^{page 79 note 1} The standard reference work is Charles Carr, T., Nominal Compounds in Germanic, St Andrews Univ. Publ. 41 (London, 1939)Google Scholar. In his introduction (pp. xx-xxix) he gives a review of valuable writings previous to his; his entire Part I, chs. I-V, pp. 3–157, is devoted to ‘The Stock of Compounds’. I have also found of value Koziol, Herbert, Handbuch der englischen Wortbildungslebre, 2nd, rev., ed. (Heidelberg, 1972)Google Scholar.

^{page 80 note 1} Die Kenningar in der angelsächsischen Dichtung, dissertation, Strassburg University (Darmstadt and Leipzig, 1886).Google Scholar

^{page 80 note 2} Die altenglischen Kenningar: ein Beitrag zur Stilkunde altgermanischer Dichtung, Schriften der Königsberger Gelehrten Gesellschaft, Geisteswissenschaftliche Klasse 14 (Halle, 1938).

^{page 81 note 1} The Art of Beowulf (Berkeley and Los Angeles), p. 15.Google Scholar

^{page 81 note 2} Beowulf with the Finnesburg Fragment (London, 1953), p. 80.Google Scholar

^{page 81 note 3} ‘Old West Germanic and Old Norse’, Studies in Philology: a Miscellany in honor of Frederick Klaeber, ed. Malone, Kemp and Ruud, Martin (Minneapolis, 1929), p. 16Google Scholar; §§4–7, pp. 16–17, for interpretation of representative words.

^{page 81 note 4} Die altenglischen Kenningar, pp. 141–2.

^{page 82 note 1} Die Kenningar, p. 55.

^{page 82 note 2} The Kenning in Anglo-Saxon and Old Norse Poetry (Oxford, 1929), p. 78.Google Scholar

^{page 82 note 3} Die altenglischen Kenningar, p. 226.

^{page 83 note 1} In his ‘Prefatory Remarks on Prose Translation of “Beowulf”’, Beowulf and the Finnesburg Fragment, trans. Hall, John R. Clark, new ed. (London, 1950), pp. ix–xliiiGoogle Scholar, esp. XX-XXV. See also Stefán Einarsson, ‘Compounds of the Mann-Skratti Type’, Studies in honor of Albert Morey Sturtevant, Univ. of Kansas Humanistic Stud. 29 (Lawrence, 1952), 47–56Google Scholar, esp. 47.

^{page 83 note 2} Meaning and Change of Meaning with Special Reference to the English Language, Göteborgs Högskolas Arsskrift 38 (1932:1), 296.Google Scholar

^{page 83 note 3} ibid. p. 293.

^{page 83 note 4} ibid. p. 297, n. 1; see also 363–9.

^{page 84 note 1} As the basic text for this investigation, and the one from which quotations are taken, I have used Beowulf and the Fight at Finnsburg, ed. Klaeber, Fr., 3rd ed. (Boston, 1950)Google Scholar. However, I kept ever at hand, for constant checking of interpretations, notes and disputed readings, Chambers's, R. W. revision of Beowulf with the Finnsburg Fragment, ed. Wyatt, A. J. (Cambridge, 1953)Google Scholar; Beowulf and Judith, ed. Dobbie, Elliott van Kirk, The Anglo-Saxon Poetic Records 4 (New York, 1953)Google Scholar; Beowulf with the Finnesburg Fragment, ed. Wrenn, C. L., 1 st ed. (London, 1953)Google Scholar and 2nd ed. (London, 1958; the third edition, fully rev. W. F. Bolton (London, 1973), has been of little service to me, since for this study the worth lay in Wrenn's own erudition and critical acumen and sensitivity); and Heyne-Schückings Beowulf, ed. Else von Schaubert, 17th ed., Bibliothek der ältesten deutschen Literatur-Denkmäler 3 (Paderborn, 1958–61). The three translations I have consulted for their methods and manners of translation are Beowulf and the Finnesburg Fragment, trans. Hall, John R. Clark, new ed. rev. C. L. Wrenn (London, 1950)Google Scholar; Donaldson, E. Talbot, Beowulf: a New Prose Translation (New York, 1966Google Scholar; far the best of all the prose translations I have seen); and Beowulf and its Analogues, trans. G. N. Garmonsway and Jacqueline Simpson (London, 1968)Google Scholar, including Hilda Ellis Davidson, ‘Archaeology and Beowulf’.

^{page 85 note 1} Compare Klaeber's note on wala (1031) with Rupert Bruce-Mitford's full explanation of the term based on the Sutton Hoo helmet in his appendix to Hodgkin, R. H., A History of the Anglo-Saxons, 3rd ed. (London, 1952) 11, 752–4Google Scholar, n. 21; also Aspects of Anglo-Saxon Archaeology: Sutton Hoo and other Discoveries (London, 1974), pp. 210–13Google Scholar. Compare also Brodeur's contention (Art, p. 20) that the most concretely described helmet in the poem, se hwita helm, ‘is merely an idealization of a typical helmet’ with Rosemary Cramp's full explanation of the description based primarily on two helmets from Valsgärde graves 6 and 8, lsquo;Beowulf and Archaeology’, MA 1 (1957), 61–2.Google Scholar

^{page 85 note 2} Cramp, ‘Beowulf and Archaeology’, p. 60.

