^{Note 1 in page 1} C. Lottner, Revue Celtique, i (1870), 55–57, an appendix to W. M. Hennessy's article, “The Ancient Irish Goddesses of War,” ibid., pp. 32–55. The traditional term “wargoddesses” is retained here as the name of a class of beings appearing in Irish literature whose nature the following remarks will perhaps help to clarify. “War-witches” or “wardemons” would be equally appropriate names, but there can be no objection to the use of the traditional term as long as it is understood that nothing regarding the nature of the beings so named is implied.

^{Note 2 in page 1} The material has been most completely collected by G. S. Lane, “The Germano-Celtic Vocabulary,” Language, ix (1933), 244–264. The cultural significance of the vocabulary has been recently discussed by C. S. Elston in The Earliest Relations between Celts and Germans (London, 1934). Elston admits mutual influence but denies “historical justification for the assumption that the Celts ever dominated the Germans” (p. 186). Most students of the subject, however, are inclined to believe that the evidence points to some sort of Celtic supremacy and even J. Pokorny, who is not an adherent of the extreme Celtic school, accuses Elston of mishandling the evidence in a severe review, ZfcPh, xx (1936), 526–527. cf. also M. L. Sjoestedt's review, Revue Celtique, li (1934), 305–309. For other points of view on the question v. H. d'Arbois de Jubainville, Les Premiers Habitants de l'Europe (Paris, 1889), ii, 336–339; O. Bremer, “Ethnographie der germanischen Stämme,” in Pauls Grundriss ² (Strassburg, 1900), iii, 797–789; R. Much, Deutsche Stammeskunde (Berlin and Leipzig, 1928), pp. 197–202.

^{Note 3 in page 2} Viking influence on the Irish was first studied by H. Zimmer in “Keltische Beiträge,” ZfdA, xxxii (1888), 195–334; xxxiii (1889), 129–220; xxxv (1891), 1–172. A pioneer in the field, Zimmer, it is now generally recognized, overstated his case. K. Meyer, “Nordisch-Irisches,” Berliner Sitzungsberichte (1918), pp. 1030, 1042–417, while he accepts Zimmer's chief conclusion that there was extensive Viking influence in Ireland, warns against his theory of the Scandinavian origin of the Finn cycle and his interpretation of Ferdiad. A critical treatment of the Scandinavian loan-words in Irish is offered by C. J. S. Marstrander, Bidrag til del norske Sprogs Historie i Irland (Kristiania, 1915). For a recent study of reminiscences of the Vikings among Celtic peoples v. R. Th. Christiansen, The Vikings and the Viking Wars in Irish and Gaelic Tradition (Oslo, 1931). S. Bugge initiated the study of Irish influence on the Scandinavians. v. his Studier over de nordiske Gude- og Heltesagns Oprindelse, Første Række (Christiania, 1881–1889); Bidrag til den Ældste Skaldedigtnings Historie (Christiania, 1894); Helge-Digtene (Kjøbenhavn, 1896); Norsk Sagaskrivning og Sagafortælling i Irland (Kristiania, 1908). S. Bugge's work was continued by A. Bugge; v. e.g. his Vesterlandenes Indflydelse paa Nordboernes og sœrlig Nordmœmdenes ydre Kultur, Levesœt, og Samfundsforhold i Vikingetiden (Christiania, 1905); “Entstehung und Glaubwürdigkeit der isländischen Saga,” ZfdA, li (1909), 25–38; Norges Historie, i², (Kristiania, 1910), pp. 298–303. In a thorough study, “Tors Färd till Utgård,” Danske Studier (1910),pp. 65–105, 145–182, C. W. von Sydow has demonstrated the Irish origin of an Icelandic myth. v. also the recent studies of A. H. Krappe, “L'Origine irlandaise d'un Episode de la Halfs Saga,” Revue Celtique, xlvii (1930), 401–405, and of F. Mossé, “Sur le Nom d'Homme ‘Ketffl’,” Revue Celtique, l (1933), 248–253.

