CORNELL UNIVERSITY LIBRARY GIFT OF James Morgan Hart DATE DUE ■?0 "''« i''i CgB=r^- dbt! — 6- gVS~-ag jf^.^ ''•^''jjS0^ m^ F£a^ msi-x- IVWH--J U9% p '■ CAVLORO PRINTED IN U.S.A. PN 685.KI2"" ""'"'■*">' '■'"'^n' IIMmillliii?iIilii.,,i?°'''aflon / The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924027098072 ARTHUR AND GORLAGON BY GEORGE LYMAN KITTREDGE / Reprinted from Studies aNd Notes in Philology and Literature Vol. VUI BOSTON, U.S.A. GINN & COMPANY, PUBLISHERS 1903 ARTHUR AND GORLAGON BY GEORGE LYMAN KITTREDGE Reprinted from Studies and Notes in Philology and Literature' Vol. VIII BOSTON, U.S.A. ^^ \' %^ GINN & COMPANY, PUBLISHERS, ""'"x^ \ 1903 *i;i^i ^, I ARTHUR AND GORLAGON. THE following text, which is here edited for the first time and seems to have eluded all investigators of Arthurian tradition,^ is contained in Rawlinson MS. B. 149 (parchment) in the Bodleian Library. The manuscript is of the end of the fourteenth century," and its contents, as catalogued by Macray,' are as follows : 1 . Historia trium Magorum. 2. Narratio de Arthuro Rege Britanniae et Rege Gorlagon lycanthropo. 3. De " Tirio Appolonio " narratio. 4. Historia Meriadoci Regis Cambriae.* 5. " Liber Alexandri Philippi Macedonum qui primus regnavit in Grecia et de preliis ejusdem." 6. Tractatus, Aristotelis dictus, de regimine sanitatis, libris decern. Arthur and Gorlagon occupies pp. 55-64, and has no title. It is written in two hands, the second hand beginning with seminecem in the last line of p. 60. I have expanded the numerous contractions of the manuscript, have regulated punctuation, capitals, and the separation of words, have divided the tale into paragraphs, and have numbered the sections. All other changes are indicated in the notes or by brackets in the text. There is no clue to the authorship of Arthur and Gorlagon ; but it was not written by the author of the Vita Meriadoci and the De 1 It is possible that this was one of the " five Latin romances " known to Sir Frederic Madden {Syr Gawayne, Introd., p. x, note). 2 Or the beginning of the fifteenth (Meyer, Alexandre le Grand, II, 392). ' Catal. Codd. MSS. Bibliothecae Bodleianae, p. v, fasc. i, cols. 500-501 (1862). * Edited by Bruce (from the Cotton MS., Faustina B. vi) in PuM. of the Mod. Lang. Association of America, XV, 326 ff . (1900). The copy of Meriadocus in the Rawlinson MS. escaped Professor Bruce's notice. 149 150 G. L. Kittredge. Ortu Waluuanii} The style is enough to make that point certain, and the whole character of the tale differs widely from those long- winded romances. The Rawlinson copy is pretty accurate ; but it shows a number of errors and at least one omission. These blunders are enough to prove that it is not the author's autograph, even if this were not immediately clear from the fact that it is the work of two diiferent scribes. A' [ARTHUR AND GORLAGON.] PUD Vrbem Legionum^ celebre festum diei Pentecostes rex Arturus agebat, ad quod totius sue dicionis magnates et nobiles inuitabat, peractisque de more solemnijs, ad instructum ^ cum omnibus perti- nentibus conuiuium. Quibus copla afHuente dapum summa cum leticia prandentibus, Arturus, in nimiam effusus leticiam, regiham sibi considentem iniectis brachijs amplexatus est, amplexusque cunctis intuentibus strictissime osculatus est. Ad hec autem ilia obstupef acta simulque rubore suffusa, ipsum respexit, et cur se loco et hora insolita osculatus fuisset quesiuit. Arturus. Quia nichil mihi in diuicijs gratius, nil in delicijs te constat suauius. Regina. Si quam asseris me adeo diligas, mentem et voluntatem meam te scire patenter existimas. Arturus. Tuam mentem erga me beneuolam habere non dubito, tuam- que voluntatem mihi prorsus patere certus existo. Regina. Arture, falleris sine dubio ; quippe agnoscas te nunquam uel ingenium mentemue femine comperisse. Arturus. Omnia cell obtestor numina, si me actenus latuere, dabo operam, nee labori indulgens nunquam cibo fruar donee ea me nosse contingat. ■i. Finito itaque conuiuio, Caium dapiferum suum Arturus aduocat atque " Kai," ait, " tu et Walwainus nepos mens ascendite et ad negotium quo propero me[c]um venite. Ceteri omnes remaneant, meos conuiuas mei loco usquedum rediero letificate." Nee mora, iussi equos ascendunt et ad regem 1 Edited by Bruce (from the same manuscript that contains the Meriadocus) in Publ. Mod. Lang. Assoc, XIII, 365 ff. (1898). Professor Bruce (XIII, 388-389; XV, 338-339) refers the Meriadocus and the De Ortu Waluuanii to the second quarter of the thirteenth century (the Cotton MS. is of the early fourteenth) and ascribes them to a single author. 2 legione MS. ' instrictu MS. Arthur and Gorlagon. 1 5 1 quendam sapientissimum in confinio regnantem, Gargol ^ dictum, cum Arturo ipsi duo tantummodo properantes, die tercia in quandam vallem lassi deuene- runt, — postquam enim a domo discesserant nee cibum nee sompnum cepe- rant, sed noctes diebus continuantes semper equitauerant. E regione autem aduersa ipsius vallis mons arduus extabat, ameno nemore eonstitutus, in cuius recessu fortissimum ex politis lapidibus eminebat eastellum. Quod vbi Arturus eminus intuitus est, Caium cursim preeedere imperat et cuius esset illud opidum renunciare festinet. Citato igitur sonipede, Caius accelerauit, intrauit, et iam vallum exterius subeunti in redeundo Arturo occurrens, regis Gorgol ad quem tendebant municipium fore renunciauit. 3. Fortuitu autem rex Gorgol tunc mense pransurus consederat; ante quem Arturus equo vectus ingressus eum lepide cum conuiuantibus saluta- bat. Cui rex Gorgol " Quis es," ait, " et vnde, et que causa te tarn precipi- temnostroingessit conspectui?" Arturus. " Arturus sum " respondit, "rex" Britannie ; artem et ingenium mentemque femineam a te discere desidero, quem in rebus huiusmodi peritum sepissime expertus sum." Gorgol. Arture, magnum est quod queris, et perpauci sunt qui illud nouerunt; sed crede nunc consilio meo, deseende et comede et hodie qui- esce, quia itinere et labore te vexatum video, et eras quod inde sciero indi- cabo tibi. Negauit Arturus, se nunquam comessurum eonstip[u]lans nisi prius quod querebat didicisset. Tandem turn ^ rege et conuiuantibus socijsque instanti- bus annuit et descendit sedeque locata ante regem discubuit. Primo autem dilueulo * Arturus, pacti non immemor, regem Gargol adijt atque " O mi rex," ait, " insinua quod te mihi hodie dicturum heri spopondisti." Gor- gol. Arture, stulticlam ventilas ; sapientem te actenus reputabam. Ars ingenium et mens femine nuUius vnquam patuere noticie, nee [p. 56] ego te scio quidquam docere. Sed est mihi frater, rex Torleil ^ dictus, vicinitate regni coniunctus, me senior et sapientior, quem latere profecto non ^ autumo, si aliquis huius rei peritus habetur, quam adeo scire affectas ; hunc pete et ut tibi indicet quod inde nouerit mea ex parte edieito. 