^{page 86 note 1} This word does not mean ‘helmet’, as glossed by, e.g., Klaeber, Wrenn and von Schaubert; it refers to the cheek-guards alone, as distinct from the parts of the helmet covering the entire face, chin, neck and crown.

^{page 86 note 2} See Aspects, p. 200 and pl. 47.

^{page 86 note 3} Hatto, A. T., ‘Snake-Swords and Boar-Helms in Beowulf. ESts 38 (1957), 156Google Scholar, and Bruce-Mitford, , Aspects, pl. 55.Google Scholar

^{page 86 note 4} ‘Beowulf and Archaeology’, p. 61.

^{page 88 note 1} Die altenglischen Kenningar, p. 155.

^{page 89 note 1} Meaning, p. 308. In all discussions of metaphors I follow the principles set forth by Stern, ibid. ch. XI, pp. 296–310, esp. 300 and 307–9.

^{page 89 note 2} ‘The Synonyms for “Sea” in Beowulf’, Studies in honor of Albert Morey Sturtevant, pp. 22–46, at 39–40.

^{page 90 note 1} See Can's classification of compounds consisting of substantive plus substantive, Nominal Compounds, pp. 321–4. In this study I am concerned only with the determinatives.

^{page 91 note 1} For the archaeological evidence bearing upon the sword I rely primarily upon Behmer, Elis, Das zweischneidige Schwert der germanischen Vōlkerwanderungszeit, Akademisk Avhandlung Stockholms Högskola (Stockholm, 1939)Google Scholar; Shetelig, Haakon and Falk, Hjalmar, Scandinavian Archaeology, trans. Gordon, E. V. (Oxford, 1937)Google Scholar; Falk, Hjalmar, Altnordische Waffenkunde, Videnskapsselskapets Skrifter 11, Hist. Fil. Kl. 1914, no. 60 (Kristiania, 1914), 9–65Google Scholar (hereinafter AW); Brown, G. Baldwin, The Arts in Early England III. Saxon Art and Industry in the Pagan Period (New York, 1915)Google Scholar; and Davidson, H. R. Ellis, The Sword in Anglo-Saxon England: its Archaeology and Literature (Oxford, 1962)Google Scholar. I also find of value May Keller, Lansfield, The Anglo-Saxon Weapon Names, Anglistische Forschungen 15 (Heidelberg, 1906)Google Scholar, although subsequent archaeological and linguistic research have rendered obsolete some of her conclusions. Unfortunately Davidson's book is marred by errors of many, often substantial, kinds; cf., e.g., the reviews by Evison, Vera I., MÆ 32 (1963), 136–40Google Scholar; Green, D. H., MLR 57 (1962) 591–2CrossRefGoogle Scholar; and Blake, N. F., N & Q 207 (1968), 451–3.Google Scholar Accordingly I have depended on it, in instances in which I have been unable to check its accuracy, only when I have had no alternative inasmuch as it is the sole comprehensive work on the subject.

^{page 91 note 2} Although this word occurs only once as a simplex, it occurs also in the compound beorosweng (1590a), which appears to be a true synonym of the sweordes sweng of 23 86a.

^{page 91 note 3} Athelstanthe Ætheling, son of King Æthelred of Wessex, in a will dating from c.1015 bequeathed to his brother Eadmund þas sweordes þe Offa cyng ahte. As Dorothy Whitelock notes (Anglo-Saxon Wills, Cambridge Stud, in Eng. Legal Hist. (Cambridge, 1930), p. 171Google Scholar): ‘Offa reigned from 757 to 796, so that the sword had been handed down for over two hundred years.’

^{page 92 note 1} Quoted from Beowulf, ed. Klaeber, app. IV. In my interpretations of the Hildebrandslied I also found useful Braune, Wilhelm, Althochdeutsche Lesebuch, 15th ed., rev. Ernst A. Ebbinghaus (Tübingen, 1969).Google Scholar

^{page 92 note 2} Keller, , Weapon Names, pp. 178–9Google Scholar, and Falk, , Scandinavian Archaeology, p. 379.Google Scholar

^{page 92 note 3} Heliand und Genesis, ed. Behaghel, Otto, 7th ed., rev. Walther Mitzka (Tübingen, 1958)Google Scholar used together with Sehrt, Edward H., Vollständiges Wörterbuch zum Heliand und zur altsächsischen Genesis, 2nd ed., Hesperia: Schriften zur germanischen Philologie (Göttingen, 1966)Google Scholar. The Gothic meki in the context in which it stands, Ephesians vi.17, means simply ‘sword’: nimaiþ … meki abmins ‘take… the sword of the Holy Spirit’.

^{page 92 note 4} Scandinavian Archaeology, p. 379.

