[546] This follows from the allusive way in which he and his wife are introduced—there must be a background to allusions. If the poet were inventing this figure, and had no background of knowledge in his audience to appeal to, he must have been more explicit. Cf. Olsen in Christiania Videnskapsselskapets Skrifter, 1914, II, 2, 107.

[547] p. 87.

[548] See Olrik, "Nordisk og Lappisk Gudsdyrkelse," Danske Studier, 1905, pp. 39-57; "Tordenguden og hans dreng," 1905, pp. 129-46; "Tordenguden og hans dreng i Lappernes myteverden," 1906, pp. 65-9; Krohn, "Lappische beiträge zur germ. mythologie," Finnisch-Ugrische Forschungen, VI, 1906, pp. 155-80.

[549] See Axel Olrik in Festgabe f. Vilh. Thomsen, 1912 (= Finnisch-Ugrische Forschungen, XII, 1, p. 40). Olrik refers therein to his earlier paper on the subject in Danske Studier, 1911, p. 38, and to a forthcoming article in the Germanisch-Romanische Monatsschrift, which has, I think, never appeared. See also K. Krohn in Göttingische gelehrte Anzeigen, 1912, p. 211. Reviewing Meyer's Altgermanische Religionsgeschichte, Krohn, after referring to the Teutonic gods of agriculture, continues "Ausser diesen agrikulturellen Gottheiten sind aus der finnischen Mythologie mit Hülfe der Linguistik mehrere germanische Naturgötter welche verschiedene Nutzpflanzen vertreten, entdeckt worden: der Roggengott Runkoteivas oder Rukotivo, der Gerstengott Pekko (nach Magnus Olsen aus urnord. Beggw-, vgl. Byggwir) und ein Gott des Futtergrases Sämpsä (vgl. Semse od. Simse, 'die Binse')." See also Krohn, "Germanische Elemente in der finnischen Volksdichtung," Z.f.d.A. LI, 1909, pp. 13-22; and Karsten, "Einige Zeugnisse zur altnordischen Götterverehrung in Finland," Finnisch-Ugrische Forschungen, XII, 307-16.

[550] As proposed by K. Krohn in a publication of the Finnish Academy at Helsingfors which I have not been able to consult, but as to which see Setälä in Finnisch-Ugrische Forschungen, XIII, 311, 424. Setälä accepts the derivation from beggwu-, rejecting an alternative derivation of Pekko from a Finnish root.

[551] This is proposed by J. J. Mikkola in a note appended to the article by K. Krohn, "Sampsa Pellervoinen < Njordr, Freyr?" in Finnisch-Ugrische Forschungen, IV, 231-48. See also Olrik, "Forårsmyten hos Finnerne," in Danske Studier, 1907, pp. 62-4.

[552] See note by K. Krohn, Finnisch-Ugrische Forschungen, VI, 105.

[553] See above, p. 87, and M. J. Eisen, "Ueber den Pekokultus bei den Setukesen," Finnisch-Ugrische Forschungen, VI, 104-11.

[554] See M. Olsen, Hedenske Kultminder i Norske Stedsnavne, Christiania Videnskapsselskapets Skrifter, II, 2, 1914, pp. 227-8.

[555] See above, p. 84.

[556] Mannhardt, Mythologische Forschungen, 332.

[557] In view of the weight laid upon this custom by Olrik as illustrating the story of Sceaf, it is necessary to note that it seems to be confined to parts of England bordering on the "Celtic fringe." See above, pp. 81, etc. Olrik and Olsen quote it as Kentish (see Heltedigtning, II, 252) but this is certainly wrong. Frazer attributes the custom of "crying the mare" to Hertfordshire and Shropshire (Spirits of the Corn, I, 292 = Golden Bough, 3rd edit., VII, 292). In this he is following Brand's Popular Antiquities (1813, I, 443; 1849, II, 24; also Carew Hazlitt, 1905, I, 157). But Brand's authority is Blount's Glossographia, 1674, and Blount says Herefordshire.

[558] Brand, Popular Antiquities, 1849, II, 24.

[559] Frazer in the Folk-Lore Journal, VII, 1889, pp. 50, 51; Adonis, Attis and Osiris, I, 237.

[560] Frazer, Adonis, Attis and Osiris, I, 238 (Golden Bough, 3rd edit.).

[561] Frazer, Spirits of the Corn and of the Wild, I, 143-4.

[562] Frazer in the Folk-Lore Journal, VII, 1889, pp. 50, 51.

[563] Mannhardt, Forschungen, 317.

[564] Frazer, Spirits of the Corn, I, 138.

[565] Mannhardt, 323; Fraser, Adonis, I, 238.

[566] Mannhardt, 330.

[567] Mannhardt, 24; Frazer, Adonis, I, 238.

[568] Frazer, Adonis, I, 237.

[569] Frazer, Spirits of the Corn, I, 217.

