splendid tales of the fathers," Snorri, by implication, says to the youthful bard; "but remember always that these old legends are to be used to point a moral or adorn a tale, and not to be believed, or to be altered without authority of ancient skalds who knew them. Belief is sin; tampering with tradition is a crime against scholarship."

The second and third sections, Skáldskaparmál and Háttatal, offer the rules of composition, and drive them home by means of models drawn, in the one case, from acknowledged masters of the craft, in the other, by the example of a complete skaldic trilogy, the work of a man who was accepted by his own time as a worthy successor of Bragi, Kormákr, and Einarr. A needed transition from the literary to the technical portion of the book is supplied by Bragarædur, which narrates, in the same spirit as Gylfaginning, further useful tales, and concludes with a mythological account of the skaldic art.

Even the Prologue, which many scholars consider spurious, is an integral part of the work-a fact established by Snorri's single address, in the character of the author, to beginners. In this apostrophe he refers to the Prologue: "Remember, these tales are to be used only as Chief Skalds have used them, and must be revered as ancient tradition, but are neither to be believed nor to be tampered with. Regard them as I have indicated at the beginning of this book." The beginning of the book is a summary of the Biblical story of the Creation and Deluge, followed by a rationalized account of the rise of the ancient pagan faith, according to which the old gods appear, not as deities, but

as men.

The word "Edda," as applied to the whole work, has long furnished scholars with material for disputation. The

different theories regarding it need not be re-stated here. It is the translator's personal opinion that Magnússon's etymology, if not established, is at least the most satisfactory one likely to be offered. Magnússon' points out that Snorri passed the interval between his third and nineteenth years at Oddi, under the fostering of the grandson of Sæmundr the Learned; that Sæmundr, who had studied at Paris, had founded a school at Oddi; that Snorri became the author of a book which was called Edda; and that this book contains, in its first section, a prose paraphrase of many of the songs from the Elder or Poetic Edda, together with a number of quotations from that work. Now the Poetic Edda was ascribed by its earliest recorded possessor, Bishop Brynjólf Sveinsson, to Sæmundr; and while it is improbable that Sæmundr composed the poem, it is highly probable that it once formed part of his library at Oddi. There Snorri may have learned to know it; and we may assume that he gave the prose edition the name of its poetical original. That original, "the mother MS.," he thinks would naturally have been called "the book of, or at Oddi," which would be expressed, in Icelandic, either as "Oddabók," or as "Edda," following, in the latter case, accepted linguistic laws.

Snorri's familiarity with the Elder or Poetic Edda is demonstrated by his frequent quotations from Völuspá, Hávamál, Grímnismál, Vafthrúdnismál, Alsvinnsmál or Alvissmál, and Grottasöngr. He knew Lokasenna as well, but confused three stanzas, apparently failing to remember the order

' Magnússon's theory, with a summary of all others in the field, was presented in a paper read before the Viking Club on November 15, 1895, published in the Saga Book of that society, and separately printed at London in 1896.

in his original. One poem that he mentions is lacking in the Poetic Edda as we know it: Heimdallargaldr, the Song or Incantation of Heimdallr; moreover, he makes seventeen citations from other poems which, although lost to us, evidently formed portions of the original Eddic collections, or belonged to the same traditional stock. The disappearance of the manuscript which Snorri used is a great loss.

The first translation of the Prose Edda was published at Copenhagen in 1665, when the complete text appeared, with Latin and Danish interpretation. This was entitled Edda islandorum an. Chr. 1215 islandice conscripta per Snorronem Sturla, nunc primum islandice, danice, et latine ex antiquis codicibus in lucem prodit opera P. J. Resenii. The standard Danish translation is that of R. Nyerup, Copenhagen, 1865. In 1746, J. Göransson printed at Upsala the first Swedish version, with a Latin translation. Göransson's original was the Codex Upsaliensis. Anders Uppström made an independent translation in 1859.

In 1755-56 there appeared at Copenhagen a work of the greatest importance for the study of Scandinavian antiquities in England: Mallet's Monumens de la Mythologie et de la Poesie des Celtes et Particulièrement des Anciens Scandinaves. This book, which comprised a general introduction on the ancient Scandinavian civilization, a translation of Gylfaginning, and a synopsis of Skáldskaparmál and Háttatal, was turned into English by Bishop Percy, under the title of Northern Antiquities. Percy claimed to know Göransson's text as well as the French. Northern Antiquities was published at London in 1770, and was reprinted at Edinburgh in 1809, with additions by Sir Walter Scott.

The best-known translation, and the only complete one which is at all trustworthy, is that in Latin, combined, with

the Icelandic text, in the Arnamagnæan edition, Copenhagen, 1848-87.

In 1842, G. W. Dasent, the translator of Njáls Saga, and a prominent scholar in the Scandinavian field, printed at Stockholm his Prose or Younger Edda, which contains a translation of Gylfaginning and of the narrative passages of Skáldskaparmál. A similarly incomplete English version was printed at Chicago, in 1880, by Rasmus B. Anderson. Professor Anderson also edited a combined translation of both Eddas, the Poetic Edda by Benjamin Thorpe, and the Prose Edda by I. A. Blackwell. Blackwell's translation, which stops with Bragarædur, had first appeared at London in 1847, together with an abstract of Eyrbyggja Saga by Scott. Samuel Laing's translation is likewise incomplete.

A French version of Gylfaginning, La Fascination de Gulfi, was published at Strassburg by F. G. Bergmann. A second edition appeared in 1871.

So far as I can ascertain, the first translation into German was the work of Friedrich Rühs, Berlin, 1812. This contains a long historical introduction, and ends with the story of the Völsungs in Skáldskaparmál. Karl Simrock's Die Jüngere Edda, published in 1851 and reprinted in 1855, although incomplete, is more accurate than any earlier translation, and is remarkable for its literary excellence. The most scholarly rendering into German is by Hugo Gering, Leipzig, 1892, but unfortunately it includes only the narrative portions of the book.

Until 1900, the best edition of Snorri's Edda was by Thórleifr Jónsson, Copenhagen, 1875. This was superseded by Finnur Jónsson's splendid Danish edition. In 1907, Professor Jónsson produced an Icelandic edition,

which forms volume xli of the Íslendinga Sögur, published at Reykjavík.

It was fortunate for me that these last two editions appeared before I began my work. Professor Jónsson provided me with an excellent text; and, second in value only to this, with an index and an invaluable Icelandic prose re-phrasing of the skaldic verses.

I regret exceedingly that the highly technical nature of Háttatal forbids translation into English. There are, to be sure, more or less-usually less-accurate translations into Scandinavian and into Latin. Even in the excellent Arnamagnæan edition, many of the glosses are purely conjectural; and any attempt to convey into English a vocabulary which has no equivalent in our language must fail. Skáldskaparmál, however, is here presented, complete, for the first time in English.

To those who have helped me I wish to express my deepest appreciation. First of all, to Professor William Henry Schofield I owe a debt of gratitude which is more than four years old, and has increased beyond computation. Dr. Henry Goddard Leach, my first instructor in Scandinavian literature, gave me my greatest single intellectual stimulus, and thereby determined the current of my work. Dr. Frederick W. Lieder, of Harvard University, deserves my thanks for his devoted assistance in reading proof, a task as dreary as it is essential. I am also indebted for valuable suggestions to Mr. H. W. Rabe, of Simmons College.

It is a great satisfaction to acknowledge these debts, incurred in the course of a labor which has been my delight for several years. I should, however, do injustice to those who have aided me, as well as to myself, if I did