Sveinnson, Bishop of Skálaholt. The text which Gray found in Bartolinus, however, was sufficiently true to enable him to make a better translation of the Vegtamskvida than any which has been attempted since, and to make us deeply regret that he did not 'imitate' more of these noble Eddaic chants. He even attempts a philological ingenuity, for, finding that Odin, to conceal his true nature from the Völva, calls himself Vegtam, Gray translates this strange word 'traveller,' evidently tracing it to veg, a way. He omits the first stave, which recounts how the Esir sat in council to deliberate on the dreams of Balder, and he also omits four spurious stanzas, in this showing a critical tact little short of miraculous, considering the condition of scholarship at that time.”

This evidence looks convincing, but unfortunately it has no basis of fact. The number of mistakes in the few sentences just quoted is surprising. What is meant by "the tolerably complete but very inaccurate edition of Sæmundar Edda which existed in [Gray's] time," it is hard to divine. There was no edition at all of the so-called Sæmund's Edda (i.e., The Poetic Edda) extant in Gray's time, for none had been published. Gray died in 1771, and the first edition of this Edda (the great Copenhagen edition) began to be published in 1787, and was not completed till 1828. Meantime Rask and Afzelius had published a complete edition in one small volume in 1818. Of the more than thirty pieces which compose the Poetic Edda only three (Voluspá, Hávamál,1 and Vegtamskviða) had been printed before Gray's death. The only edition of the

1 The Voluspá and the Hávamál had been edited by Resenius (1665) in a volume which was usually bound up with his edition of the Prose Edda. The poems were accompanied by a Latin translation by Stefán Ólafsson. The stanzas of the Hávamál concerning runes were printed by Resenius as a separate poem, and this division accounts for Mallet's words "le petit poème intitulé le chapitre runique, ou la magie d'Odin" (Introd. à l'Hist. de Dannemarc, 2d ed., II, 285; Northern Antiquities, 1770, II, 216). When Mallet first published his second volume (1756), he did not even know whether any part of the Poetic Edda was in existence except Voluspá, Hávamál and this so-called Runic Chapter, and this uncertainty was shared by the English translator in 1770 (Northern Antiquities, II, 201); but before he published his third edition (1787), Mallet had learned better (II, 264 ff.). The Hávamál was re-edited by Göransson in 1750. A pretty complete list of the contents of the Poetic Edda is given in Peringskjöld's catalogue of Copenhagen MSS. appended to Wanley's Catalogus (1705), p. 310. Wanley's Catalogus (which Gray knew) forms, with Hickes's Thesaurus, the Antiqua Literatura Septentrionalis.

Vegtamskviða (the original of The Descent of Odin) was that in Bartholin's book.1 We are forced to conjecture that Mr. Gosse has for the moment confused the Poetic Edda with the Prose Edda of Snorri, which was edited by Resenius in 1665, and by Göransson in 1746.2

All this, to be sure, though it may shake one's confidence in Mr. Gosse's accuracy, does not affect the validity of his arguments from Gray's "philological ingenuity" in translating Vegtamr by Traveller and from the omission of the four spurious stanzas. But these arguments themselves rest on other misapprehensions. As to the four spurious stanzas, they are omitted by Gray simply because he? was unaware of their existence. They do not stand in the ArnaMagnæan MS. or in Bartholin's text, and were not printed at all until 1787, when the Copenhagen editors inserted them in the text from late paper manuscripts. The rendering of Vegtamr by Traveller would doubtless be significant but for a fact which Mr. Gosse neglects to mention, though it is of the first importance in settling the main question: Bartholin appended to each stanza of the original a literal translation into Latin, and that translation, which renders Vegtamr by Viator, relieved Gray of the necessity of "attempting a philological ingenuity" in interpreting this strange word. The fact that Bartholin's texts of the two poems which Gray translated are, like all his other Norse quotations, accompanied by

1 The Vegtamskviða is not contained in the manuscript discovered by Bishop Brynjólfr Sveinsson, probably in 1643,- the so-called Codex Regius. It is preserved (except for some paper copies of no consequence) only in the Arna-Magnæan MS. 748, which contains, to be sure, merely a somewhat different redaction of what is essentially the same collection. The poem was doubtless made known to Bartholin by Arni Magnússon himself.