[570] See Björkman in Anglia, Beiblatt, XXX, 1919, p. 23. In a similar way Sceaf appears twice in William of Malmesbury, once as Sceaf and once as Strephius.

[571] Vol. LII, p. 145.

[572] MS Cott. Vesp. B. XXIV, fol. 32 (Evesham Cartulary). See Birch, Cart. Sax. I, 176 (No. 120); Kemble, Cod. Dipl. III, 376. Kemble prints þæt æft for þā æft (MS "þ¯ æft"). For examples of "þ¯" for þā, see Ælfrics Grammatik, herausg. Zupitza, 1880; 38, 3; 121, 4; 291, 1.

[573] There are two copies, one of the tenth and one of the eleventh century, among the Crawford Collection in the Bodleian. See Birch, Cart. Sax. III, ..7 (No. 1331); Napier and Stevenson, The Crawford Collection (Anecdota Oxoniensia), 1895, pp. 1, 3, 50.

[574] MS Cotton Ch. VIII, 16. See Birch, Cart. Sax. II, 363 (No. 677); Kemble, Cod. Dipl. II, 172.

[575] A nearly contemporary copy: Westminster Abbey Charters, III. See Birch, Cart. Sax. III, 189 (No. 994), and W. B. Sanders, Ord. Surv. Facs. II, plate III.

[576] A fourteenth to fifteenth century copy preserved at Wells Cathedral (Registr. Album, f. 289 b). See Birch, Cart. Sax. III, 223 (No. 1023).

[577] MS Cotton Aug. II, 6. See Birch, Cart. Sax. III, 588 (No. 1282).

[578] Brit. Mus. Stowe Chart. No. 32. See Birch, Cart. Sax. III, 605 (No. 1290).

[579] Cf. the Victoria History, Middlesex, II, p. 1.

[580] "Grendeles gate har väl snarast varit någon naturbildning t. ex. ett trångt bergpass eller kanske en grotta": C. W. von Sydow, in an excellent article on Grendel i anglosaxiska ortnamn, in Nordiska Ortnamn: Hyllningsskrift tillägnad A. Noreen, Upsala, 1914, pp. 160-4.

[581] Près du Neckersgat molen, il y avait jadis, antérieurement aux guerres de religion, des maisons entourées d'eau et appelées de hoffstede te Neckersgate: Wauters (A.), Histoire des Environs de Bruxelles, 1852, III, 646.

[582] Peg Powler lived in the Tees, and devoured children who played on the banks, especially on Sundays: Peg o' Nell, in the Ribble, demanded a life every seven years. See Henderson (W.), Notes on the Folk-Lore of the Northern Counties of England, 1879 (Folk-Lore Society), p. 265.

[583] See Kisch (G.), Vergleichendes Wörterbuch der siebenbürgischen und moselfränkischluxemburgischen Mundart, nebst siebenbürgischniederrheinischem Ortsund Familiennamen-verzeichnis (vol. XXXIII, 1 of the Archiv des Vereins f. siebenbürg. Landeskunde, 1905).

[584] See Grindel in Förstemann (E.), Altdeutsches Namenbuch, Dritte Aufl., herausg. Jellinghaus, II, 1913, and in Fischer (H.), Schwäbisches Wörterbuch, III, 1911 (nevertheless Rooth legitimately calls attention to the names recorded by Fischer in which Grindel is connected with bach, teich and moos).

[585] There is an account of this by G. Kisch in the Festgabe zur Feier der Einweihung des neuen evang. Gymnasial Bürger- und Elementar-schulgebäudes in Besztercze (Bistritz) am 7 Oct. 1911; a document which I have not been able to procure.

[586] Such a connection is attempted by W. Benary in Herrig's Archiv, CXXX, 154. Alternative suggestions, which would exclude any connection with the Grendel of Beowulf, are made by Klaeber, in Archiv, CXXXI, 427.

[587] A very useful summary of the different etymologies proposed is made by Rooth in Anglia, Beiblatt, XXVIII (1917), 335-8.

[588] So Skeat, "On the significance of the monster Grendel," Journal of Philology, Cambridge, XV (1886), p. 123; Laistner, Rätsel der Sphinx, 1889, p. 23; Holthausen, in his edition.

[589] So Weinhold in the SB. der k. Akad. Wien, Phil.-Hist. Classe, XXVI, 255.

[590] Cf. Gollancz, Patience, 1913, Glossary. For grindill as one of the synonyms for "storm," see Edda Snorra Sturlusonar, Hafniae, 1852, II, 486, 569.

[591] This will be found in several of the vocabularies of Low German dialects published by the Verein für Niederdeutsche Sprachforschung.

[592] See grand in Falk and Torp, Etymologisk Ordbog, Kristiania, 1903-6.

[593] See Feist, Etymol. Wörterbuch der Gotischen Sprache, Halle, 1909; grunduwaddjus.