4. Regi igitur Gorgal valedicto, Arturus discessit, iter arripuit, et quatri- duano confecto itinere ad regem Torbeil peruenit, ipsumque casu pranden- tem inuenit. A quo resalutatus et quis esset inquisitus, se Arturum regem 1 Gargolu (u nearly erased) MS. On the variations in the names of the char- acters, see pp. 201, 203. ' r«/ Then Eochaid sent his messengers throughout Erin to seek for him the most beautiful woman among the maidens of Erin. Also he declared that he would marry no wovian whom any 07ie of the vien of Erin had known, before him? The resemblance between Melion's vow and Eochaid's needs no emphasis. The one might almost be a translation of the other. But the parallel does not stop here. One day, while hunting in the woods, Melion sees a maiden riding towards him. She is richly dressed and of surpassing beauty. He salutes her and asks her of what kindred she is and what brings her thither ; Dites inoi dont vos estes nee Et que ici vos a menee (vv. 103-104). Compare the Tochmarc Elaine : King Eochaid's messengers traverse all Erin until they learn of a maiden, who is a fitting match for him.^ They return to Tara with their report, ' Of course the king was disgraced by this refusal, as Melion was by the ladies' sending him to Coventry. ^ The original may be added on account of the significance of the passage : •' Al asbert, ni biad ina farrad acht ben nad fesser nech do feraib hErend riam " (Lebor na h-Uidre, Windisch, Irische Texte, I, 119). The version in Egerton MS. 1782 (Windisch, ibid^ has the same requirements, but includes also the proviso that the woman shall be Eochaid's equal "in form and beauty and family." The version prefixed to the Togail Bruidne Da. Derga (Stokes, Revue Celtique, XXII, 13 £f.) omits Eochaid's feast and his vow, and begins with his meeting with Etain at Brig Leith. A part of the Tochmarc Etaine is translated by Thurneysen, Sagen aus dem alien Irland, pp. 77 f£. 3 What follows is not in the Lebor na h-Uidre, which concludes the first epi- sode of the Tochmarc Etaine with the bald statement (immediately following the king's declaration, just quoted) : " There was found for him [one], at Inbir Chichmaine, namely, Etain, daughter of Etar, and Eochaid took her home then, and she was a match for him in shape and form and family,'' etc. (Windisch, I, 119). It then proceeds directly to the love of Ailill for the queen. The details of Eochaid's meeting with Etain are preserved not only in Egerton MS. 17S2 (edited by Windisch, Irische Texte, I, 113 ff.), but in three manuscripts of The Destruction of Di Derga' s Hostel (Togail Bruidne D& Derga), including the Yello-M Book of Lecan (see the edition and translation of Stokes, Revue Celtique, XXII, Arthur and Gorlagon. 193 and the king sets out to win lier. He finds the damsel at Brig Leith, on the margin of a spring. Her beauty and the splendor of her attire are described in florid language. The king accosts her with these words : " Whence art thou, maiden, and whence comest thou ? " ' Let us return to M for a moment : Cele respont: "J el vos dirai, Que ja de mot ne mentirai. Je sui ass& de haut parage, Et nde de gentil lignage ; D'Yrlande sui a vos venue ; Sachids que je sui mout vo drue; Onques home fors vos n'amai, Ne jamais plus n'en amerai. Forment vos ai 01 loer; Onques ne voloie altre amer Fors vos tot seul, ne jamais jor Vers nul autre n'avrai amor '' (vv. 105 £E.). Melion takes the lady with him to his castle and marries her. Again we must compare the Wooing of Etain : " It is not hard [to reply to thy question]," the maiden answers. " I am Etain, daughter of the king of the horsemen, from the side [i.e. the fairy folk]." " Shall I lie with thee now ? " asks Eochaid. " For that have I come here into thy protection," says the maid. " It is twenty years since I was bom in the sid\\.e.. fairy hill], and men of the sid, both kings and fair men, a- wooing me, and no man of them has known me, because I have loved thee and set affection and desire upon thee since I was a child and capable of speech, on account of thy fame and thy glory ; and I have never seen thee before this time, and I recognized thee immediately by thy 9 ff.). Though Egerton 1 782 and the Dd Derga manuscripts are later than the Lebor na h-Uidre, there is no doubt that they afford us a very old version. According to Stokes, the Yellow Book "preserves some Old-Irish forms which have been modernised in the elder copy" (i.e. in the Lebor na h-Uidre). The Lebor na h-Uidre copy seems to have been condensed at this point. Compare the relation between the longer and the shorter version of The Wooing of Emer (Tochmarc Entire) : K. Meyer, Archaological Review, I, 68 ff., etc. ; id., Revue Celtique, XI, 442 ff. ; Hull, Cuchullin Saga, pp. ^5 ff. 1 " ' Can deit iarum a ingen,' ar Eochaid, ' ocus can dollot ? ' " (Irische Texte, I, 120, 1. 16); cf. Revue Celtique, XXII, 16-17. 194 G. L. Kittredge. description, and it is thou to whom I have come." The king promises . to forsake all other women and have her for his sole wife, and she goes to Tara with him, where she is warmly welcomed and the feast takes place.^ The parallels just given sufficiently justify the conjecture (p. 176) that X contained a passage corresponding to the introductory inci- dent in M and including the rash vow of the hero," his meeting with 1 Irische Texte, I, 120; Stokes, Revue Celtique, XXII, 16-17 (except the last sentence). " King Adier, in a curious little poem (midway between romance and ballad) found only in the Percy MS. (Hales and Furnivall, II, 296 ff.), declares: There were not that woman this day aliue, I kept [i.e. should care\ to bee my wedded wiffe, Without she [MS. the\ were as white as any milke Or as soft as any silke, And the(y) royall rich wine ran downe her brest bone, And lord! shee were and a leal [MS. leatK\ maiden (w. 5 if.). He is informed that King Estmere has such a paragon, and proceeds to win her, against heavy odds. The story, as Professor Child has noted (Ballads, II, 50), is that of Hugdietrich in the Heldenbuch (von der Hagen, 1855, 1, 169 ff. ; Amelung and Janicke, I, 167 ft.), and there is some relation between Kinge Adler (as the romance is called in the manuscript) and the superb ballad of King Estmere (Child, no. 60, II, 49 ff.). The impossible tasks in King Adler are like those which adventurers must undertake in mdrchen (and elsewhere) to win the daughters of supernatural beings. It seems to be the rule that mortals who make vows of this kind -mnfees, and sometimes have trouble with them. See, for example, Richard Coer de Lion, vv. 43 ff. (Weber, Metrical Romances, II, 5 ff.), with regard to the demon wife of Henry II (and cf. Child, Ballads, IV, 463). I take this opportunity of comparing Richard's eccentric method of killing the lion {Richard Coer de Lion, vv. 1063 ff., Weber, II, 43-44) with that followed by Cuchulinn in disposing of the sea-monster in the Fled Bricrend (§ 86, Irische Texte, I, 298; Henderson, pp. 106-107) and of Conall's hound Conbel in the Aided Guill ocus Aided Gairb, 40 (edited and translated from the Book of Leinster by Stokes, Revue Celtique, XIV, 422-3; cf. also The Pursuit after Diarmuid, ed. O'Grady, Ossianic Soc. Trans., Ill, 102-103). The incident, as well as the account of King Henry's demon wife, occurs in a part of the poem which is foreign to the Auchinleck text and which Paris thinks has no French source {Rom., IX, 54^ ff ■ XXVI, 356-7, note 3). Arthur and Gorlagon. 