^{page 92 note 5} See Evison's explanation, MÆ 32, 139. Wiglaf also apparently used his bill with a thrusting motion to penetrate the dragon's soft underside (2699–700), but the poet does not use ord in this passage.

^{page 92 note 6} Behmer, , Das zweischneidige Schwert, pl. XXXIIGoogle Scholar:6; however, Behmer (ibid. p. 83) believes it to be a deviation from the typical narrow northern type v, an offshoot of the continental broad blade Indeed, the broad sword is not typical in either the north or England; it is primarily a mainland German type. The plate shows the Fairford sword sheathed, unfortunately, so that we cannot ascertain the type of the tip. See Behmer's discussion of English swords, ibid. pp. 157–8. See also Wilson, David M., The Anglo-Saxons, 2nd ed. (Harmondsworth, 1975), pp. 108–13.Google Scholar

^{page 93 note 1} Das zweischneidige Schwert, p.38.

^{page 93 note 2} ibid. pl. XXX; in its restored form, with scabbard concealing the blade, pl. XXIX.

^{page 93 note 3} Cramp, , ‘Beowulf and Archaeology’, p. 67Google Scholar.

^{page 93 note 4} The Sword, p. 147.

^{page 93 note 5} ‘Snake-Swords’, p. 148.

^{page 94 note 1} Falk, , Scandinavian Archaeology, p. 379Google Scholar, and Evison, , MÆ 32, 137–8Google Scholar, with her comments on Davidson's wavering and uncertainty.

^{page 94 note 2} Davidson, , The Sword, pp. 379–80 and pl. 60Google Scholar. Falk's use of the term ‘false damascening’ results, of course, from the fact that his original Altnordische Waffenkunde was published in 1914 and his co-operative work with Shetelig, , Scandinavian Archaeology, was published in 1937Google Scholar, whereas it was not until 1947 that Maryon, Herbert (Proc. of the Cambridge Ant. Soc. 41, 73–4Google Scholar) introduced the current and more accurate term ‘pattern welding’.

^{page 94 note 3} See above, pp. 81 and 88, and below, pp. 136–7.

^{page 95 note 1} Bosworth, J. and Toller, T. N., An Anglo-Saxon Dictionary (Oxford, 1882–1898)Google Scholar; Toller, T. N., Supplement, with Enlarged Addenda and Corrigenda by Campbell, A. (Oxford, 1972).Google Scholar

^{page 95 note 2} Falk, , Scandinavian Archaeology, p. 380Google Scholar, n. 1, notes that ‘except when loosely applied to the handle and guard together, bilt in English corresponds normally to ON meðalkafli “middle-piece”;’, that is, the handle or grip.

^{page 95 note 3} ibid. p. 386.

^{page 95 note 4} The Sword, p. 61.

^{page 95 note 5} Bruce-Mitford, Rupert, The Sutton Hoo Ship Burial: a Handbook, 2nd ed. (London, 1972Google Scholar; hereinafter 1972 Handbook), p. 76; also fig. 9, p. 32.

^{page 96 note 1} The Heroic Legends of Denmark, trans. Lee M. Hollander, American-Scandinavian Foundation (New York, 1919), pp. 145–6.Google Scholar

^{page 96 note 2} The argument concerning the interpretation of 1136b–45b, particularly 1I42a–4b, is typified and perhaps best set forth by the Malone-Brodeur feud, esp. Malone, Kemp, ‘The Finn Episode in Beowulf’, JEGP 25 (1926), 157–72Google Scholar, particularly 158–9; Brodeur, Arthur G., ‘The Climax of the Finn Episode’, Univ. of California Pub. in Eng. 3.8 (1943), 285–362Google Scholar, particularly 330–55; ‘Design and Motive in the Finn Episode’, Essays and Stud., Univ. of California Pub. in Eng. 14 (1943), 1–42Google Scholar; and Malone, Kemp, ‘Hildeburg and Hengest’, ELH 10 (1943), 257–84CrossRefGoogle Scholar, particularly 276–81. A good summary of their arguments is given by Fry, Donald K., Finnsburb Fragment and Episode (London, 1974), pp. 9–11.Google Scholar

^{page 96 note 3} The Origin of the English Nation (Cambridge, 1907), p. 52, n. 1.Google Scholar

^{page 97 note 1} ‘Skjoldungasaga i Arngrim Jonssons Udtog’, ed. Olrik, Axel, Aarboger for Nordisk Oldkyndigbed og Historic, ser. II 9 (1894), 83–164Google Scholar. Ch. 4 of ‘“Rerum Danicarum fragmenta” og “Regum Sveciæ catalogus”’, p. 107 gives a genealogical table showing the line of Herleifus, the fourth king of Denmark. It lists the six sons of Leifus cogn. hinn frekne, second son of Herleifus, as Herleifus, Hunleifus [MS Humleifus], Aleifus, Oddleifus, Geirleifus, Gunnleifus. Chadwick, who first noted a possible connection between this Hunleifus and Hunlafing, did not suggest Hunlafing to be the son of Hunleifus. All he says is that the names Oddleifus and Gunnleifus ‘are identical with those of two of Hengest's warriors … while Hengest himself is in possession of a sword called Hunlafing … This can hardly be accidental coincidence.’ The connection between Arngrímur's Hunleifus and Hunlafing was made by Huchon, R. in his review of Chadwick, Revue Germanique 3 (1907), 626Google Scholar, n. 1; and he identified Hunlafing only as ‘un homme, l'un des compagnons de Hengest’. Rudolph Imelmann in his review of the Heyne-Schücking Beowulf, 8th ed., Deutsche Literaturzeitung 30 (1909), 999Google Scholar, carried the connection further and identified Hunlafing as the brother of Guthlaf and Oslaf; he believed that the three brothers came to Frisia together to carry out the vengeance and take Hildeburh home.