[594] With Grendel, thus explained, Rooth would connect the "Earth man" of the fairy-tale "Dat Erdmänneken" (see below, p. 370) and the name Sandhaug, Sandey, which clings to the Scandinavian Grettir- and Orm-stories. We have seen that a sandhaug figures also in one of the Scandinavian cognates of the folk-tale (see above, p. 67). These resemblances may be noted, though it would be perilous to draw deductions from them.

[595] Schweizerisches Idiotikon, II, 1885, p. 776.

[596] See above, pp. 43, etc.; below, p. 311.

[597] Duignan, Warwickshire Place Names, p. 22. Duignan suggests the same etymology for Beoshelle, beos being "the Norman scribe's idea of the gen. plu." This, however, is very doubtful.

[598] Engl. Stud. LII, 177.

[599] Heltedigtning, II, 255. See above, pp. 81-7.

[600] Binz in P.B.B. XX, 148; Chadwick, Origin, 282. So Clarke, Sidelights, 128. Cf. Heusler in A.f.d. A. XXX, 31.

[601] A.-S. Chronicle.

[602] Historia Brittonum.

[603] "hrædlan" (gen.), Beowulf, 454.

[604] "hrædles," Beowulf, 1485.

[605] A.-S. Chronicle.

[606] Beowulf, Ethelwerd.

[607] Geata, Geta, Historia Brittonum; Asser; MS Cott. Tib. A. VI; Textus Roffensis.

[608] A.-S. Chronicle.

[609] Charter of 931.

[610] A.-S. Chronicle, Ethelwerd.

[611] Origin, 273.

[612] Origin, 282.

[613] Some O.H.G. parallels will be found in Z.f.d.A. XII, 260. The weak form Gēata, Mr Stevenson argues, is due to Asser's attempt to reconcile the form Gēat with the Latin Geta with which he identifies it (Asser, pp. 160-161). See also Chadwick, Heroic Age, 124 footnote. Yet we get Gēata in one text of the Chronicle, and in other documents.

[614] This is the view taken by Plummer, who does not seem to regard any solution as possible other than that the names are missing from the Parker MS by a transcriber's slip (see Two Saxon Chronicles Parallel, II, p. xciv).

[615] Plummer, II, pp. xxix, xxxi, lxxxix.

[616] Plummer, II, p. lxxi. Note Beowi for Bedwig.

[617] This table shows the relationship of the genealogies only, not of the whole MSS, of which the genealogies form but a small part. MS-relationships are always liable to fluctuation, as we pass from one part of a MS to another, and for obvious reasons this is peculiarly the case with the Chronicle MSS.

[618] Origin, 295.

[619] Origin, 292.

[620] Origin, 296.

[621] The absence of the West-Saxon pedigree may be due to the document from which the Historia Brittonum and the Vespasian MS derive these pedigrees having been drawn up in the North: Wessex may have been outside the purview of its compiler; though against this is the fact that it contains the Kentish pedigree. But another quite possible explanation is, that Cerdic, with his odd name, was not of the right royal race, but an adventurer, and that it was only later that a pedigree was made up for his descendants, on the analogy of those possessed by the more blue-blooded monarchs of Mercia and Northumbria.

[622] See M.L.N. 1897, XII, 110-11.

[623] It is prefixed to the Parker MS of the Chronicle, and is found also in the Cambridge MS of the Anglo-Saxon Bede (Univ. Lib. Kk. 3. 18) printed in Miller's edition; in MS Cott. Tib. A. III, 178 (printed in Thorpe's Chronicle): and in MS Add. 34652, printed by Napier in M.L.N. 1897, XII, 106 etc. There are uncollated copies in MS C.C.C.C. 383, fol. 107, and according to Liebermann (Herrig's Archiv, CIV, 23) in the Textus Roffensis, fol. 7 b. There is also a fragment, which does not however include the portion under consideration, in MS Add. 23211 (Brit. Mus.) printed in Sweet's Oldest English Texts, p. 179. The statement, sometimes made, that there is a copy in MS C.C.C.C. 41, rests on an error of Whelock, who was really referring to the Parker MS of the Chronicle (C.C.C.C. 173).

[624] p. 73.

[625] See above, p. 70.

[626] Brandl in Herrig's Archiv, CXXXVII, 12-13.

[627] Origin, p. 272.

[628] So Ethelwerd (Lib. I) sees in Woden a rex multitudinis Barbarorum, in error deified. It is the usual point of view, and persists down to Carlyle (Heroes).

[629] Origin, p. 293.

[630] Beowulf, p. 5. For a further examination of this "Beowa-myth" see Appendix A, above.

[631] Cf. Tupper in Pub. Mod. Lang. Assoc. Amer. XXVI, 275.

[632] P.B.B. XLII, 347-410. A theory as to the date of Beowulf, in some respects similar, was put forward by Mone in 1836: Untersuchungen zur Geschichte der teutschen Heldensage, p. 132.