^{Note 4 in page 2} A. Schullerus, “Zur Kritik des altnordischen Valhollglaubens,” PBB, xii, 221–282.

^{Note 5 in page 2} G. Neckel, Walhall, Studien über germanischen Jenseitsglauben (Dortmund, 1913).

^{Note 6 in page 2} A. H. Krappe, “The Valkyries,” MLR, xxi (1926), 55–73.

^{Note 7 in page 2} W. Golther “Studien zur germanischen Sagengeschichte. i Der Valkyrjenmythus,” Münchener Abhandlungen, philosophisch-philologische Classe, xviii (1890), 399–438.

^{Note 8 in page 2} Op. cit., p. 402. cf. the insistence of J. de Vries on a sharp distinction between Scandinavian and southern material, Altgermanische Religionsgeschichte, i (Berlin and Leipzig, 1935), Vorwort.

^{Note 9 in page 2} Op. cit., p. 423.

^{Note 10 in page 3} Op. cit., p. 404.

^{Note 11 in page 3} Altgermanische Religionsgeschichte, ii (Berlin and Leipzig, 1937), p. 384. J. Grimm, Deutsche Mythologie iv (Berlin, 1876), p. 354, noticed the similarity between the and the valkyries. The Chadwicks, The Growth of Literature (Cambridge, 1932), pp. 207, 641–642, were also struck by the resemblance between the war-goddesses and the valkyries.

^{Note 12 in page 3} R. Much, “Der germanische Himmelsgott,” Festgabe Heinzel (Halle, 1898), p. 247; cf. also Deutsche Stammeskunde, p. 45.

^{Note 13 in page 3} A similar method has been used by W. Krause in his study, Die Kenning als typische Stilfigur der germaniscken und keltischen Dichtersprache (Halle, 1930). Krause shows that the kenning as we find it in Germanic is to be found elsewhere only in Celtic. True kennings are found in early runic inscriptions where there can be no question of Irish influence. Consequently the kenning in Celtic and Germanic must have originated when the two peoples were in contact on the continent.

^{Note 14 in page 3} For a collection of the Old English material v. R. Jente, “Die mythologischen Ausdrücke im altenglischen Wortschatz,” Anglistische Forschungen, lvi (1921), p. 209.

^{Note 15 in page 3} T. Wright, Anglo-Saxon and Old English Vocabularies. Second edition by R. P. Wülcker (London, 1884), 19.44, 25.27, 417.12.

^{Note 16 in page 3} Idem. 50.40, 189.11.

^{Note 17 in page 3} Idem, 347.32, 533.26.

^{Note 18 in page 3} Idem, 360.3, 527.17.

^{Note 19 in page 3} Idem, 189.12, and note. Wülcker, ibid., rejects the suggestion.

^{Note 20 in page 3} Neckel, op. cit., pp. 5–6.

^{Note 21 in page 4} v. Neckel, op. cit., p. 79. One might add that þórrn Hornklofi in his Haraldskvœoi presents a valkyrie in conversation with a raven. For the poem v. F. Jonsson, Den norskislandske Skjaldedigtning (København og Kristinia, 1912–1915), IB, 22–25.

^{Note 22 in page 4} Exodus, 1.164, ed. G. P. Krapp, The Junius Manuscript (New York, 1931). Cf. Neckel, op. cit., p. 78; Golther, op. cit., p. 415.

^{Note 23 in page 4} A. S. Napier, Wulfstan (Berlin, 1883), pp. 165, note, and 298.

^{Note 24 in page 4} Idem., p. 298.

^{Note 25 in page 4} Edited by J. Jonsson, Skjaldedigtning, IB, 389–391.

^{Note 26 in page 4} W. Braune, Althochdeutsches Lesebuch (Halle ^a/S., 1875), p. 83.

^{Note 27 in page 4} Grimnismál, 36.5, ed. G. Neckel in Edda (Heidelberg, 1927).