195 they?!?, and a conversation between them much like that preserved in M. Thus these parallels enable us to supply certain signifi- cant details in our reconstruction of x, the common source of M and y (GI). The resemblance between the general outline of x and that of the Tochmarc Etaine, as well as the particular correspondences which present themselves in so remarkable a manner, suggests the next step in our investigation and enables us to take it with a feeling of security. We are now in a position to understand the make-up of X, the source of all our versions except B. It was manifestly a complex tale. In its main outlines, it was a fairy mistress story of the type exemplified in ancient Irish literature by the Wooing of Stain. A fie abandons the Other World and marries a mortal. Her fairy lover or husband follows her and takes her back with him. Her mortal husband visits the Other World and recovers his wife. Into a story of this type has been worked an anecdote of an entirely different character, — The Werewolf's Tale proper. In this the hero was a born werewolf, forced by his very nature to spend a definite portion of his life in the shape of a wolf. His wife induced him to disclose his secret, and, with the help of her lover (or of a rejected suitor whom she promised to reward with her hand), forced the hero to retain his wolfish shape for a long time. At last, however, he took refuge with a certain king, who disenchanted him. The faithless wife was discarded, and her lover was punished. The result of combining these two stories has been to disguise somewhat the original plot of the former ; yet we can still recognize the character of that plot in two versions of the composite story, — the " Breton lay " of Melion (M), and the mdrchen I, still current through- out a large part of Ireland and well known, until recently, in the Scottish Highlands. Where did this amalgamation take place ? The almost inevitable answer is, — in Ireland. There, and nowhere else, the composite in question is still thoroughly at home and in active circulation as a folk-tale. There, too, we find the Tochmarc Etaine, with its startling correspondences to the Lai de Melion, preserved (in part) in a manu- script of about 1 1 00 (nearly a century earlier than Marie's time), and 196 G. L. Kittredge. antedating by two or three hundred years the manuscript that con- tains it.* We need not hesitate, then, in pronouncing for Irish as the language of x, and for Ireland as the country in which that version originated. This Irish x was not a mere floating folk-tale, in all prob- ability ; it was a pretty definite piece of literary work (oral or written), composed at a time considerably antecedent to the earliest French ver- sions of Arthurian romances. We should never forget that the Irish legends which we know from the Lebor na h-Uidre (c. iioo), and others of similar character, are simply the debris of a great literature, often betraying centuries of redaction by the form in which we find them at that early time. Fixation by literary means is a sufficient 1 The Tochmarc Etaine is not cited as one of the sources of our tale, but merely as an extant and very early example of a type of Irish saga to which that tale owes its outline (exclusive of the werewolf material) and 4o6. Loth (Chrestomathie, I, 171, note 2) speaks respectfully of Rhjs's theory, but in Revue Celtique, XV. 224, suggests an etymology of his own. He regards Breton -uualart, Welsh ■waladr (in Cat-uualart, Cat-waladr, etc.) as cognate with Old Norse Valfa&ir, a name of Odin, and as coming from a form *valii-{p)atir. This would make -uual cognate with the Germanic *walu- (O.H.G. wal, seen in Ger. wahlstatt ; A.S. ivcel ; O.N. valr, val-kyrjd). Stokes (Bezzenberger's Beitr., XXIII, 41) mentions Loth's equation of -uualart and Valfa&ir, but without committing him- self. It is impossible, however, to attach any weight to Loth's etymology. It is altogether improbable that the Scandinavian Valfa&ir is old enough to be cognate with the Celtic -waladr, and there are other difficulties. ' Fbrstemann, 2d ed., col. 799. 8 ij^ jqI 1453. Arthur and Gorlagoti. 203 Gurgjwl ^ith Waraidf} These correspondences ^ make out a strong case, in the lack of any other satisfactory explanation. If Rh^s's etymology of uual (jguaT) is correct, the name Gorgol in our Latin text not only means " werewolf " but is etymologically identical with that word.' If it is not correct, we are face to face with an amazing coincidence : more than twenty years ago, on linguistic grounds alone, Professor Rh^s equated Gurgual with werewolf, and now the name turns up as that of an actual werewolf (or his brother) in the Latin text which we are studying.* Let us turn to Gorlagon.^ This word occurs also in the prose Perceval, in the forms Gorgalan, Gurgalain, etc.,' as the name of a heathen king of " Albania." Its etymology is not beyond conjecture. 1 Id., col. 1537. 2 Rhys also equates the Welsh Budgual [Breton Butgual'\ with Botolf, but this is an error. Botolf, Badulf (Forstemann, col. 230) are from O.H.G. badu-, A.S. beadu, " battle," which is not cognate with Breton but-, bud-, Irish buaid, " victory." On Butgual, cf. Zimmer, Ztsch. f. franz. Sprache, XIII, 51. In sense, however, Butuual (-gual) may be compared with A.S. Sigewulf. ' On werewolf as " man-wolf," see the decisive remarks of Mogk and Napier (Paul and Branne's Beitrdge, XXI, 575-576; XXIII, 571 ff.) in reply to Kogel (Paul's Grundriss, ist ed., I, 1017, note). * No one, it is to be hoped, will maintain that Gorgol is a corruption of Garulf which Marie gives as the Norman equivalent of bisclavret (" Garulf [var. garwal'\ I'apelent li Normun "). This would doubtless be maintained by any one who wished to derive G from Marie's lay. Such a theorist, however, would have to account not only for G, but for M and I ; for it has been proved, beyond cavil, that G and I have a common source (y) and that y and M have a common source (x). In other words, the theorist in question would have to derive x from B. This hypothesis would encounter many difficulties, already pointed out in the course of the argument, and finaily, it would force its upholder to explain why Gorgol occurs (or something that may certainly be identified with it) as the name of certain actual Welshmen in the twelfth century and of actual Bretons in the eighth and ninth. To be sure, Gorgol shows some similarity to Garulf, but that is not strange if Rh^s's theory of the Welsh name Gurguol as = werewolf is cor- rect, iox garulf is the Germanic wariwulf. 