[633] See above, p. 103; and Brandl in Pauls Grdr. (2) II, 1000, where the argument is excellently stated.

[634] See Olrik, Sakses Oldhistorie, 1894, 190-91.

[635] See Björkman, Eigennamen im Beowulf, 77.

[636] Sarrazin's attempt to prove such corruption is an entire failure. Cf. Brandl in Herrig's Archiv, CXXVI, 234; Björkman, Eigennamen im Beowulf, 58 (Heaðo-Beardan).

[637] A few Geatic adventurers may have taken part in the Anglo-Saxon invasion, as has been argued by Moorman (Essays and Studies, V). This is likely enough on a priori grounds, though many of the etymologies of place-names quoted by Moorman in support of his thesis are open to doubt.

[638] P.B.B. XLII, 366-7.

[639] History of England to the Norman Conquest, I, 245.

[640] Heroic Age, 52-6. I have tried to show (Appendix F) that these accounts of cremation are not so archaeologically correct as has sometimes been claimed.

[641] Oman, England before the Norman Conquest, 319.

[642] Bede, Hist. Eccles. IV, 26.

[643] "Nunc qui Roma veniunt idem allegant, ut qui Haugustaldensem fabricam vident ambitionem Romanam se imaginari jurent." William of Malmesbury, Gesta Pontificum, Rolls Series, p. 255.

[644] Baldwin Brown, The Arts in Early England, II, 1903, p. 325.

[645] p. 407.

[646] Beowulf, ll. 201, 601-3.

[647] Cf. Beowulf, l. 1018.

[648] Bede, Eccles. Hist. III, 21.

[649] See Oman, pp. 460, 591, for the honour done to this saint by converted Danes.

[650] p. 393.

[651] Æneid, X, 467-9.

[652] In the two admirable articles by Klaeber (Archiv, CCXVI, 40 etc., 399 etc.) every possible parallel is drawn: the result, to my mind, is not complete conviction.

[653] Chadwick, Heroic Age, 74.

[654] "Litteris itaque ad plenum instructus, nativae quoque linguae non negligebat carmina; adeo ut, teste libro Elfredi, de quo superius dixi, nulla umquam aetate par ei fuerit quisquam. Poesim Anglicam posse facere, cantum componere, eadem apposite vel canere vel dicere. Denique commemorat Elfredus carmen triviale, quod adhuc vulgo cantitatur, Aldelmum fecisse, aditiens causam qua probet rationabiliter tantum virum his quae videantur frivola institisse. Populum eo tempore semibarbarum, parum divinis sermonibus intentum, statim, cantatis missis, domos cursitare solitum. Ideo sanctum virum, super pontem qui rura et urbem continuat, abeuntibus se opposuisse obicem, quasi artem cantitandi professum. Eo plusquam semel facto, plebis favorem et concursum emeritum. Hoc commento sensim inter ludicra verbis Scripturarum insertis, cives ad sanitatem reduxisse." William of Malmesbury, De gestis pontificum Anglorum, ed. Hamilton, Rolls Series, 1870, 336.

[655] "Reverentissimo patri meaeque rudis infantiae venerando praeceptori Adriano." Epist. (Aldhelmi Opera, ed. Giles, 1844, p. 330).

[656] Faricius, Life, in Giles' edition of Aldhelm, 1844, p. 357.

[657] Letter of Cuthbert to Cuthwine, describing Bede's last illness. "Et in nostra lingua, hoc est anglica, ut erat doctus in nostris carminibus, nonnulla dixit. Nam et tunc Anglico carmine componens, multum compunctus aiebat, etc." The letter is quoted by Simeon of Durham, ed. Arnold, Rolls Series, 1882, I, pp. 43-46, and is extant elsewhere, notably in a ninth century MS at St Gall.

[658] "quid Hinieldus cum Christo."

[659] "Þæt ǣnig prēost ne bēo ealuscop, ne on ǣnige wīsan glīwige, mid him sylfum oþþe mid ōþrum mannum"—Thorpe, Ancient Laws and Institutes of England, 1840, p. 400 (Laws of Edgar, cap. 58).

[660] "avitae gentilitatis vanissima didicisse carmina." This charge is dismissed as "scabiem mendacii." Vita Sancti Dunstani, by "B," in Memorials of Dunstan, ed. Stubbs, Rolls Series, 1874, p. 11. Were these songs heroic or magic?

[661] The Heroic Legends of Denmark, New York, 1919, p. 32 (footnote).

[662] Ibid. p. 39.

[663] Thus, much space has been devoted to discussing whether "Gotland," in the eleventh century Cotton MS of Alfred's Orosius, signifies Jutland. I believe that it does; but fail to see how it can be argued from this that Alfred believed the Jutes to be "Geatas." Old English had no special symbol for the semi-vowel J; so, to signify Jōtland, Alfred would have written "Geotland" (Sievers, Gram. §§ 74, 175). Had he meant "Land of the Geatas" he would have written "Geataland" or "Geatland." Surely "Gotland" is nearer to "Geotland" than to "Geatland."