^{Note 28 in page 4} F. Jonsson, Lexicon Poeticum (København, 1931), s.v.

^{Note 29 in page 4} Passages illustrating this meaning are cited by K. Maurer in “Die Valkyrjen Hlökk und Herfjötr,” ZfdM, ii (1885), 341–343. v. also G. Neckel, op. cit., p. 75.

^{Note 30 in page 4} Jacob Grimm first connected the name Herfj$ootur with the Merseburg Charm, v. his Deutsche Mythologie iv, p. 352. The valkyrie-name Hl$ookk Grimm translates catena. He is followed by Neckel, loc. cit.; Jonsson, Lexicon, s.v. however, gives another meaning for Hl$ookk. Herfj$ootur alone is enough to connect the idisi of the Charm with the valkyries. Cf. also R. Kögel's attempt to ascertain the linguistic relation between Norse dís and German idis, “Idis und Walküre,” PBB, xvi (1892), 502–508.

^{Note 31 in page 5} The English may have had similar stories. The fettering of warriors is one of the tricks attributed to the devil in the Old English Solomon and Saturn: hwilum he [folme] gefeterao / faeges mannes, / handa gehefegao, / oonne ne æt hilde sceall / wio lao werud / lifes tiligan. (ll. 158–160, ed. Wülker, Bibliothek, iii² (Leipzig, 1898).

^{Note 32 in page 5} Eiríksmál, i. 9, 10, ed. Jonsson, Skjaldedigtning, IB, 164.

^{Note 33 in page 5} Cf. H. M. Chadwick, The Heroic Age (Cambridge, 1912), p. 412.

^{Note 34 in page 5} E. Mogk in J. Hoops, Reallexicon der germanischen Altertumskunde (Strassburg, 1911–1913), s.v. Alp.

^{Note 35 in page 5} Op. cit., p. 77.

^{Note 36 in page 5} R. Thurneysen, Die irische Helden- und Königsage (Halle ^a/S., 1921), pp. 96, 110–113. I suppose by Grundtext (p. 96) Thurneysen means the text found in the lost Urhandschrift (p. 111) which he places in the first half of the eighth century (ibid.).

^{Note 37 in page 5} The text, according to Thurneysen, “Zur Tain Bo Cuailnge,” ZfcPh, x (1915), 208, reads as follows: Cein bater di(du) intsloig octochim maige breg forrumai allechtu colleic nochisi inmorrigansom indeilb euin comboi forsin chorthi hitemair cualngi 7 asbert frisintarb. Cf. J. Strachan and J. G. Keefe, The Tain Bo Cuailnge from the Yellow Book of Lecan (Dublin, 1912), p. 32.

^{Note 38 in page 5} Thurneysen conjectures (“Zur Tain,” p. 208) that he had in mind Vergil's description of Allecto (Æneid, vii, 323 ff.). The Old English glossator explained Allerto as a valkyrie, cf. supra, note 18.

^{Note 39 in page 5} In the Tain Bo Regamain, ed. E. Windisch, Irische Texte, ii (Leipzig, 1887), 245, the Morrigu transforms herself bto a black bird. H. Hubert, basing his opinion on traditions reported in Livy, believed that the Celts in Italy had stories of divinities who appeared on battlefields in the form of crows, v. his Les Celtes depuis l'époque de la Tène (Paris, 1932), pp. 37–38.

^{Note 40 in page 6} On these “rhetorics” v. Thurneysen, Königsage, pp. 54–55. They are probably the oldest elements in the Tain. v. idem., pp. 110–112. All versions of this prophecy together with some suggestions as to the interpretation are printed by Windisch, Tain Bo Cualnge (Leipzig, 1905), pp. 186–189.

^{Note 41 in page 6} The texts in Windisch, Tain Bo Cualnge, pp. 828–833. Although it must have been introduced early, this “rhetoric” may not have belonged originally to the Tain Bo Cualnge. v. Thurneysen, Konigsage, p. 209, note 2.

^{Note 42 in page 6} Windisch, Tain Bo Cualnge, pp. 838–841.