^{page 97 note 2} See the varying and contradictory accounts of the Scyldings in the translated analogues, including further chapters from Arngrimur's Skjǫldunga saga, in Beowulf and its Analogues, trans. Garmonsway and Simpson, pp. 118–211. Saxo Grammaticus, a reasoning historian trying to reconcile conflicting traditions yet wishing to give a complete history of Denmark from its beginnings to his own day, conscientiously gathered together all the current traditions, Icelandic and Norwegian as well as Danish. For his method in treating them, see my The Legends of Ermanaric (Berkeley and Los Angeles, 1943)Google Scholar, ch. 3, esp. pp. 136–43.

^{page 97 note 3} Deutsche Literaturzeitung 30, 999, and Chambers, R. W., Widsitb: a Study in Old English Heroic legend (Cambridge, 1912)Google Scholar, app. K, p. 254.

^{page 98 note 1} Brodeur, endeavouring to support the contention that Hunlafing means ‘son of Hunlaf’, did not make his case that the patronymic in -ing could in Old English poetry stand alone without the personal name of the individual; for all his ‘supporting evidence’ from Old English and Old Norse (‘The Climax’, pp. 330–55, esp. 340–1) of the use of constructions like sunu Healfdenes is beside the point, since these are not patronymics in the true sense of being formed by means of prefix or suffix. In point of fact, sunu Healfdenes is the very term that in The Art (p. 25 2) he selects to illustrate the viðkenning, a type of non-kenning which he defines as ‘a characterization in terms of a specific person or named thing … [it] is pronominal in function: it stands for a name… [it] identifies a specific individual and stands for him alone’ (pp. 251–2). This construction is not at all the same thing as a patronymic. On the other hand Falk, , AW, p. 65Google Scholar, asserts that in the formation of Old Norse sword-names, besides pure adjective constructions and suffix-less nomina agentis, the suffixes -(n)ir and -(n)ingr predominate.

^{page 98 note 2} Heroic Legends, p. 145.

^{page 98 note 3} ‘Hildeburg and Hengest’, p. 277.

^{page 98 note 4} AW, p. 65.

^{page 98 note 5} ibid. pp. 64–5.

^{page 98 note 6} See above, p. 95.

^{page 99 note 1} AW, pp. 31 and 57, and Scandinavian Archaeology, pp. 377–8; see also Shetelig, Scandinavian Archaeology, p. 204.

^{page 99 note 2} Davidson, , The Sword, pp. 185–6, 76–7 and 210.Google Scholar

^{page 99 note 3} See below, pp. 106–7.

^{page 99 note 4} Woolf, Rosemary, ‘The Ideal of Men Dying with their Lord in the Germania and in The Battle of Maldon’, ASE 5 (1976), 69–76.Google Scholar

^{page 99 note 5} Vickrey, John F., ‘The Narrative Structure of Hengest's Revenge in Beowulf, ASE 6 (1977), 91–103Google Scholar, esp. 96, has advanced the view that the poet did not mean ‘winter’ and ‘spring’ in 1131b–7a to be taken as real, but rather as symbolical: ‘The description … does not show how much time really elapsed but instead symbolizes the full course of Hengest's revenge, implying the deed itself behind the description of his mood.’ This last phrase I do not fully comprehend. But more disturbing is the current trend toward finding symbolism at every turn in Beowulf; not inapplicable to this trend are Theodore Silverstein's comments on the allegorical approach delivered to the Old English Group at the convention of the Modem Language Association of America on 27 December 1971 and printed in ASE 2 (1973), 296–7Google Scholar. Much more convincing than Vickrey's view are those of Stanley, E. G., ‘Old English Poetic Diction and the Interpretation of The Wanderer, The Seafarer and The Penitent's Prayer’, Anglia 73 (1956), 439Google Scholar, and Irving, Edward B. Jr, A Reading of Beowulf (New Haven, 1968), p. 172Google Scholar. To me the winter is quite real, and the storms and ice-bound waves exist in Hengest's mind and heart as well as in the sea.

^{page 100 note 1} Fry, , Finnsburh, pp. 23–4 and 44Google Scholar, n., would see torngemot as referring to ‘an angry Danish war-council which would result in a battle …’ Yet he himself emphasizes that the poem offers no evidence that the Danes had shut Hengest off into ‘tragic isolation’ by mistrust and reproach. Hengest had but one aim: to carry through to an effective conclusion his revenge for the slaying of Hnæf. He might have wanted a consultation with his men, but not ‘an angry confrontation’.