[664] P.B.B. XII, 1-10.

[665] See above, p. 8. Fahlbeck has recently revised and re-stated his arguments.

[666] Danmarks Riges Historie, I, 79 etc.

[667] Beowulf, übersetzt von H. Gering, 1906, p. vii.

[668] See above, also Nordisk Aandsliv, 10, where Olrik speaks of the Geatas as "Jyderne." His arguments as presented to the Copenhagen Philologisk-historisk Samfund are summarized by Schütte, J.E.G. Ph. XI, 575-6. Clausen also supports the Jute-theory, Danske Studier, 1918, 137-49.

[669] J.E.G. Ph. XI, 574-602.

[670] Beowulf, et Bidrag til Nordens Oldhistorie af Chr. Kier, København, 1915.

[671] This is admitted by Bugge, P.B.B. XII, 6. "Geátas ... ist sprachlich ein ganz anderer name als altn. Jótar, Jútar, bei Beda Jutae, und nach Beda im Chron. Sax. 449 Jotum, Jutna ... Die Geátas ... tragen einen namen der sprachlich mit altn. Gautar identisch ist."

[672] From a presumed Prim. Germ. *Eutiz, *Eutjaniz. The word in O.E. seems to have been declined both as an i-stem and an n-stem, the n-stem forms being used more particularly in the gen. plu., just as in the case of the tribal names, Seaxe, Mierce (Sievers, § 264). The Latinized forms show the same duplication, the dat. Euciis pointing to an i-stem, the nom. Euthio to an n-stem, plu. *Eutiones. For a discussion of the relation of the O.E. name to the Danish Jyder, see Björkman in Anglia, Beiblatt, XXVIII, 274-80: "Zu ae. Eote, Yte, dän. Jyder 'Jüten'."

[673] I regard it as simply an error of the translator, possibly because he had before him a text in which Bede's Iutis had been corrupted in this place into Giotis, as it is in Ethelwerd: Cantuarii de Giotis traxerunt originem, Vuhtii quoque. (Bk. I: other names which Ethelwerd draws from Bede in this section are equally corrupt.)

Bede's text runs: (I, 15) Aduenerant autem de tribus Germaniae populis fortioribus, id est Saxonibus, Anglis, Iutis. De Iutarum origine sunt Cantuarii et Victuarii; in the translation: "Comon hi of þrim folcum ðam strangestan Germanie, þæt [is] of Seaxum and of Angle and of Geatum. Of Geata fruman syndon Cantware and Wihtsætan": (IV, 16) In proximam Iutorum prouinciam translati ... in locum, qui uocatur Ad Lapidem; "in þa neahmægðe, seo is gecegd Eota lond, in sume stowe seo is nemned Æt Stane" (Stoneham, near Southampton). MS C.C.C.C. 41 reads "Ytena land": see below.

[674] Two Saxon Chronicles, ed. Plummer, 1899. Introduction, pp. lxx, lxxi.

[675] The O.E. version of Bede's Ecclesiastical History, ed. Miller, II, xv, xvi, 1898.

[676] Florentii Wigorn. Chron., ed. Thorpe, II, 45; I, 276.

[677] It cannot be said that this is due to textual corruption in our late copy, for the alliteration constantly demands a G-form, not a vowel-form.

[678] See pp. 8, 9 above, §§ 2-7.

[679] Just as, for example, in Heimskringla: Haraldz saga ins hárfagra, 13-17, the Götar are constantly mentioned, because the kingdom of Sweden is being attacked from their side.

[680] Procopius tells us that there were in Thule (i.e. the Scandinavian peninsula) thirteen nations, each under its own king: βασιλεῖς τέ εἰσι κατὰ ἔθνος ἕκαστον ... ὧν ἔθνος ἓν πολυάνθρωπον οἱ Γαυτοί εἰσι (Bell. Gott. ii, 15).

[681] On this alliteration-test, which is very important, see above, pp. 10-11.

[682] Geta was the recognized Latin synonym for Gothus, and is used in this sense in the sixth century, e.g. by Venantius Fortunatus and Jordanes. And the Götar are constantly called Gothi, e.g. in the formula rex Sueorum et Gothorum (for the date of this formula see Söderqvist in the Historisk Tidskrift, 1915: Ägde Uppsvearne rätt att taga och vräka konung); or Saxo, Bk. XIII (ed. Holder, p. 420, describing how the Gothi invited a candidate to be king, and slew the rival claimant, who was supported by the legally more constitutional suffrages of the Swedes); or Adam of Bremen (as quoted below).

[683] Folknamnet Geatas, p. 5 etc.