5 The MS. has both Gorlagon and Gorlogan, and Gorlogam occurs once (c. 23). « Gurgalanz (Potvin, p. 65), Gurgalain (p. 72), Gorgalan (pp. 73, 74), Gorga- ranz (p. 74). The episode is curious. Gorgalan has the sword with which John the Baptist was beheaded. Many have sought to win it of him, but nobody has ever returned. Gawain essays the quest at a favorable moment, when Gorgalan's 204 ^- ^- Kittredge. Side by side with Breton -uual {-gual), which does not occur out of composition, we have the form Uuallon, which is found a good many times in the Redon Cartulary both as a proper name by itself ^ and as the second part of compound proper names. It corresponds exactly to the Welsh -guallaun. We may cite the following pairs '^ : Breton Welsh Cat-uuallon Cat-gual Cat-guallaun Drid-uual Drid-uallon Dum-uual Dum-uallon Dum-gual Dun-guallaun larn-uual larn-uallon lud-uual lud-uallon lud-gual ' lud-guallaun lun-uual lun-uallon Tut-uual Tut-uallon Tuta-gual (Tut- wal) Clearly uuallon (Welsh -guallaun) is a derivative of -uual (Welsh -gual), probably with the adjective suffix -Ion (Welsh -laun)^ or, at all events, uuallon was early associated with -uual in the Welsh-Breton etymological consciousness, and names in -uallon (^-guallaun) were son has been carried off by a giant. Gawain recovers the son's body, which is then cooked by the king and eaten by his men. The grateful heathen gives Gawain the sword and receives baptism. Evidently something pretty savage has been imperfectly toned down. I do not know of the name elsewhere in French. Gorgalians (nom.), the name of a brother of Julien li Gros (prose Perceval, ed. Potvin, p. 3) is perhaps a different word. Gargeolain is the name of Ruvalen's amie in an intensely Celtic episode in the prose Tristan of MS. Bibl. Nat.fr. 103. But this is still another name. Eilhart, who draws from the same source (Beroul) as MS. 103, calls the lady Garidle. See Bedier's edition of a long passage from the manuscript, Romania, XV, 496 ff. (especially p. 484). I cannot refrain from comparing Tristan's Sport with the rushes in this episode with Cuchulinn's needle- feat in the Fled Bricrend, 65 (Windisch, I, 286-287 i Henderson, p. 82). 1 See De Courson's index, and cf. Loth, Chrestotnathie, I, 171-172, 207-208. There is also a name Uuallonic. 2 The Breton forms are from the Redon Cartulary (see De Courson's index, and cf. Loth, I, 171-172) ; the Welsh forms are from the genealogies in MS. Harl. 3859 (end of nth or beginning of 12th century), thought to have been collected in the loth century, ed. Phillimore, Cymmrodor, IX, 169 ff. (see Anscombe's index, Archiv f. Celtische Lexicographie, I, 187 ff.), or from charters in ^& Book of Llan Ddv, ed. Evans and Rhys. ^ Cf. Loth, Revue Celtique, XI, 145. * See, however, Gliick, Die bet Caesar vorkommenden Keltischen Namen, pp. 49, 164-5, '78 ff-; cf. Kossina, Idg. Forsch., II, 181. Arthur and Gorlagon. 205 freely formed from those in -ual {-gual). Nothing hinders us, there- fore, from adding to the pairs already cited : Breton : — Uur-uual, U ur-gual : * Uur-uallon. Welsh : — Gurgal, Guorguol, Guruol, Gurguol : * Gur-guallaun. Thus the Gorlagon {Gorlogati) of our version G appears to be an easy metathesis for *Gorgolan, corresponding to the Gorgalan of the prose Perceval and to a lost Welsh * Gurguallaun. And if Gorgol means "werewolf," Gorlagon (*Gorgolari) means practically the same thing. The other name Gorbeil, Gorleil, or Gorliel has so uncertain a form that it is idle to dogmatize about it. The second syllable, -beil, is con- ceivably a corruption of Welsh Beli (Breton Bili, -bili, -uuili), in •which case Gorbeil may be compared with the Breton name Uuor- uili (Guor-uili, Uur-uili)} Another possibility is that we have in the second syllable {-beil?) a corruption of the Welsh bela {bald), ■" wolf," " wolf's cub." ^ This would give us " werewolf " as the mean- ing of Gorbeil as well as of Gorgol and Gorlagon, and the condition of things would agree exactly with our inference (made on other grounds than those of etymology) that the " three brothers " of G •were originally three separate manifestations of one and the same person. But Gorbeil is a rather dubious reading to operate with, and the case is good enough without it. Whether the etymologies suggested for Gorgol, Gorbeil (?), and Gor- lagon are right or wrong, the names are unquestionably not Irish, and therefore cannot have stood in y. They are either Welsh or Armori- can. Between these two languages it is impossible to decide on the basis of the forms preserved in the Latin G. G is certainly rendered from a prose text, either Welsh or Breton, which was similar in style and general character to the " Four Branches " of the Welsh 1 For these names, see De Courson's indexes; Loth, I, no, 178 (cf. 191, note 2) ; index to Book of Llan Ddv ; Red Book of Hergest, ed. RhJ^s and Evans, II, index. 2 So defined by Silvan Evans, Dictionary, following Owen Pughe. The word is rare, however, and its meaning doubtful. Professor Robinson refers me to Loth's discussion of bala (Archiv f. kelt. Lexicographic, I, 457-458). where " fox " is suggested. 2o6 G. L. Kittredge. Mabinogi. No such texts are preserved in Armorican. The passage from the Irish y to Wales would be a shorter journey than the passage from y to Brittany ; indeed, the latter itinerary would involve, in all likelihood, an actual transit through Wales. It is, then, much easier and more natural to regard the immediate original of the Latin G as a Welsh than as an Armorican version of the Irish y ; and accord- ingly I have adopted that hypothesis. It accords with the well-known influence of Irish literature upon that of Wales. The Welsh hypothesis may perhaps be strengthened by certain special considerations. The rationalization of G in one significant particular has already been mentioned and accounted for. In x the wife was certainly z. fke and she was not punished for abandoning her mortal husband. These features were not abandoned by y, as is shown by their preservation in I. In G, however, the insertion of an Oriental tale, The Dog and the Lady, has necessitated a complete change in the lady's nature, — she is no longer a fie, but a mere woman, conceived after the cynical manner of the East. The inser- tion of this tale and the consequent rationalization of the story may be ascribed to the Latin translator, who was doubtless a clerk,* and therefore likely to be familiar with such material ; The Dog and the Lady, we should remember, was afterwards made a part of the Gesta Romanorum,^ a monkish collection of exempla for the use of preach- ers. In the Welsh text, then, of which G is a translation, the Other- World character of the lady was probably preserved. With this in view, it is pleasant to find in the mabinogi of Pwyll Prince of Dyvet a tale which resembles, in general outline, as well as in some particular features, the fairy mistress story of x and y. Rhiannon, Pwyll's wife, is certainly z-fte, and his first interview with her is similar to Melion's with the (fairy) Princess of Ireland and Eochaid's with Etain of the side. Pwyll sees Rhiannon as the result of sitting on a marvellous mound (or seat).' He asks whence she comes, and why, and who she is. She replies that she has come to seek him, mentions her par- entage, and adds : '' Je n'ai voulu d'aucun homme, et cela par amour 1 Note his quotation from Catonis Disticha in chap. 14 (see p. 157, above). ^ See p. 247, below. ' Lady Guest, Mabinogion, III, 46; Loth, I, 38 ff. Arthur and Gorlagon. 