^{page 100 note 2} I follow Mitchell, Bruce, A Guide to Old English, 2nd ed. (Oxford, 1975)Google Scholar, §168, þat 3(b).

^{page 100 note 3} Kaske, R. E., ‘The Eotenas in Beowulf’, Old English Poetry: Fifteen Essays, ed. Creed, Robert P. (Providence, 1967), pp. 285–310Google Scholar, brought the ‘giants’ back into Finnsburg. He (at p. 289, with Fry following him) would take the eotena beam of 1088a and 1141a and the eotenum of 1145a as ‘an insulting figurative epithet for “enemies”’. Neither Kaske nor Fry, however, is able to produce substantial support for this conclusion, as Kaske is the first to acknowledge: ‘A good bit of the evidence presented above is, admittedly, a grabbing at straws’ (p. 299). Indeed, the Beowulf poet has his own word which he uses more than once to express his own attitude (or that of one of his characters) toward such people as these Jutes. That word is freca, ON freki ‘the greedy one, the voracious one’, the name of one of Othin's wolves. The poet uses five nominal compounds – three of them nonce-words – with -freca as base-word to describe the Franks, the Swedes, mercenaries, the dragon and – fascinatingly – Beowulf, the warrior also characterized as lofgeornost. The Finn episode is exactly the type of context in which we could justifiably expect to find a-freca compound. And we find one: the Bahuvrihi compound ferbðfrec ‘voracious in spirit’ applied to Finn himself (1146a). A full discussion of the meanings of the simplex and the five compounds I give in a paper in progress. I must continue to agree with Chambers, R. W., Beowulf: an Introduction, 3rd ed. with a Supplement by Wrenn, C. L. (Cambridge, 1959), p. 289Google Scholar: ‘That the King of Frisia should have had Jutes under his rule is likely enough.’ See also Beowulf, ed. Klaeber, p. 233, n. 3.

^{page 100 note 4} Mitchell, , A Guide, §168Google Scholar, swa 2(d) and §186, 2. The same construction appears in the Fragment 41b.

^{page 100 note 5} Mitchell, , A Guide, §168Google Scholar, þonne 2(a). I am grateful to Professor Stanley B. Greenfield for calling my attention to this reference.

^{page 100 note 6} According to Mitchell, A Guide, §193.7, ‘A pronoun subject is frequently not expressed. Often the subject not expressed is that of the preceding clause.’ The unexpressed subject of dyde (1144b) is the same as that expressed by he in 1138b, 1140a, 1141a and 1142a.

^{page 100 note 7} I follow Klaeber's edition, but retain MS readings, follow von Schaubert in placing a comma instead of a full stop at the end of 1141b and add a comma after Hunlafing as well as Fry's after 1144a.

^{page 101 note 1} On this type of semantic pattern see Fred C. Robinson, ‘Variation: a Study in the Diction of Beowulf’, unpublished dissertation, University of North Carolina, 1961, p. 138.

^{page 101 note 2} ibid. p. 142.

^{page 101 note 3} The only comparable word in Old English is the swurdleoma of Finnsburg 35b, and here the first element itself establishes the connection of -leoma with the sword. The ON gunnlogi cited sometimes, e.g. by Klaeber, is not a cognate, since the second element is -logi, not -Ijómi.

^{page 101 note 4} A detailed discussion of the meaning of -mal in these two and four other compounds in Beowulf is given by Davidson, , The Sword, pp. 121–6Google Scholar; see also Cramp, ‘Beowulf and Archaeology’, pp. 63–5.

^{page 101 note 5} I discuss atertan in a study of Old English kennings which is in preparation.

^{page 102 note 1} Cassiodori senatorls Variae, ed. Theodoras Mommsen, Monumenta Germaniae Historica, Auct. Antiq. 12 (Berlin, 1894), v. 1. ‘There shines a polished brightness, so that they mirror with faithful clarity the faces of those looking at them; their sides come down to an edge with such smoothness that they could be thought to be not finished with files but cast from fiery furnaces. Their midribs, hollowed out in fine grooves, appear as though fretted by some sort of little worms, where light and shadow play together so variously that you might rather believe the shining metal inwoven with various colours.’

^{page 102 note 2} ‘Snake-Swords’, p. 148.

^{page 102 note 3} ibid. p. 152, n. 12.

^{page 102 note 4} Lines 2255–7.

^{page 102 note 5} Davidson, , The Sword, pp. 28 and 107.Google Scholar

^{page 103 note 1} The meaning of guðaine in the latter context has been a subject of some controversy. A few of the older editors and some translators and critics take it to be ‘sword’; but other editors – Dobbie and von Schaubert, apparently Klaeber and Wrenn – take it as ‘warrior’. I follow these later editors.

^{page 103 note 2} Ernst Kock, A., ‘Interpretations and Emendations of Early English Texts, IV’, Anglia 42 (1918) 111–12.Google Scholar

^{page 104 note 1} Davidson gives a good summary of this material from Norse and continental sources as well as English in her concluding chapter, The Sword, pp. 211–16.

