is the poems, 2367 is Snorri. A Greek New Testament majuscule of the 8th century is also called Codex Regius (so the German Wikipedia's disambiguation page); it is unrelated.

The sister manuscript AM 748 I a 4to

The closest relative of the Codex Regius is a fragment from the collection of the Icelander Árni Magnússon, hence the abbreviation AM. According to our Edda overview it consists of six vellum leaves from the beginning of the 14th century that have remained in Copenhagen; I did not check these details again against a catalogue for the present article. The fragment contains poems that are also in the Codex Regius, some complete and some in part, among them Grímnismál, Hymiskviða and parts of Hárbarðsljóð.

It permits comparison: where both manuscripts have the same text, the differences show where a scribe misread or altered something. It is also the only medieval manuscript to preserve Baldrs draumar, Odin's ride to a dead seeress. That poem is absent from the Codex Regius. Why the compiler left it out is an open question.

What is not in the King's Book

“Poetic Edda” and “Codex Regius” are not the same thing. Modern editions include a number of poems that belong with the collection in form and subject but are transmitted elsewhere. The institute lists them: Baldrs draumar in the fragment AM 748, Hyndluljóð with the “short Völuspá” in Flateyjarbók, Grottasöngr and Rígsþula in manuscripts of Snorri's Edda, the Battle of the Goths and Huns (Hlöðskviða) inside Hervarar saga. Völuspá itself exists in a second, divergent version in Hauksbók (AM 544 4to), and fragments of a third are embedded in Snorri's quotations.

For most poems in the Codex Regius there is no second witness. A corrupt word there can only be mended by conjecture. Where a second witness does exist, as with Völuspá, the versions differ so much that modern editions produce a composite text. The text read today is therefore, in many places, an editors' decision.

The way back: 21 April 1971

Icelandic requests for the return of manuscripts go back, according to the University of Copenhagen's account, to the 1830s and grew louder after the separation from Denmark in 1944. In May 1965 the Danish parliament decided that items in the Arnamagnæan Collection which could count as “Icelandic cultural property” should be handed over to the Icelandic manuscript institute. This meant works composed or translated by an Icelander and wholly or chiefly concerned with Iceland. The same criteria applied to manuscripts in the Royal Library.

Two books needed a special clause: the Codex Regius of the Poetic Edda and Flateyjarbók. They did not necessarily meet the definition, since poems about gods and heroes and histories of Norwegian kings are not chiefly about Iceland, but their significance for Iceland was so great that they were named individually. Their ceremonial delivery opened the transfer. The Árni Magnússon Institute gives 21 April 1971 as the day of arrival. According to the English Wikipedia the volumes travelled by ship with a military escort, because air travel was not trusted with so precious a cargo. The ship was the Danish warship “Vædderen”: according to a retrospective by the Icelandic government it docked at Miðbakki in Reykjavík harbour on 21 April 1971, where a large crowd was waiting.

The handover took a quarter of a century. By the University of Copenhagen's figures, 1,666 manuscripts from the Arnamagnæan Collection went to Reykjavík, along with all the Icelandic charters, plus 141 manuscripts from the Royal Library. The last two were handed over in June 1997.

Where the book is today

The Codex Regius belongs to the holdings of the Árni Magnússon Institute for Icelandic Studies in Reykjavík. The institute is housed on the campus of the University of Iceland in a building called Edda, Arngrímsgata 5. The exhibition “World in Words” opened there in November 2024, showing medieval manuscripts in the original; reports from its first weeks mention Möðruvallabók and Flateyjarbók besides the Codex Regius. The items are rotated regularly because they are fragile. A report from December 2024 calls the King's Book “quite a small book” and names periods in which it was off display. Anyone travelling for this one manuscript should check with the institute beforehand whether it is currently on show.

What a single manuscript proves, and what it does not

The Codex Regius is a witness of the late 13th century. It shows which poems an Icelandic compiler of that time knew and how he ordered them. Iceland had by then been Christian for more than two and a half centuries.

What the book does not show is how old each poem is, where it was composed, how it sounded before it was written down, or how much the compiler changed. The institute words it cautiously: many of the mythological poems were probably composed in the late pagan period and then reshaped in oral transmission. For the heroic poems it notes that Atli and Jörmunrekr go back to historical figures of the Migration Period (see our article on the Burgundians), but that Helgi and Sigurd cannot be verified and that the poems are not reliable historical sources.