207 pour toi, et je ne voudrai jamais de personne, k tnoins que tu ne me repousses." This is remarkably like the Tochmarc Etaine and the Lai de Mellon. We may also compare Pwyll's reply with that of Eochaid : " If I had my choice of all the women and maidens in the world," says Pwyll, " it is thou that I should choose." '^ Nor does the parallel end here. Pwyll is subsequently deprived of Rhiannon by Gwawl, an old suitor of hers, but follows her to Gwawl's abode and wins her back. All these correspondences between the mabinogi on the one hand and M and the Tochmarc Etaine on the other, point to such a resemblance between native Welsh fairy- literature and the Irish story of which y was a version as would have made the naturalization of y in Wales (according to our hypothesis) the simplest thing imaginable. Add to all this the fact that the mabinogi of Pwyll likewise contains the incident of the Hand and the Child (which, as we have seen,* stood in y, though not in x) and the hypothesis that favors a Welsh (rather than an Armorican) trans- lation of y as the immediate source of the Latin G must be admitted as extremely probable.' Finally, the resemblance between extant Welsh literature and the story that we are investigating extends even to some of the werewolf elements. In the fourth branch of the Mabinogi {Math, Son of Mathonwy), Math strikes with his enchanted ring his nephews Gwydyon and Gil- vaethwy, sons of Don, and transforms them successively into a doe 1 Lady Guest, III, 51 ; Loth, I, 44. With the year's postponement in Pwyll, cf. Tochmarc Entire, Arch. Rev., I, 304; Haupt's Ztsch., XXXII, 240-241; Hull, Cuchullin Saga, p. 82. ^ See p. 168 ; cf. pp. 222 ff. ' It might even be contended that the elements in Pwyll which we have been comparing with x and y are borrowings from Irish (perhaps from y itself). This, however, is not my opinion (except, perhaps, with reference to the Hand and the Child). For our immediate purpose, the point is of no moment. The presence in Welsh literature of these close parallels to the remoter original of G must certainly give us confidence in choosing between Wales and Armorica as the country in which the immediate original of G was written, — it being remembered that the proper names in G (which cannot have stood in the Irish y) point either to Brit- tany or to Wales. On the resemblance between the story of Rhiannon and the Tochmarc Etaine, cf. Anwyl, Ztsch. f. celt. Phil., I, 288-289. 2o8 G. L. Kittredge. and a deer, a boar and a sow, a pair of wolves.^ Each transforma- tion is accompanied by a speech, in terms practically identical, and always including the provision that they shall have the instincts of the animals in question : " Vous aurez les instincts des animaux dont vous avez la forme " ; " Vous aurez les m^mes instincts que les pores des bois " ; " Ayez les instincts des animaux dont vous avez la forme." Compare the formula in Arthur and Gorlagon: " Sis lupus et habeas sensum lupi " (c. 6). Math restores his nephews to human form with a stroke of the same magic ring.^ In view of what has been said, the fact that G is a translation or adaptation from the Welsh can hardly be denied. The lost Welsh document, if we had it, would prove to be very similar to some of the tales in the extant mabinogion. It may almost be described as a lost "branch" of that collection, though actually it was a translation or adaptation from the Irish, like some things in the extant mabinogion 1 Loth, Mabinogion, I, 132 fl[. On the remark "vous avez eu la grande honte d'avoir des enfants I'un de I'autre " (1, 134), cf. Lokasenna, sts. 23, 33 ; Helgakvi&a Hundingshana I, sts. 38, 39 (Bugge) ; Hyndlulj6^ , st. 40 (Bugge) ; Gylfaginning, c. 42. Observe that the hero consorts with a she-wolf in G and that the pair, and their two whelps, correspond to the band of wolves in M and in some versions of I (see p. 180). Probably G here preserves an old incident which has been softened or suppressed in all other versions. Doubtle.ss it stood not only in y and x but in the Irish Werewolf's Tale proper (0), the source of B. In B, however, there is no mention of a band of wolves. This may be taken as further testimony that x is not derived from B, if more evidence is needed. Paris has subjected the ignominious punishment inflicted on the enchanter flliavres by the elder Caradoc to a learned and ingenious examination, with happy results (Rom., XXVIII, 217, note). He has overlooked, however, the excessively curious episode in the mabinogi of Math Son of Mathonwy, which furnishes a striking parallel. In the mabinogi, as in the Perceval, the initial offence is an intrigue with the prince's favorite or wife. All this may go to support the present text of the Livre de Caradoc and to vacate Paris's conjecture that the incident has been transferred from an earlier portion of the poem and shifted from Caradoc Senior to filiavres. At all events, though no one would think of deriving the incident in Math from that in the Perceval, or vice versa, the parallel certainly aids in establishing not only a Celtic but a specifically Welsh source as that from which the adventures of Caradoc (or some of them) made their way into French, and so confirms the arguments of Lot (Rom., XXVIII, 578) ; see also Rhys, Celtic Folklore, II, 689-690, 694 (note to pp. 579-580). ^ There is also a werewolf incident in Kulhwch and Olwen (Loth, I, 266). Arthur and Gorlagon. 209 themselves.^ Whether the Arthurian elements were added by the Welsh redactor or by the author of the Latin G cannot be determined, — very likely by the latter. The date at which the Welsh version was made cannot be fixed. If Gorgol means " werewolf," however, the Welsh version must have arisen before the etymological signification of the term had lapsed from the Welsh consciousness, since the name was the insertion of the Welsh redactor. We do not know how early this sense was lost, but, on the other hand, nothing prevents our putting the Welsh version pretty far back. The Latin text itself is preserved in a manuscript of about the age of the Red Book of Hergest. V. THE FRAME-STORY IN VERSION y. The most striking distinction between x and y is, as we have seen, the fact that in y The Werewolf's Tale, which stands by itself in x (see M and cf. B), is told by the Werewolf, under com- pulsion, to a quester who is in duty bound to learn it. This method of introducing a story is not unexampled. It is familiar to all readers of Irish mdrchen. Nor are instances wanting in which a tale occurs both without such an introduction (like B and x) and with it (like y). An excellent example is the favorite Irish story called The Fairy Palace of the Quicken Trees, This is found without the frame in no less than fourteen versions.'