^{page 104 note 2} Meaning, p. 305.

^{page 104 note 3} ‘Interpretations’, p. 113.

^{page 104 note 4} The personification of weapons, particularly the sword, is a salient characteristic of Germanic story. Falk (A.W, 99, 41–65) presents a thorough treatment of the symbolic, supernatural and human qualities of the sword in Old Norse poetry and sagas together with his list of 176 swordnames, mentioning any peculiarities of each. Cherniss, Michael D. in his ‘The Cross as Christ's Weapon: the Influence of Heroic Literary Tradition on The Dream of the Rood’, ASE 2 (1973), 241–52Google Scholar, offers some comments on the personification of weapons in Beowulf, Waldere and The Fight at Finnsburg; however, he gives no sound evidence that personification does genuinely exist in such examples as þeab þæt wæpen duge ‘though that weapon [Hrunting] be good’ in 1660b or in scurheard ‘hardened by showers [or quenching]’ in 1033a, nor yet for his apparent belief that the convention was more or less all-pervasive in Beowulf; his survey of the convention in the weapon-riddles and in Riddle 30 is of value; and his treatment of the personification of the cross in The Dream of the Rood is interesting and illuminating. To date, however, no scholar has proffered a well-conceived, well-formulated, lucid and sound comprehensive treatment of the personification of weapons in Beowul or Old English poetry as a whole.

^{page 106 note 1} The Anglo-Saxon Poems of Beowulf, the Scop or Gleman's Tale, and the Fight at Finnsburg, ed. Thorpe, B. (Oxford, 1855; repr., 1875).Google Scholar

^{page 106 note 2} The Sword, pp. 143–4. It must be noted, however, that Davidson's conjectures based upon Beowulf mast always be cautiously tested against all evidence available to us, for she is capable of error, some of such gravity as to suggest a most cursory reading of the poem as a whole: in concluding her presentation of this conjecture she states (p. 144) that Beowulf, having prevented Dæghrefn from carrying off the torque and presumably saving Hygelac's sword, ‘could have carried it as his own from that time; be would have had the right to do so, since be succeeded Hygelac as ruler of the Geats’ (italics mine) – a statement that flies directly in the face of the facts.

^{page 107 note 1} No more satisfying is the interpretation proffered by Rosier, James L., ‘Icge Gold and Incge Lafe in Beowulf’, PMLA 81 (1966), 346CrossRefGoogle Scholar. He would take the word as the preposition in + the prefix ge-+ lafe and interpret ‘native remnant’. Now, *gelaf would be a possible word; but since ge- when prefixed to a noun usually retains something of its original pre-Primitive Germanic meaning of ‘together’ it is normally used to form collective nouns and to convey a sense of close relationship or cohesion. The three in- words in Exodus – ingefolc, 142b; ingemen, 190a; and ingeðeod, 444b – all fit this pattern: all refer to cohesive groups of individuals. * Gelaf, in turn, could mean ‘remnant’ only in the sense of ‘a small number of persons or things’; it could not refer to an individual object, a single sword. Of in- Rosier states, ‘when prefixed to nouns … [it] most commonly has the meaning “in” or “among”, and by reference to what is located or belongs in something may mean “native” … or “household”’ (p. 543). Here, too, the Exodus words fit the pattern: Edward Burroughs Irving, Jr, in his edition (The Old English Exodus (New Haven, 1955), p. 78Google Scholar, n. to line 142), states that ‘the dominant idea in all these words … is the “inland” or native quality of the Egyptians. The poet is consistently contrasting them in this respect with the Israelites, who are neither natives nor “land-men” …’ But to what is Nagling ‘native’ or in what environs does it ‘belong’? The context gives no hint. The connection between Nægling's age and its ‘native’ quality completely eludes me. Rosier's theory simply does not stand up under close scrutiny.

^{page 108 note 1} ‘Beowulf and Archaeology’, p. 65.

^{page 108 note 2} The Audience of Beowulf (Oxford, 1958), pp. 86–7.Google Scholar

^{page 108 note 3} ibid. p. 87.

^{page 109 note 1} Scandinavian Archaeologf, pp. 176–8.

^{page 109 note 2} Most editors follow Thorpe's emendation to beaðoscearpe ‘battle-sharp,’ although the MS reading beaðoscearde ‘battle-scarred’ could be defended. Ætheling Athelstan bequeathed in his will two swords scearde, one to his discþene and one to his chaplain.

^{page 110 note 1} Baldwin Brown notes (The Arts III, 235) that all spear-heads found in Anglo-Saxon cemeteries are of hammered iron.

^{page 111 note 1} On this description and the details comparable with those of the Snartemo sword, sec Cramp, ‘Beowulf and Archaeology’, pp. 66–7.