[684] Speaking of the Götaelv, Adam says "Ille oritur in praedictis alpibus, perque medios Gothorum populos currit in Oceanum, unde et Gothelba dicitur." Adami Canonici Bremensis, Gesta Hamm. eccl. pontificum, Lib. IV, in Migne, CXLVI, 637. Modern scholars are of the opinion that the borrowing has been rather the other way. According to Noreen the river Götaelv (Gautelfr) gets its name as the outflow from Lake Væner. (Cf. O.E. gēotan, gēat, "pour.") Götland (Gautland) is the country around the river, and the Götar (Gautar) get their name from the country. See Noreen, Våra Ortnamn och deras Ursprungliga Betydelse, in Spridda Studier, II, 91, 139.

[685] The Scholiast, in his commentary on Adam, records the later state of things, when the Götar were confined to the south of the river: "Gothelba fluvius a Nordmannis Gothiam separat."

[686] Heimskringla, cap. 17.

[687] "Hann [Haraldr] er úti á herskipum allan vetrinn ok herjar á Ránríki" (cap. 15). "Haraldr konungr fór víða um Gautland herskildi, ok átti þar margar orrostur tveim megin elfarinnar.... Síðan lagði Haraldr konungr land alt undir sik fyrir norðan elfina ok fyrir vestan Væni" (cap. 17). Heimskringla: Haraldz saga ins hárfagra, udgiv. F. Jónsson, København, 1893-1900.

[688] Baltzer (L.), Glyphes des rochers du Bohuslän, avec une préface de V. Rydberg, Gothembourg, 1881. See also Baltzer, Några af de viktigaste Hällristningarna, Göteborg, 1911.

[689] Guinchard, Sweden: Historical and Statistical Handbook, 1914, II, 549.

[690] See Chadwick, Origin, 93; Heroic Age, 51.

[691] ll. 2910-21. See Schütte, 579, 583.

[692] ll. 2922-3007.

[693] ll. 3018-27.

[694] ll. 3029-30.

[695] pp. 575, 581.

[696] The reason for locating the Eudoses in Jutland is that the name has, very hazardously, been identified with that of the Jutes, Eutiones. Obviously this argument could no longer be used, if the Eudoses were the "Wederas."

[697] See e.g. Schütte, 579-80.

[698] Beowulf, 1856.

[699] Beowulf, 1830 etc.

[700] Beowulf, 2394. See Schütte, 576-9.

[701] Sēo ēa þǣr wyrcþ micelne sǣ. Orosius, ed. Sweet, 12, 24.

[702] See above, p. 7.

[703] As Miss Paues, herself a Geat, points out to me.

[704] Kier, 39; Schütte, 582, 591 etc.

[705] See above, pp. 99, 100.

[706] Vendel och Vendelkråka in A.f.n.F. XXI, 71-80: see Essays, trans. Clark Hall, 50-62.

[707] This grave mound is mentioned as "Kong Ottars Hög" in Ättartal för Swea och Götha Kununga Hus, by J. Peringskiöld, Stockholm, 1725, p. 13, and earlier, in 1677, it is mentioned by the same name in some notes of an antiquarian survey. That the name "Vendel-crow" is now attached to it is stated by Dr Almgren. These early references seem conclusive: little weight could, of course, be carried by the modern name alone, since it might easily be of learned origin. The mound was opened in 1914-16, and the contents showed it to belong to about 500 to 550 A.D., which agrees excellently with the date of Ohthere. See two articles in Fornvännen for 1917: an account of the opening of the mound by S. Lindqvist entitled "Ottarshögen i Vendel" (pp. 127-43) and a discussion of early Swedish history in the light of archaeology, by B. Nerman, "Ynglingasagan i arkeologisk belysning" (esp. pp. 243-6). See also Björkman in Nordisk Tidskrift, Stockholm, 1917, p. 169, and Eigennamen im Beowulf, 1920, pp. 86-99.

[708] See Appendix F: Beowulf and the Archæologists, esp. p. 356, below.

[709] By the Early Iron Age, Engelhardt meant from 250 to 450 A.D.: but more recent Danish scholars have placed these deposits in the fifth century, with some overlapping into the preceding and succeeding centuries (Müller, Vor Oldtid, 561; Wimmer, Die Runenschrift, 301, etc.). The Swedish archæologists, Knut Stjerna and O. Almgren, agree with Engelhardt, dating the finds between about 250 and 450 A.D. (Stjerna's Essays, trans. Clark Hall, p. 149, and Introduction, xxxii-iii).

[710] Essays on questions connected with the O.E. poem of Beowulf, trans. and ed. by John R. Clark Hall, (Viking Club), Coventry. (Reviews by Klaeber, J.E.G.Ph. XIII, 167-73, weighty; Mawer, M.L.N. VIII, 242-3; Athenæum, 1913, I, 459-60; Archiv, CXXXII, 238-9; Schütte, A.f.n.F. XXXIII, 64-96, elaborate.)