^ In five 1 See p. 245, below. 2 (i) J. F. Campbell, Revue Celtique, I, 193 ft. ; (2) MacDougall, Folk and Hero Tales, pp. 56fiE. ; (3) MS. of 1600 in the Advocates' Library, Edinburgh, cited by MacDougall, p. 270 ; (4) Dunstaffnage MS. of 1603, in' the same, cited ibid., p. 271, printed by J. F. Campbell, Leabhar na FHnne, pp. 86 ff., and imperfectly summarized by the same, Pop. Tales of the West Highlands, II, 187, and fully by MacDougall, pp. 271 ff.; (5) J. F. Campbell, Fof. Tales, II, 168 ff.; (6) another MS. in the Advocates' Library, cited by MacDougall, p. 271 ; (7) Joyce, Old Celtic Romances, pp. 177 ff., from three Irish MSS., of 1733, 1766, 1841 (p. xiv) ; (8) Curtin, Hero-Tales, ^^. 407 ff. ; (9) J. G. Campbell, The Fians, pp. 233 ff. ; (10) the same, p. 74 (summarized) ; (11) Curtin, Myths and Folk-Lore, pp. 221 ff. ; (12) the same, pp. 281 ff.; (13) the same, pp. 292 ff. ; (14) Kennedy, Bardic Stories of Ireland, pp. 1 16 ff. Cf. also The Chase of Slieve Fuad, Joyce, pp. 362 ff. (and the poem published by O'Daly, Transactions of the Ossianic Society, VI, 20 if.). 2IO G. L. Kittredge. versions,^ however, it has precisely the kind of introduction that we are investigating : a quester is compelled to discover " what has kept the King of Erin cheerless and laughterless for the last seven years " (or the like) ; the king objects to telling his story, since it involves the disclosure of a disagreeable and humiliating experience, but he yields to force majeure. The pressure exerted to elicit the story may be physical, or consist in threats of death or violence. On the other hand, it may assume a finer form (as in G).^ King Arthur comes upon Gorlagon at table, but refuses to dismount and join in the feast until he gets the story.' Gorlagon pauses several times in his narrative to repeat his invita- tion : " Arture, descende et comede," but to no purpose ; he is forced to continue. The compulsion is involved in the disgrace that befalls a host whose hospitality is rejected.* It is a kind of ceremonial interdict : he must not go on with the banquet till his guest is con- tent to share it. A striking instance of this method of moral suasion may be seen in one of Larrainie's West Irish tales ^ : A great feast has been prepared for Finn by Pampogue, but he declares that " he will not eat a bit until Pampogue grants him a request." Pampogue replies that she "will grant any request except to let her husband go to fight with the Blauheen Bloye," — an expedition which she is sure will be his death. " Unless you grant me that," says Finn, " I will not eat any food." " Sooner than you should be without eating, I 1 (15) MacInnes,i'o//5a«(^^^ytf7fl/«, pp. 72-73; {id) Cwxixa, Myths and Folk- Zore.-pp. 121 ff. ; (17) the same, pp. 256 ff. ; (18) the same, pp. 428 ff . ; (19) the same, Hero-Tales, pp. 477 ff. '^ For other instances of extorted stories, see Larminie, pp. 45, 151, 171. 8 For the heroic lengths to which a king might be expected to go when his hospitality was impugned, see The Destruction of Da Derga's Hostel, cap. 63, ed. by Stokes, Revue Celtique, XXII, 60, 61. * The mounted messenger (or the like) riding into the hall and refusing to get off his horse till his boon is granted is a familiar figure in Arthurian romance. We should also remember King Arthur's habit of refusing to eat, on a high day, until some adventure had happened (Child, Ballads, I, 257, note % ; III, 51, and note §). There is a striking Irish parallel in the shorter Fled Bricrend (from the Yellow Book o/Z(f<:fl«), Windisch, Irische Texte, II, i, 174, 188: "It is not fitting to consume this feast of mine without a brave deed of the Ulstermen in return for it." 6 j^gji /„-j^ Folk-Tales, pp. 76-77. A rthur and Gorlagon. 211 ■will grant even that," replies Pampogue. A stronger instance, or one more thoroughly Celtic, could hardly be required.' We may safely infer that in y the compulsion exerted to get the story from the Werewolf consisted in refusing his hospitality (as in G).^ This accords with the quester's vow not to eat (in G) — and with the requirement (in I) that he shall not eat twice at the same table — until he has learned what he wishes to know. King Arthur's adventures on his way to the abode of Gorlagon have already been described: he visits successively Gorlagon's younger brothers, Gorgol and Gorbeil (?), and is cajoled by them both. We have seen reason to believe that Gorgol, Gorbeil, and Gorlagon are really one and the same person, — the Werewolf, who, to avoid telling his story, attempts to delude Arthur by meeting him at three different times and under three different names, — a device common in Irish legend (p. 201). In these preliminary adventures, G probably follows the plot of y,^ though the actual names of the masquerading Werewolf must first have made their appearance in the Welsh (p. 205). The similarity of the names may indicate that the Welsh redactor, who is responsible for them, understood the identity of the three "brothers "; and if all three of the names (or even two of them) mean "werewolf," 1 A very curious instance of moral pressure is in Coise CHn (Kiati's Leg), Maclnnes, pp. 235 ff. Here the hero refuses to allow his broken thigh to be treated until he has elicited story after story from the would-be healer. " Stretch forth your leg, Kian, that I may apply to it leaves of herbs and healing. Pressure and business are upon me ; and I am under the necessity of going to the big church of Rome to-morrow to listen to joy.'' " I will not stretch forth my leg . . . until you tell me why . . . ." And so on, time and time again, in this extraordinary conglomerate of stories. See J. G. Campbell, Superstitions of the Highlands and Islands of Scotland, p. 132. 2 In I the quester first steals the sword of light and then threatens to kill the Werewolf with it unless he shall tell the story. The sword of light (as we shall see presently) was not in y. I, then, has certainly departed from y in its account of the means which the quester adopts to make the Werewolf answer his question. In J the Werewolf gives up the sword and tells the story readily enough when the quester has passed two dragon guards. 3 I here leaves us in the lurch. It has taken in an independent tale, The Sword of Light (see pp. 2 1 3 ff .), and the combination obscures the original course of the narrative. Yet even I affords a trace of the situation in y : in KJ, the Werewolf is the brother of the quester's father-in-law and of the challenger. 212 G. L. Kittredge. there can be no doubt that he comprehended the situation perfectly. The Latin translator, however, failed to grasp the device. He took the three " brothers " for three distinct characters, and accordingly equated Gorgol with the king who befriended the Werewolf^ and effected his restoration to human form. In both G and I the faithless wife is present while the tale is told. Much is made of her presence in I, and it was doubtless a feature of y.