^{Note 43 in page 6} W. Stokes and J. Strachan, Thesaurus Palaeohibernicus (Cambridge, 1901), i, 2.

^{Note 44 in page 6} Ed. by G. Calder (Cambridge, 1922), 1. 88.

^{Note 45 in page 6} Lecan version, ed. Strachan and O'Keeffe, 1. 1748.

^{Note 46 in page 6} W. Stokes, Revue Celtique, xii (1891), 128; F. Holthausen, Altenglisches etymologisches Wörterbuch (Heidelberg, 1934), s.v. mare; A. Walde, Vergeichendes Wörterbuch . . . herausgegeben von J. Pokorny (Berlin and Leipzig, 1930–1932), ii, 277.

^{Note 47 in page 6} Thus war-goddess and valkyries had common differences from the who brought all kinds of evils—filth, sickness, and blindness as well as death in battle. They seem to have been connected with death in battle only in so far as they were connected with all kinds of deaths, v. Malten in Pauly-Wissova, Suppl. iv, 883–900.

^{Note 48 in page 7} The assumptions defended here, that the close connection of Celts and Germanic peoples led to the common development of material inherited, originally, perhaps, from the Indo-Europeans by both peoples is borne out by the fact that a similar process seems to have taken place in the case of a large number of words which appear in both languages in forms indicating that they were, in each language, regularly developed from Indo-European forms, but which, nevertheless, show similarities in meaning and use, as opposed to their cognates in other Indo-European languages, e.g. cf. Irish orbe, Gothic arbi, inheritance, with Greek orphaned; Irish gae, Gallo-Latin gaesum, Old English gar, spear, with Greek os, shepherd's staff; Irish run, secret, Old English run, secret, whisper, rune, with Sanskrit rauti, howls, cf. Lane, op. cit. et infra, note, 51. A similar assumption is made by de Vries, op. cit., i, 195–196, concerning the matres in Celtic and Germanic. There is no evidence that the valkyries are the result of ideas developed exclusively by the Celts and borrowed by the Germanic peoples at a comparatively late period. Some believe, however, that such was the case with the matres. cf. K. Helm, Altgermanische Religionsgeschichte, i (Heidelberg, 1913), pp. 391–411.

^{Note 49 in page 7} v. Lane, op. cit., p. 246; Walde-Pokorny, op. cit., ii, 126. Old English beadu and Old Icelandic b$ooo, battle, are cognate. There are no words certainly cognate outside of Celtic and Germanic. There are no means of telling whether the word was borrowed from one people by the other before the Germanic sound-shift or whether it was inherited from Indo-European by both independently.

^{Note 50 in page 7} Tacitus, Annates, iv, 73.

^{Note 51 in page 7} The suggestion of Th. von Grienberger, PBB, xix (1894), 532–533, that the word is a Roman graph for Germanic *badu-wenna, “the violent in battle,” the second element of which is to be connected with Old High German winna, strife, is accepted by M. Schönfeld, Wörterbuch der altgermanischen Personen- und Völkernamen (Heidelberg, 1911), s.v., and tentatively by R. Much, Himmelsgolt, p. 274. K. Müllenhoff, ZfdA, ix (1853), 240–241, likewise had considered the h due to a Roman scribe but held that the name was uncompounded and the -enna a suffixal element. T. Siebs's suggestion, ZfdPh, xxiv (1892), 147, that Baduhenna was a god, the second element of whose name is to be connected with the Germanic root *hen-, *han-, which appears also in the inscription to Mercurius Channo, has apparently not found any supporters, cf. S. Gutenbrunner, “Die germanischen Götternamen der antiken Inschriften,” Rheinische Beiträge und Hülfsbücher zur germanischen Philologie und Volkskunde, xxiv (Halle, 1936), p. 57. H. Jaekel's, ZfdPh, xxii (1890), 268, equation of Baduhenna and the Bede of the Mars Thingsus inscription is wholly untenable on linguistic grounds. O. Bremer, ZfdPh, xxii(1890), 252, suggests, because of the ending, that the name is Celtic. If this is true, it must have been borrowed before the Germanic change of o to a. The existence of the common noun Old English beadu, battle, in Germanic makes it more likely that the goddess was named with a native word. There were, however, bilingual persons on the Celto-Germanic border (cf. Much, Stammeskunde, p. 48), and it is even possible, as Marstrander suggests, Symbolae Osloenses, iii (1925), 62, that Celtic and Germanic were mutually comprehensible as late as the first century B.C. Celtic *bodu- and Germanic *badu- would be recognized as forms of the same word. The fact that the Celts used their form in the name of a goddess might have led to a similar use of the Germanic form, or the reverse may have taken place. This process is distinct from linguistic borrowing.