^{page 111 note 2} Bruce-Mitford, , 1972 Handbook, p. 37.Google Scholar

^{page 111 note 3} Engelhardt, Conrad, Denmark in the Early Iron Age, Illustrated by Recent Discoveries in the Peat Mosses of Slesvig (London, 1866), p. 46Google Scholar. Emphasized also by Müller, Sophus, Nordische Altertumskunde (Strassburg, 1898) 11, 128Google Scholar, and Brøndsted, Johannes, Danmarks Oldtid (Copenhagen, 1940) III, 204.Google Scholar

^{page 111 note 4} Shown by Stjerna, both Knut, Essays on Questions connected with the Old English Poem of Beowulf, trans. Hall, John R. Clark, Club, Viking extra ser. 3 (Coventry, 1912), 34Google Scholar and fig. 16, and Hildebrand, Hans, ‘Vendelfyndet. Naågra Anmärkningar’, Antiqvarisk Tidskrift för Sverige 8 (Stockholm, 1884–1891), 43.Google Scholar

^{page 112 note 1} Grein, C. W. M., Sprachsatz der angelsächsischen Dichter, new ed. Köhler, J. J. (Heidelberg, 1912).Google Scholar

^{page 112 note 2} Die altenglischen Kenningar, pp. 196–7.

^{page 112 note 3} AW, pp. 155–85, and Scandinavian Archaeology, pp. 401–4.

^{page 113 note 1} Altmestnordische Kleiderkunde mil besonderer Berücksichtigung der Terminologie, Videnskapsselskapets Skrifter II, Hist. Fil. Kl. 1918, no. 3 (Kristiania, 1919Google Scholar; hereinafter AKl).

^{page 113 note 2} Keller gives a great deal of valuable information, but she had not the benefit of Falk's later work or of later archaeological discoveries. Professor Eric Stanley has informed me that the new edition of Johannes Hoops's Reallexikon der germanischen Altertumskunde will contain all such material, but I have been unable to see fascicles thus far published.

^{page 113 note 3} Represented in Stjerna, , Essays, fig. 122, p. 258Google Scholar, and described, in part, by Engelhardt, , Denmark, p. 46.Google Scholar

^{page 113 note 4} It is pictured by Engelhardt, Denmark, Thorsbjerg pl. 1 and fully described at p. 40; pictured by Falk, , AKl, fig. 11, p. 140Google Scholar, and described at pp. 140–1.

^{page 113 note 5} Danmarks Oldtid, fig. 207, p. 210.

^{page 113 note 6} ‘Armour of the Vendel Period’, Acta Archœlologica 10 (1939), fig. 2, p. 35.Google Scholar

^{page 113 note 7} AW, p. 179.

^{page 114 note 1} Die angelsächsischen Prosabearbeitungen der Benedictinerregel, ed. Arnold

^{page 114 note 2} AW, p. 180.

^{page 115 note 1} See esp. Ilkow, Peter, Die Nominalkomposita der altsächsischen Bibeldichtung: ein semantisch-kulturgeschichtliches Glossar, ed. Wissmann, W. and Rosenfeld, H.-Fr. (Göttingen, 1968), p. 118.Google Scholar

^{page 115 note 2} AW, p. 179.

^{page 116 note 1} Robinson, ‘Variation’, p. 71.

^{page 118 note 1} The scyldas-guðgewadu of 2850–1 are to be taken as enumeration.

^{page 118 note 2} AW, p. 175, n. 1.

^{page 118 note 3} Die Gotiscbe Bibel, ed. Streitberg, Wilhelm, 4th ed. (Heidelberg, 1960).Google Scholar

^{page 120 note 1} Holthausen, F., Altenglisches etymologiscbes Wörterbuch, 3rd ed. (Heidelberg, 1974).Google Scholar

^{page 121 note 1} ‘Beowulf and Archaeology’, p. 65.

^{page 122 note 1} ‘Beowulf's Retreat from Frisia: some Textual Problems in LI. 2361–2362’, SP 62 (1965), 1–16Google Scholar, esp. 2–3; also ‘Elements of the Marvellous in the Characterization of Beowulf: a Reconsideration of the Textual Evidence’, Old English Studies in honour of John C. Pop, ed. Burlin, Robert B. and Irving, Edward B. Jr (Toronto, 1974), pp. 124–6.CrossRefGoogle Scholar

^{page 123 note 1} Ilkow, , Die Nominalkomposita, p. 323.Google Scholar

^{page 123 note 2} Cf. The Bayeux Tapestry: a Comprehensive Survey, ed. SirStenton, Frank et al. (London, 1957)Google Scholar, detail from no. 71.

^{page 123 note 3} Die altenglischen Kenningar, P. 219.

^{page 124 note 1} See the various wills in Anglo-Saxon Wills, ed. Whitelock, passim, for the proportion of helmets, byrnies and swords to shields and spears bequeathed to heirs or paid as heriot to lords; see Sophus Müller's statement (Nordische Altertumskunde, p. 128) that the Vimose byrnie of 20,000 rings would have required a full year's work by a single smith; and see Chambers, , Beowulf: an Introduction, p. 354.Google Scholar

^{page 124 note 2} Maryon, Herbert, ‘The Sutton Hoo Shield’, Antiquity 20 (1946), 21–30CrossRefGoogle Scholar; Lindqvist, Sune, ‘Sutton Hoo and Beowulf’, Antiquity 22 (1948), 135–6CrossRefGoogle Scholar; and Bruce-Mitford, , 1972 Handbook, pp. 24–6.Google Scholar

^{page 125 note 1} ‘Variation’, p. 70.