^ The occasion of the quest for " the cause of the one story " in y must remain a matter of conjecture. In I the adventurer is required to learn the story as the penalty for losing a game to a supernatural challenger. This is an excellent Irish incident, but the evidence of I is worthless here, since the gambling incident is ^borrowed from a distinct tale which I combines with y to make the frame-story.^ In G we find an exceedingly lively and picturesque introduction : Arthur is holding his Pentecostal feast at Caerleon. After dinner, in the joy of his heart, he throws his arms about his wife, as she sits by his side, and kisses her in the presence of the assembled court. Scandalized at such a breach of decorum, the queen blushes furiously and asks the reason for his undignified behavior. "Quod nichil mihi in diuiciis gratius," " nil in deliciis te constat suavius," is the amatorious response. " If you love me so much," rejoins the queen, " you must believe yourself well acquainted with my mind and will." Says Arthur, " Your mind, I am confident, is well-disposed towards me, and I am sure I understand your will." " You are mistaken ! " she replies, " You have never known a woman's nature (ingenium) or her mind." " Omnia cell obtestor numina," cries the angry king, " if I have hitherto been ignorant of these matters, I will never taste food till I discover them ! " And he sets out, with Kay his steward {Caius) and Gawain {Walwainus) his nephew, to visit King Gorgol and learn the secret. This reminds one of the introductory incident in the Pelerinage Charlemagne,^ but it would be overhasty to infer that it is borrowed 1 Cap. 23, p. 162. 2 See p. 220. s gee pp. 214 ff. * The Pilerinage is here (as elsewhere) closely parallelled by the fragmentary English ballad of King Arthur and King Cornwall (Child, no. 30, I, 274 ft.). Child at first regarded the ballad as " an imitation or a traditional variation " of the French chanson (I, 274), but was induced to change his mind and to refer the Arthur and Gorlagon. 213 from that poem. In the first place, we are not to suppose that the feature in question first came into existence when the Pehrinage was composed. It is rather an incident which the author of the poem knew independently of the story of the Pelerinage and which he utilized (with superb effect) to motivate Charlemagne's journey.^ And, in the second place, there is considerable difference between the incident in the Pelerinage and that in Arthur and Gorlagon. The two incidents simply belong to the same general type of popular legend. VI. PECULIARITIES OF THE IRISH VERSION (I). We have seen that the Irish mdrchen (I) is distinguished from all other versions in three ways: (i)the husband passes through a suc- cession of metamorphoses ''; (2) the frame-story is complicated by a quest for the Sword of Light, and (3) the incident of the Defence of the Child is expanded in a manner quite out of proportion to the modest place which it occupied in y. The first of these peculiar- ities needs no discussion ; it is an easy and natural elaboration of the single transformation which stood in y. The other two special features of I, however, require particular study. VII. THE QUEST FOR THE SWORD OF LIGHT. In y, as we have seen, The Werewolf's Tale of x was inserted in a frame-story: a quester is required to learn the "cause of the one two to a common source by the arguments of Paris, Hist. Litt., XXX, iio-ni (see Ballads, III, 503). More recently. Dr. W. D. Briggs has argued strongly in favor of Professor Child's first opinion (Journ. of Germanic Philology, III, 342 ff.). For parallel stories see Paris, Romania, IX, 8 f£., and Hist. Litt., XXX, 94; Child, I, 279, 282-283. In the unpublished French Rigomer and in Heinrich von dem Turlin's Crdne (vv. 3313 ff.), there is a, somewhat similar scene attached to the Arthurian cycle. Dr. K. G. T. Webster, who is investigating the history of Guinevere, has subjected Arthur and Cornwall to a searching examination and finds new grounds for referring it and the PHerinage to a common source ; but I must not anticipate his results. 1 Paris (Rom., IX, 8) has pointed out that the PHerinage combines two stories, originally distinct, — (i) the king who visits his rival, and (2) the pilgrimage proper. " So in KJLHO'FCiCj, but not in S. 214 G. L. Kittredge. story about women," or he puts himself under bonds to learn it. In I this frame-story is complicated by an additional quest, — the adventurer must secure the Sword of Light. This weapon turns out to be in the possession of the same person who knows the story. The quester secures the sword first, and uses it as a means of compelling the Werewolf to tell the tale. Fortunately, The Quest for the Sword of Light occurs,^ in a form almost identical with that in I, but quite out of connection with The Werewolf s Tale, in a Scottish Gaelic mdrchen. Two ver- sions of this mdrchen have been printed, both in Gaelic and in English : J. F. Campbell's Young King of Easaidh Ruadh (c) ^ and Maclnnes's Herditig of Cruachan (m).' The hero plays (shinty m ; a game not specified c) with a wizard- champion {gruagach). He wins the first game and takes as his prize a " little untidy, swarthy woman cleaning the byre " (" cropped rough-skinned girl behind the door " c : in c, but not in m, she becomes beautiful when she reaches his house). The second time he wins and takes a "dun shaggy filly." The third day he goes to play (against his new wife's advice, who has informed him that his opponent is her father c) and loses. The wizard-champion requires him to get " the white sword of light that the King of Sorcha has " (" the Glaive of light of the King of the Oak Win- dows " c).* 1 Sword-quests occur everywhere, and the Sword of Light is a familiar weapon in fairy-tales. We are here concerned, however, with a particular form of this quest. Still, it may be worth while to compare Larminie, pp. 206 ff. ^ Popular Tales of the West Highlands, no. 1, I, I ff. In another version still, summarized by J. F. Campbell, I, 18 ff., the Quest of the Swordis wanting, probably from a lapse of memory. It is barely possible, however, that we have in this version the tale as it existed before the Quest of the Sword was compounded with it. The recovery of a stolen wife from a giant whose soul is out of his body, and the capture of the soul by the aid of animals, form a well-known incident in folk-literature. See, for example, Nutt's note to Maclnnes, pp. 455 ff., and Kohler's remarks in Orient and Occident, II, lOI— 102. ' Folk and Hero Tales, no. 4, pp. 94 ff. Curtin's Son of the King of Erin and the Giant of Loch Liin, Myths and Folk-Lore, pp. 32 ff., begins as if it belonged to this set, but goes on later with a different type of story (Maclnnes's no. i, pp. 2 ff. ; see Nutt's note p. 431). * "Claidheamh soluis righ nan uinneagan daraich " (Campbell, p. 13); "Clai- dheamh geal soluis a th' aig rlgh na Sorcha" (Maclnnes, p. 102). Arthur and Gorlagon. 