^{Note 52 in page 8} v. Hennessy, op. cit., pp. 32–33; A. Holder, Altcellischer Sprachschatz (Leipzig, 1896–1907), i, 841.

^{Note 53 in page 8} e.g., Hariasa and Harimella, the first element of whose names is almost certainly to be connected with Germanic *hari-, Old English here, army. v. de Vries, op. cit., i, 205; Gutenbrunner, op. cit., pp. 100–102. Although the connection of the name Vihansa with Old Icelandic víg, battle, is, because of the voiceless spirant where, according to Verner's law, a voiced spirant is to be expected, not certain, the dedication of spear and shield seems in any case to connect her with war. v. de Vries, op. cit., i, 209; Gutenbrunner, op. cit., pp. 101–102; Much, Bimmelsgott, p. 247.

^{Note 54 in page 8} T. Siebs, “Beiträge zur deutschen Mythologie, ii, Things und die Alaisiagen,” ZfdPh xxiv (1892), 433–456; “Neues zur germanischen Mythologie, ”Mitteilungen der schlesischen Gesellschaft für Volkskunde, xxv (1924), 1–17; Krappe, loc. cit. supra note 6.

^{Note 55 in page 8} These are the most recent readings, those of Gutenbrunner, op. cit., p. 25, on the evidence regarding the dating of the inscriptions (third century a.d.) and the dedicants (cives Tuihanti cunei Frisiorum), given most fully in the second inscription v. de Vries, op. cit., i, 171. cf. Gutenbrunner, op. cit., p. 48.

^{Note 56 in page 9} Before the publication of the third inscription in 1921 (v. de Vries, op. cit. I, 201), R. Heinzel, Wiener Sitzungsberichte, philosophisch-historische Classe, cxix (1889), Abhandlung iii, p. 52, drew from the second inscription the not unnatural conclusion that there were only two alaisiagae. Because of the third inscription, Siebs, Neues, p. 11, concluded that there were many alaisiagae but that they appeared in pairs. Gutenbrunner, op. cit., pp. 41–43, and de Vries, op. cit., i, 202, retain Heinzel's opinion that there was only one pair of alaisiagae and believe that each goddess had more than one name. Gutenbrunner even composes some Gothic verses to support this opinion.

^{Note 57 in page 9} The most striking case is the opening of the Hákonarmál. v. Jonsson, op. cit., IB, 57.5 Additional examples are cited by Siebs, Neues, pp. 14–15; Krappe, Valkyries, p. 57; cf. Gutenbrunner, op. cit., p. 42, pp. 113–114.

^{Note 58 in page 9} Siebs, Neues, pp. 12–13, accepted by Gutenbrunner, op. cit., p. 43. The explanation of the first element, a common one in Germanic names, is difficult. Siebs, loc cit., connects it with the Germanic roto *beud-, to offer, and translates “die Kampfbietende.” This explanation assumes a position of nominal and verbal element wholly unusual in Germanic compounds. Gutennbrunner, op. cit., p. 43, considers the word cognate with Irish buaid, victory. This involves the assumption that the d of the Celtic and Germanic words comes from Indo-European dh and means, therefore, that the connection of Irish buaid with Low German bute, booty, and its cognates, which implies an Indo-European d, must be abandoned. It is hard to see why buaid should be torn from connections which it fits in both form and meaning in order to explain a word of whose meaning nothing is known.