Survival is also a matter of chance: that this book was preserved and its exemplar was not says nothing about their relative worth. Anyone who reads “the faith of the Vikings” straight off the Codex Regius skips several stages: a compiler, at least one exemplar, an oral transmission of unknown length, and the fact that poetry is not the same thing as cult.

What is secure

  • The manuscript GKS 2365 4to has 45 vellum leaves; a gathering of eight leaves is missing and was already missing in 1643.
  • It was written by a single hand not known elsewhere and is dated by its script to the late 13th century (about 1270–1280).
  • It bears no title in the first hand and names no author.
  • Bishop Brynjólfur Sveinsson entered his monogram and the year in 1643 and sent the book to King Frederik III in 1662.
  • The manuscript lay in the Royal Library in Copenhagen until 1971 and came to Iceland on 21 April 1971; it is kept at the Árni Magnússon Institute in Reykjavík.
  • The poems are ordered: mythological poems first, then heroic poems, linked by prose passages.

Myth check

  • “The Codex Regius is the Edda of Sæmundr the Learned.” That was Brynjólfur's assumption in the 17th century. The manuscript does not name Sæmundr and was written long after his death; according to Pétursson he had nothing to do with it.
  • “There is only one Codex Regius.” There are at least two Icelandic ones: GKS 2365 4to with the poems and GKS 2367 4to with Snorri's Edda. Details of date, extent and return are easily swapped between them.
  • “The manuscript dates from the Viking Age.” It was written in the late 13th century in Christian Iceland. The poems may be older; the book is not.
  • “The Codex Regius contains the whole Poetic Edda.” Baldrs draumar, Hyndluljóð, Rígsþula and Grottasöngr are transmitted in other manuscripts.
  • “Völsunga saga fills the gap.” It retells the content of the lost poems in prose and quotes a few stanzas. The wording of the eight leaves is lost.

Frequently asked questions

What is the Codex Regius? An Icelandic vellum manuscript of the late 13th century with the shelfmark GKS 2365 4to. It contains the mythological and heroic poems that form the core of the Poetic Edda and is the only medieval witness for most of them.

Why is the manuscript called the “King's Book”? Because Bishop Brynjólfur Sveinsson sent it to the Danish king Frederik III in 1662, after which it lay in the Royal Library in Copenhagen. The name has nothing to do with the contents.

What is missing from the Codex Regius? A gathering of eight leaves in the Sigurd section. The text breaks off in Sigrdrífumál and resumes with the end of a Sigurd poem. The content of the lost part is preserved in prose in Völsunga saga.

Where can the Codex Regius be seen today? It belongs to the Árni Magnússon Institute in Reykjavík and is shown in the Edda building in the exhibition “World in Words”, though not continuously. It is worth asking the institute before a visit.

What is the difference between GKS 2365 4to and GKS 2367 4to? 2365 is the manuscript of the poems from the late 13th century; 2367 is a manuscript of Snorri's Prose Edda from the early 14th century. Both lay in the Royal Library in Copenhagen, which is why both are called Codex Regius.

Further reading on this site

tealight collection Tordensønn with Thor, Odin, Freya, Loki and Heimdall
From our workshop: tealight collection “Tordensønn” with Thor, Odin, Freya, Loki and Heimdall – laser-engraved on crystal glass. To the shop →

Read more with us

The Edda · Völuspá – the Seeress's Prophecy · Grímnismál: Odin between the fires · Gylfaginning: the Tricking of Gylfi · The Völsunga saga.

Sources & literature

Visit the Shop More posts

More from History & Finds

All History & Finds articles →

The Corieltauvi: a British tribe known mainly through its finds

No Roman historian tells the story of the Corieltauvi. None of their rulers appears in Tacitus, and no battle bears …

Read more →

The Cotini and Osi – When Tacitus' "Germany" Gets Too Small

In the hill country north of the Danube, behind the Marcomanni and Quadi, Tacitus names two peoples he expressly …

Read more →

The Cugerni – a Germanic Tribe on the Lower Rhine in Rome's Shadow

The Cugerni lived around Xanten, Kleve and Krefeld on the left bank of the Lower Rhine. They were Germani, but Germani …

Read more →

The Curmsun Disc: a piece of gold bearing Harald's name and a find story nobody can check

It measures 4.5 centimetres across, weighs a little over 25 grams and names, in Latin letters, a King Harald, son of …

Read more →