[711] An account of these was given at the time by H. Stolpe, who undertook the excavation. See his Vendelfyndet, in the Antiqvarisk Tidskrift för Sverige, VIII, 1, 1-34, and Hildebrand (H.) in the same, 35-64 (1884). Stolpe did not live to issue the definitive account of his work, Graffältet vid Vendel, beskrifvet af H. Stolpe och T. J. Arne, Stockholm, 1912.

[712] Also added as an Appendix to his Beowulf translation, 1911.

[713] Clark Hall's Preface to Stjerna's Essays, p. xx.

[714] J.E.G.Ph. XIII, 1914, p. 172.

[715] Essays, p. 239: cf. p. 84.

[716] p. 39.

[717] Germania, cap. XV.

[718] ll. 378, 470.

[719] Cassiodorus, Variae, V, 1.

[720] Walter, Corpus juris Germanici antiqui, 1824, II, 125.

[721] Heimskringla, Haraldz saga, cap. 38-40.

[722] "The idea of a gold hoard undoubtedly points to the earlier version of the Beowulf poem having originated in Scandinavia. No such 'gold period' ever existed in Britain." Essays, p. 147.

[723] Cottonian Gnomic Verses, ll. 26-7.

[724] l. 14.

[725] Exeter Gnomic Verses, l. 126.

[726] Baldwin Brown, III, 385, IV, 640.

[727] B. l. 19.

[728] l. 339.

[729] l. 991.

[730] Cf. Falk, Altnordische Waffenkunde, 28.

[731] I would suggest this as the more likely because, if the ring were inserted for a practical purpose, it is not easy to see why it later survived in the form of a mere knob, which is neither useful nor ornamental. But if it were used to attach the symbolical "peace bands," it may have been retained, in a "fossilized form," with a symbolical meaning.

[732] Most editors indeed do take it in this sense, though recently Schücking has adopted Stjerna's explanation of "ring-sword." In l. 322, Falk (27) takes hring-īren to refer to a "ring-adorned sword," though it may well mean a ring-byrnie.

[733] Actually, I believe, more: for two ring-swords were found at Faversham, and are now in the British Museum. For an account of one of them see Roach Smith, Collectanea Antiqua, 1868, vol. VI, 139. In this specimen both the fixed ring and the ring which moves within it are complete circles. But in the Gilton sword (Archæologia, XXX, 132) and in the sword discovered at Bifrons (Archæologia Cantiana, X, 312) one of the rings no longer forms a complete circle, and in the sword discovered at Sarre (Archæol. Cant. VI, 172) the rings are fixed together, and one of them has little resemblance to a ring at all.

[734] At Concevreux. It is described by M. Jules Pilloy in Mémoires de la Société Académique de St Quentin, 4^e Sér. tom. XVI, 1913; see esp. pp. 36-7.

[735] See Lindenschmit, "Germanisches Schwert mit ungewöhnlicher Bildung des Knaufes," in Die Altertümer unserer heidnischen Vorzeit, V Bd., V Heft, Taf. 30, p. 165, Mainz, 1905.

[736] Salin has no doubt that the Swedish type from Uppland (his figure 252) is later than even the latest type of English ring-sword (the Sarre pommel, 251) which is itself later than the Faversham (249) or Bifrons (250) pommel. See Salin (B.), Die Altgermanische Thierornamentik, Stockholm, 1904, p. 101. The same conclusion is arrived at by Lindenschmit: "Die ursprüngliche Form ist wohl in dem, unter Nr. 249 von Salin abgebildeten Schwertknopf aus Kent zu sehen"; and even more emphatically by Pilloy, who pronounces the Swedish Vendel sword both on account of its "ring" and other characteristics, as "inspirée par un modèle venu de cette contrée [Angleterre]."

[737] The Benty Grange helmet; see below, p. 358.

[738] Depicted by Clark Hall, Stjerna's Essays, p. 258.

[739] Clark Hall's Beowulf, p. 227.

[740] "Von Skandinavien gibt es aus der Völkerwanderungszeit und Wikingerepoche keine archäologischen Anhaltspunkte für das Tragen des Panzers, weder aus Funden noch aus Darstellungen," Max Ebert in Hoops' Reallexikon, III, 395 (1915-16). But surely this is too sweeping. Fragments of an iron byrnie, made of small rings fastened together, were found in the Vendel grave 12 (seventh century). See Graffältet vid Vendel, beskrifvet af H. Stolpe och T. J. Arne, pp. 49, 60, plates xl, xli, xlii.

[741] 54-I. Liebermann, p. 114.

[742] Essays, 34-5.

[743] Elene, 264.

[744] Engelhardt, Denmark in the Early Iron Age, p. 66.

[745] Andreas, 303.

[746] l. 2869.

[747] "Few have corslets and only one here and there a helmet" (Germania, 6). In the Annals (II, 14) Tacitus makes Germanicus roundly deny the use of either by the Germans: non loricam Germano, non galeam.

[748] See above, p. 124.