^{Note 59 in page 9} Siebs, who is the principal defender of the theory that the alaisiagae were valkyries, connects alaisiagae with Old Icelandic eisa, to rage, storm, and translates “those who storm furiously on,” (Thingsus, p. 442). He admits in the later article, Neues, pp. 7–8, that he is doubtful about the interpretation of the names Beda and Fimmilena, which he had previously (Thingsus, pp. 447–449) tried to connect with words suggesting the violence of the wind. Friagabi he translates “the freedom-giver” and recalls the third group of idisi in the Merseburg Charm, those who broke the chains (Neues, p. 14). He connects the word Thingsus with Gothic peihs, time, and considers it to have been originally a title of *Tiwaz as god of the sky (Thingsus, p. 450; Neues, p. 7). With him the valkyries, originally stormgoddesses, were connected.

It is generally agreed that the Mars of the inscription is Germanic *Tiwaz, but there is disagreement as to the meaning of Thingsus. W. van Helten, PBB, xxvii (1902), 137–153, sees in the word a roman graph for a Germanic noun which he believes means “warrior.” Most investigators are agreed, however, that Thingsus has some connection with the Germanic thing, the popular assembly. Kauffmann, PBB, xvi (1892), pp. 208–209, suggests that Germanic *þing- may have been the translation for Latin cuneus, since the tribes composing the Roman cavalry unit were, in their civil capacity, gathered in the thing. Mars Thingsus, then, was the guardian genius of the cuneus Frisiorum mentioned in the second inscription. Th. von Grienberger, Zeitschrift für die öslerreichischen Gymnasien, xlvii (1896), 1007, holds that Thingsus is not a Germanic word at all but a Latin formation from Germanic *things. De Vries, op. cit., i, 172–173, is inclined to accept the suggestion of W. Scherer, Berliner Sitzungsberichte (1884), p. 574, that the word reflects a Germanic title formed from *things and indicates that the Mars-*Tiwaz of the inscription was god of the thing. De Vries, op. cit., i, 174–175, adds that, in the Germanic mind, war and the thing were intimately connected and that the same god might well be connected with both. Heinzel, op. cit., p. 52; Much, Himmelsgott, p. 195 et passim; and T. E. Karsten, Acta Societatis Scientiarum Fennicae, xlv (1915), no. 2, p. 15, connect Thingsus with Gothic þeihjo, thunder, and consider it to have been originally a title of *Tiwaz as god of the sky and the thunder. None of these investigators considers such an origin of the term incompatible with the theory that Mars Thingsus was at the time of the inscription god of the thing, although Much and Karsten differ in their explanations of the development from sky-god to thing-god. Gutenbrunner, op. cit., p. 30, while insisting that the Mars Thingsus of the inscription was god of the thing, is willing to accept Much's explanation of the origin.

Explanations of the term alaisiagae have been, in part, determined by the theories held regarding Thingsus. Heinzel, op. cit., pp. 50–52 considers the word a kenning, alaisi-agae, “Schrecken der Erle,” which fits weather goddesses attendant on the sky-god. Weinhold, Zfd Ph, xxi (1896), 7, amends to alaisagiae, “die grosse Gesetzsprecherinnen.” Grienberger, op. cit., p. 1007, sees in alaisiagae a place-name with the Celto-Latin suffix -iacus. Henning. ZfdA, xlii (1898), 193–194, recalls Old English ae- and Gothic leiseigs and translates “ad edocendum idoneae.” All these explanations are open to serious objections, and the most widely held theory is the one adumbrated by Scherer, op. cit., pp. 579–580, developed by Kauffmann, op. cit., pp. 201–204, and perfected by Sievers, PBB, xvi (1892), 259. This theory connects alaisiagae with the Germanic root *aiz- found in Old English arian, to honor, and translates “the all-honored ones.” It has been accepted by de Vries, op. cit., i, 202, and Gutenbrunner, op. cit., p. 47, who adds that the ending may be connected with Old Icelandic -gen in such words as systken, brother and sister.