2 Göransson's Latin version of the first tract in the Prose Edda (The Deception of Gylfi) was reprinted in Northern Antiquities, 1770, II, 273-352. 3 See Bugge, Norræn Fornkvæði, pp. 138-140. Mason reprinted Bartholin's Latin translation in his note at the end of The Descent of Odin. Mitford's remark that "the first five stanzas of this Ode are omitted" by Gray is repeated by Dr. Bradshaw without investigation: the fact is that Gray omitted but one stanza (the first, also omitted by Bartholin) of the genuine text, the other four are those four spurious strophes that misled Mr. Gosse. Ten other spurious lines, which are in Bartholin, are accepted by Gray. 4 The Fatal Sisters is extracted by Gray from Torfæus, but Torfæus refers to Bartholin, from whom he repeats both the original and the Latin version. Gray too adds a reference to Bartholin. In his preface Bartholin acknowledges his indebtedness to the famous Icelander Arni Magnússon with

literal Latin renderings, has, as has just been suggested, an important bearing on the general question of Gray's knowledge of the Old Norse tongue. By the use of these literal renderings, Gray could have written both The Fatal Sisters and The Descent of Odin without reading a word in the original language. What we know of his scholarly habits and of his insatiate love of investigation makes it incredible, however, that he should have contented himself with so humble a process. It is more probable that he carefully compared the Latin text with the Old Norse and by this means made out something of the originals. In this he would be assisted by the striking similarity of many of the Scandinavian words to their English cognates. Since he is likely to have done this with many other interesting texts in Bartholin and elsewhere, he may perhaps have arrived at a halting knowledge of the language: that he ever "mastered" it, there is at any rate no evidence in his two "Norse odes."

The notes and introductions which Gray wrote for these Norse odes seem at first sight to supply some of the evidence which is lacking in the odes themselves. But an examination of this apparatus reduces this testimony also to a minimum. Of all the comments which Gray wrote on The Descent of Odin, including the long note on the seeresses,1 the material is furnished in Latin by Bartholin and some of it was also accessible to all Europe in Mallet. The Preface to The Fatal Sisters is from Torfæus 2 (Orcades, i, 10, pp. 33 ff.) and the other notes are, as before, due to Bartholin. Of only two bits of information is this not true: the translation of

regard to the Latin versions of the Norse poems that he quotes (“ex versione, quæ accurate docti Islandi Arnæ Magnæi industriæ complementum suum debet ").

1 Not printed by Gray, but added by Mason from Gray's papers (see p. 169, below).

2 Gray's date for the Battle of Clontarf,- "Christmas day" (Preface to The Fatal Sisters), is a mere slip. Torfæus (Orcades, i, 10, p. 35) clearly puts the battle on Good Friday: "Die Veneris, qvi, in diem memoriæ passionis Servatoris, Zwrŋpía dictum," etc. Gray's note on the conversion of the Orcadians is also derived in the main from Torfæus, who (Orcades, i, 10, p. 33) gives an account of the heroic measures adopted in 995 by King Olave Tryggvason to secure the baptism of Earl Sigurd. All the knowledge which he shows of the "history of Olaus Tryggueson" in the same note, he could easily have collected from the Orcades and from Torfæus's Historia Rerum Norvegicarum (1711 ff., see vi, 7 ff., II, 246 ff.), supplemented perhaps by Bartholin or by Peringskjöld's Latin version of the Heimskringla (1697).

fiolkunnug (read fjęlkunnig) by multi-scia and of vísinda-kona by oraculorum mulier in the long note inserted by Mason from Gray's papers (see Descent of Odin, v. 51, p. 170, below) is not based on anything in Bartholin and seems to betray a knowledge of the component parts of those words. But we cannot build much on this. Fjol is a very common prefix and Gray may have found it explained in various places. In the Glossary to Verelius's edition of the Hervarar Saga (Upsala, 1672), for example, all-fiolkunnugur (put under f) is glossed by multiscius. In the same Glossary

vísinda menn is explained by oraculorum interpretes: this would furnish Gray with the meaning of vísinda-, and kona is a very common word, the meaning of which he would infallibly discover by the process of comparison suggested above. We have no real evidence that Gray knew this work of Verelius, but the presumption is perhaps that way. It was certainly known to Warton1 and Percy,2 and there was a copy in the Bodleian Library.3