^{page 125 note 2} Engelhardt, , Denmark, pp. 48–50Google Scholar. Baldwin Brown shows (The Arts III, pl. XXI.2) a restored example of just such a Thorsbjerg shield, with the hole plugged with a small boss ‘of a plain hemispherical form that Dr Salin rightly notes is of the usual type’ (p. 198).

^{page 125 note 3} Scandinavian Archaeology, p. 397.

^{page 125 note 4} Brown, Baldwin, The Arts III 197.Google Scholar

^{page 126 note 1} Bruce-Mitford, , 1972 Handbook, pp. 24–5.Google Scholar

^{page 126 note 2} Scandinavian Archaeology, p. 398.

^{page 126 note 3} I follow the reading first suggested, for metrical reasons, by Pope, John Collins, The Rhythm of Beowulf (New Haven, 1942), p. 320Google Scholar, and since accepted by Klaeber, (Beowulf, 2nd Supplement, p. 470Google Scholar), Wrenn, (Beowulf, Supplement, p. 319Google Scholar), Dobbie, von Schaubert and others.

^{page 126 note 4} Nominal Compounds, pp. 334–6.

^{page 127 note 1} Brown, Baldwin, The Arts III, 197.Google Scholar

^{page 127 note 2} Greenfield, Stanley B., The Interpretation of Old English Poems (London, 1972), p. 73.Google Scholar

^{page 127 note 3} Beowulf, ed. Wrenn, Introduction, p. 81; see my ‘Óðinn and the Norse Jörmunrekkr-Legend’, PMLA 55 (1940), 913–14.Google Scholar

^{page 127 note 4} Brown, Baldwin, The Arts III, 201.Google Scholar

^{page 127 note 5} The Interpretation, p. 69; at pp. 69–73 he discusses various types of variation in which a shift in perspective is involved.

^{page 128 note 1} Meaning, p. 404; see also 72–4.

^{page 128 note 2} I am indebted to Professor Fred C. Robinson for drawing my attention to the parallel semantic development of the early-fourteenth-century loan-word buckler: originally the French word for a certain type of shield was escu boucler ‘shield having a boss’, but later boucle ‘boss’ came to mean ‘shield’ (as a whole).

^{page 128 note 3} See above, pp. 81 and 88, and below, pp. 136–7.

^{page 129 note 1} Falk, , Scandinavian Archaeology, pp. 385 and 388.Google Scholar

^{page 129 note 2} ibid. fig. 26, p. 385; Ilkow, , Die Nominalkomposita, pp. 144–5Google Scholar; and Keller, , Weapon Names pp. 137–44.Google Scholar

^{page 130 note 1} The Art, p. 30.

^{page 130 note 2} Engelhardt, , Denmark, p. 78Google Scholar; see also 56–7 and pls. II and 12.

^{page 130 note 3} Brown, Baldwin, The Arts III, 241 and 236 and pl. XXXIGoogle Scholar. Although it is true, as Wilson, (The Anglo-Saxons, p. 120Google Scholar) observes, that no complete Anglo-Saxon spear survives, because the wooden shafts, like the wood of the shields, rotted in the ground, Baldwin Brown explains that archaeologists simply measured the distance between the head and the ferule where they lay buried in the ground. Wilson indicates that the length of spears in graves at Chessel Downs and Petersfinger was similarly measured.

^{page 131 note 1} Brown, Baldwin, The Arts 111, 241.Google Scholar

^{page 132 note 1} The Bayeux Tapestry, ed. Stenton, pl. 40; see also Wilson, , The Anglo-Saxons, fig. 29, p. 124.Google Scholar

^{page 132 note 2} Die altenglischen Kenningar, p. 226; see also Ilkow, , Die Nominalkomposita, pp. 279–80Google Scholar, on OS megin and meginfard in the Heiland.

^{page 133 note 1} I am indebted to Professor Peter Clemoes for calling my attention to the human power implied in magen in this context.

^{page 133 note 2} See Ilkow, , Die Nominalkomposita, pp. 304–5, on OS modthraka.Google Scholar

^{page 133 note 3} Meaning, p.405.

^{page 136 note 1} See above, pp. 80–1.

^{page 136 note 2} See above, p. 81.

^{page 138 note 1} Carr, , Nominal Compounds, pp. 344–54.Google Scholar

^{page 141 note 1} I am deeply grateful to Professor Peter Clemoes, who generously gave his encouragement, suggestions and time; Professors Stanley B. Greenfield, Fred C. Robinson and Margaret Schlauch, who read an earlier draft and offered valuable criticism; Dr Bruce Mitchell, who read the Hunlafing section and made helpful suggestions; and the librarians of the University of California, Irvine, who by their dedication and pertinacity made possible the completion of my research for this article.