The names Bede and Fimmilene have offered the greatest difficulties. Earlier investigators were inclined to see a connection between the names of the goddesses and the expressions bodthing and fimelthing found in the late medieval Frisian laws. Heinzel, op. cit., p. 52, who interprets the names with reference to the supposed original character of the alaisiagae as weather-goddesses, implies that the legal terms are derived from the names of the goddesses. Weinhold, op. cit., p. 14, suggests that the goddesses were originally personifications of legal processes whose names have been preserved in bodthing and fimelthing. Such theories are attacked on legal and historical grounds by Kauffmann, op. cit., p.203, and by Heck in an addition to Siebs's earlier article. The publication of the third inscription with its clear evidence on warlike alaisiagae might be expected to have given the coup de grace to all theories connecting the alaisiagae with the Frisian things, and de Vries, op. cit., i, 202, does abandon the connection. Gutenbrunner, however, has revived the old theory in a more subtle form. The goddesses, he holds (op. cit., p. 32) were not personifications of the things, nor were the things named after the goddesses, but the names of both goddesses and things were derived from a pair of old legal expressions which together covered the entire juridical process, *bed-, the invitation, the calling to the thing, and *fimil-, the conclusion, the judgment (op. cit., p. 44). Even though Gutenbrunner's explanation is granted to be the most likely hitherto offered, however, there is still reason to assume that the alaisiagae were originally valkyries who, as companions of *Tiwaz, god both of war and of the thing, had acquired a specialized function and legal names. Even Gutenbrunner, op. cit., p. 45, admits that the alaisiagae of the third inscription were associated with *Tiwaz as god of war. If it is assumed that this was the original character of the alaisiagae, then there is a source for them in material already existing in the Germanic imagination. If one insists on the juridical character as primary, however, and at the same time denies, as Gutenbrunner does, that the goddesses were personifications of the bodthing and fimelthing, there is no such source. The names may have come from the legal sphere, but what is the thing that was named? Siebs's theory, therefore, at least in its insistence that the alaisiagae were connected primarily with war, still seems to be the most plausible.

^{Note 60 in page 11} The statement occurs in a poem on the deaths of the Tuatha De Danann in the Book of Leinster, facsimile edition (Dublin, 1880), p. 11b. Flann was a late poet who, according to O'Curry, On the Manners and Customs of the Ancient Irish (London, 1873), ii, 149, died in 1056. The possibility of Viking influence must, therefore, be considered here. Since two of the names, Neit and Bodb, are unquestionably old and Gaelic, however, it seems reasonably certain that Flann is here reporting an old Celtic tradition. Windisch, Tain, p. 338, note 8, calls attention to the passage.

^{Note 61 in page 11} cf. supra note 59 ad init. This was formerly the most common explanation of the valkyries. v. Mogk, Literaturblatt für germanische und romanische Philologie, vii (1887), 291–292.

^{Note 62 in page 11} Valkyries, p. 70.

^{Note 63 in page 11} Cf. supra notes 7, 5, and 11.

^{Note 64 in page 11} Since the first element in mor-rigu is probably an Indo-European word (cf. supra, note 46) and, except in the name, is not found in Irish, it follows that the naming of the goddess must have taken place very early and probably on the continent. The name Morrigu proves the demonic and witch-like character of the continental ancestresses of the Irish war-goddesses. The Cathubodua inscription proves their divine character. But the close association of Morrigu and Bodb, from which it is here inferred that the continental Celts knew of a class of beings who were at once witches, demons, and goddesses, is, of course, demonstrable only from the late Irish material where the two figures are clearly portrayed as belonging to the same class of beings, called bodbai after the goddess Bodb. It is assumed here that this close and constant association of the two figures in the Irish material must have its roots far back in the Celtic past.