Important direct evidence of Gray's interest in Scandinavian study is contained in his well-known letter to Mason, Jan. 13, 1758 (Works, II, 350 ff.). In this he insists on the distinction between Celtic and Northern antiquities and expresses himself as follows about Odin : "Woden himself is supposed not to have been older than Julius Cæsar; but let him, have lived when he pleases, it is certain that neither he nor his Valhalla were heard of till many ages after. This is the doctrine of the Scalds, not of the Bards; these are the songs of Hengist and Horsa, a modern new-fangled belief in comparison with that which you ought to possess." 4 "In short," he remarks in

1 History of English Poetry, ed. Hazlitt, I, 120. 2 See Five Pieces of Runic Poetry, 1763, p. 4.

3 Catalogus Librorum Septentrionalium appended to Jónsson (Runolphus Jonas), Grammatica Islandica Rudimenta, Oxford, 1688, p. 179. Of course, as Mr. Gosse suggests, Gray may have used the Index Linguæ veteris ScythoScandica sive Gothica of Verelius (Upsala, 1691).

4 I.e., the Druidical belief, which Mason needed for his Caractacus. play, Mason makes the chorus of Druids present a sword to the hero:

"Caractacus!

Behold this sword: The sword of old Belinus,

In his

Stain'd with the blood of giants, and its name Trifingus."
(Poems, 1764, p. 257.)

In a note (p. 316) he says: "TRIFINGUS. The name of the inchanted sword in the Hervarer Saga." The reader will recognize the famous Tyrfing.

the same letter (II, 352), "I am pleased with the Gothic Elysium. Do you think I am ignorant about either that, or the hell before, or the twilight? I have been there, and have seen it all in Mallet's Introduction to the History of Denmark (it is in French), and many other places." The first volume of Mallet's Introduction had appeared in 1755, the second (containing a translation of much of the Prose Edda) in 1756. If Gray's attention, like that of Europe in general, was first called to the Norse mythology by this work, it is interesting to observe that he had so far extended his reading by the beginning of 1758 as to have seen the chief doctrines of the Odinic system "in many other places." The letter gives us no inkling as to what these places were, but it does inform us what one of them was not: Gray expressly disclaims having read " Keysler," i.e., the Antiquitates Selectae Septentrionales et Celticae of Johann Georg Keysler, Hannover, 1720.1

In 1761 Gray translated The Fatal Sisters and The Descent of Odin, and these show that he had been reading the treatise of Thomas Bartholin De Causis contemnenda Mortis (Copenhagen, 1689) with much attention and that he had at any rate consulted the Orcades of Torfæus (1697). If, as there seems little reason for doubting, the Observations on the Pseudo-Rhythmus (Works, I, 361 ff.), intended like the Norse Odes as material for the projected History of English Poetry, were written at about the same time,3 we may

2

1 Mr. Gosse's suggestion (Gray's Works, II, 351) that the "Keysler" mentioned by Gray "was probably the second English edition, of 1757, of Johann Georg Keysler's Travels through Germany, Hungary, Bohemia, Switzerland, Italy, and Lorrain" was made in momentary forgetfulness of Mason's letter of Jan. 5, 1758 (Correspondence of Gray and Mason, ed. Mitford, p. 120), to which this of Gray's was a reply: Mason mentions the title of the book, and in his reply to Gray (Jan. 16, 1758; id., p. 130), he returns to the subject and gives an extract from Keysler. The work was in its day well-known among antiquaries and is still worth consulting.

2 Torfæus (pormóðr Torfason), b. 1636, d. 1719, was a learned Icelander and one of the founders of the science of Northern Antiquities. His most important works (chiefly historical) were written while he was the King of Denmark's historiographer royal for Norway. His History of the Orkneys (Historia Orcadum, Copenhagen, 1867), is the work here referred to.

3 The essay contains a note as to Death and Life and Scottish Field, two poems in the Percy MS.: "I read them in a MS. collection belonging to the Rev. Mr. Thomas Piercy in 1761" (Works, I, 371); but this note proves little. In a letter to Montagu, May 5, 1761 (Letters, ed. Cunningham, III, 399), Walpole writes: "Gray and Mason were with me, and we listened to