[749] See Chifflet, J. J., Anastasis Childerici I ... sive thesaurus sepulchralis, Antverpiæ, Plantin, 1655.

[750] That both sword and scramasax were buried with Childeric is shown by Lindenschmit, Handbuch, I, 236-9: see also pp. 68 etc.

[751] l. 2762-3.

[752] Worsaae, Nordiske Oldsager, Kjøbenhavn, 1859; see No. 499; Roach Smith, Collectanea Antiqua, 1852, II, 164; Montelius, Antiq. Suéd. 1873, No. 294 (p. 184).

[753] Essays, p. 198. See also above, p. 124. Mr Reginald Smith writes to me: "Unburnt objects with cremated burials in prehistoric times (Bronze, Early and late Iron Ages) are the exception, and are probably accidental survivals from the funeral pyre. In such an interpretation of Beowulf I agree with the late Knut Stjerna, who was an archæologist of much experience."

[754] Forming vols. 3 and 4 of The Arts in Early England, 1903-15.

[755] It was, however, necessary to leave over for a supplementary volume some of the contributions most interesting from the point of view of the archæology of Beowulf: e.g. spatha, speer, schild.

[756] B. E. Hildebrand, Grafhögarne vid Gamla Upsala, Kongl. Vitterhets Historie och Antiqvitets Akademiens Månadsblad, 1875-7, pp. 250-60.

[757] Fasta fornlämningar i Beovulf, in Antiqvarisk Tidskrift för Sverige, XVIII, 48-64.

[758] Heimskringla: Ynglingasaga, cap. 25, 26, 29.

[759] See B. Nerman, Vilka konungar ligga i Uppsala högar? Uppsala, 1913, and the same scholar's Ynglingasagan i arkeologisk belysning, in Fornvännen, 1917, 226-61.

[760] Heimskringla: Ynglingasaga, cap. 27.

[761] A discovery made by Otto v. Friesen in 1910: see S. Lindqvist in Fornvännen, 1917, 129. Two years earlier (1675) "Utters högen i Wändell" is mentioned in connection with an investigation into witchcraft. See Linderholm, Vendelshögens konunganamn, in Namn och Bygd, VII, 1919, 36, 40.

[762] For a preliminary account of the discovery, see Ottarshögen i Vendel, by S. Lindqvist in Fornvännen, 1917, 127-43, and for discussion of the whole subject, B. Nerman, Ottar Vendelkråka och Ottarshögen i Vendel, in Upplands Fornminnesförenings Tidskrift, VII, 309-34.

[763] Baldwin Brown, III, 216.

[764] 213.

[765] 218.

[766] So Baldwin Brown, III, 213; Lorange, Den Yngre Jernalders Sværd, Bergen, 1889, passim.

[767] Baldwin Brown, III, 215.

[768] It is somewhat similar in Norse literature, where swords are constantly indicated as either inherited from of old, or coming from abroad: cf. Falk, 38-41.

[769] Beowulf, 1489, wǣgsweord; cf. Vægir as a sword-name in the Thulur. In ll. 1521, 1564, 2037, hringmǣl may refer to the ring in the hilt, and terms like wunden- are more likely to refer to the serpentine ornament of the hilt. This must be the case with wyrm-fāh (1698) as it is a question of the hilt alone. Stjerna (p. 111 = Essays, 20) and others take āter-tānum fāh (1459) as referring to the damascened pattern (cf. eggjar ... eitrdropom innan fáþar; Brot af Sigurðarkviðu). It is suggested however by Falk (p. 17) that tān here refers to an edge welded-on: the Icelandic egg-teinn.

[770] The only certainly Anglo-Saxon helmet as yet discovered: traces of what may have been a similar head-piece were found near Cheltenham: Roach Smith, Collectanea Antiqua, II, 1852, 238.

[771] Coll. Ant. II, 1852, 239; Bateman, Ten Years' Diggings, 30; Catalogue of the Antiquities preserved in the Museum of Thomas Bateman, Bakewell, 1855.

[772] A very good description of these continental "Spangenhelme" is given in the magnificent work of I. W. Gröbbels, Der Reihengräberfund von Gammertingen, München, 1905. These helms had long been known from a specimen (place of origin uncertain) in the Hermitage at Petrograd, and another example, that of Vézeronce, supposed to have been lost in the battle between Franks and Burgundians in 524. Seven other examples have been discovered in the last quarter of a century, including those of Baldenheim (for which see Henning (R.), Der helm von Baldenheim und die verwandten helme des frühen mittelalters, Strassburg, 1907, cf. Kauffmann, Z.f.d.Ph. XL, 464-7) and Gammertingen. They are not purely Germanic, and may have been made in Gaul, or among the Ostrogoths in Ravenna, or further east.

[773] Stjerna, Essays, p. 11 = Studier tillägnade Oscar Montelius af Lärjungar, 1903, p. 104: Clark Hall, Beowulf, 1911, p. 228.