2 1 5 From this point the order of events differs in c and m, though the incidents themselves are the same to all intents and purposes. I fol- low c first, returning to m later. His wife consoles him, and he sets out on the filly, who, the lady tells him, will give him all necessary instruction. The filly carries him to the castle of the King of the Oak Windows and tells him what to do. The king is at dinner, and the sword is in his chamber. The hero steals the sword, which gives a sort of sgread as it comes out of the sheath. There is a great pursuit, but all fall behind except the King of the Oak Windows, mounted on the brother of the filly, who is swifter even than she. As the pursuer is passing, the hero, acting under his filly's instructions, strikes off his head with the sword. He then mounts the swifter horse, the filly fol- lows, and they reach home in safety. His wife receives him gladly and tells him what to do when he meets the gruagach on the morrow. The gruagach is the brother of the King of the Oak Windows. He will ask the hero how he got the sword. The latter must answer : " If it were not the knob that was on its end, I had not got it.'' When the gruagach "gives himself a lift" to look at the alleged knob, the hero will see a mole on the right side of his neck, and he must then stab the gruagach in the mole. The hero does as he is told, and ^t. gruagach falls dead. When the hero returns home after this encounter, he finds that his wife and the two horses have been carried off by a giant. He sets out in pursuit, and falls in successively with a dog, a hawk, and an otter, who direct him on his way.'^ At last he finds his wife and the horses in a chasm [which is the giant's den]. The woman hides her husband and cajoles the giant when he returns and smells human flesh. She induces the giant to tell her where his soul resides [for it appears that he is one of those monsters, familiar in folk- lore, who have no soul in their body].^ There is a flagstone under the threshold ; under the stone is a wether ; there is a duck in the wether's belly, and an egg in the belly of the duck ; in the egg is the giant's soul. The hero and his wife remove the stone. They catch the animals and get the egg, by the help of the dog, the hawk, and the otter. The lady crushes the egg, and the giant, who is on his way home, falls dead. Then the couple return to their own country, taking with them much of the giant's gold and silver. ^ I have condensed the tale very much at this point. 2 Here, too, I have condensed. It takes three days to carry out the lady's plot. The details follow a well-known type of mdrchen (see p. 214, note 2). 2i6 G. L. Kittredge. Maclnnes's Herding of Cruachan (m), as I have already observed, has the adventures in a different order : When the hero returns after his third game with the wizard-champion, he finds that " the big giant, King of Sorcha," has stolen his wife and the filly. Consequently the quest for the Sword of Light and the search for the stolen wife are included in a single journey. The hero is assisted by animals (as in Campbell): four, however, instead of three, — a hawk, a duck, a fox, and an otter, each of whom inhabits a little house. The concealment of the giant's soul is more elaborate than in c, and all four of the helpful animals, as well as one of the giant's horses (which seems to correspond to the swifter of the two steeds in c), are needed to get it. When the giant is dead, the hero and his wife return home, taking with them " all the gold and silver that the giant had, his white sword of light, the big dappled horse, and the shaggy dun filly." On their reaching home, the hero's wife tells him how to outwit the wizard-champion. He is to give him the sword. The champion will then boast of the weapon, and the hero is to reply that it has a flaw. The wizard-champion will say, " Show me the flaw." The hero is then to take the sword and cut off the champion's head, with the remark " This is the flaw that it has." The programme is duly carried out. Thus the outwitting of the champion comes at the end of the tale in m and not (as in c) in the middle. This story, whether in Campbell's version or in Maclnnes's, mani- festly consists of two independent tales, more or less skilfully welded together: (i) The Quest for the Sword of Light, and (2) The Abduc- tion of the Wife, and her rescue, with the death of the giant. It is the first of these that furnished I with the frame in which The Were- wolf s Tale is set.'' 1 Whether the author of I (that is, the person who inserted in y The Quest of the Sword of Light) knew The Quest in combination with the Abduction of the Wife (substantially as in Campbell and Maclnnes) is not to be determined. Prob- ably he did not ; at all events, he did not utilize the Abduction. One version of I (O'Foharta's) shows practically the whole of the combined mdrchen (Quest of the Sword plus Abduction of Wife). O'F, indeed, affords a version of this tale which is in some respects better preserved than either Campbell's Young King or Mac- lnnes's Herding of Cruachan, for it motivates the gratitude of the beasts. It also shows a trace of an incident found elsewhere in Maclnnes (p. iii) only, the dancing of the helpful animals (p. 488). It is clear, however, that O'F departs from Arthur and Gorlagon. 2 1 7 The manner in which The Qicest for the Sword of Light has been utihzed to complicate the frame-story of The Werewolf's Tale in I is rather ingenious. The introductory incident of the Quest is adopted in its entirety. The hero ^ is thrice victorious in gaming with a mysterious stranger : he wins a beautiful wife, a magic horse, a castle, etc.'' He loses the fourth game, and the stranger requires him, as a penalty, never to eat two meals I in thus including the Abduction of the Wife, for the inclusion disorders the story. We may infer that O'F was made up by some reciter who knew I (The Werewolf s Tale combined with the Quest of the Sword) and was also familiar with the double mdrchen represented by 'Campbell's Young King and Mac- Innes's Herding (Quest of Sword combined with Abduction of Wife), and who chose (or chanced) to increase the complexity of I by including the whole of the double mdrchen. 1 In L the hero is called Morraha (cf. p. 254, note l) ; in O'F he is Murrogh, son of Brian Boru ; in Ci, he is " Art, the king's son " ; in C2, Arthur, a cotter's son ; in HS he is son of the king of Ireland (but no name is given him) ; in KJ he also has no name but is described as ?l sgolog ox "small farmer." 2 The versions of I differ slightly. In S there are two winning games, the prizes being the woman who is riding behind the challenger, and the horse ; the third game is lost. S agrees pretty closely here with Campbell's Young King and Maclnnes's Herding of Cruachan, and may perhaps be more correct than the other versions of I; three games in all, two won and one lost, seem to accord with reason and sym- metry. It is not impossible that this gambling adventure was in some form a part of y ; it presents a striking parallel to the chess-play between Mider and Eochaid in the Tochmarc Etaine. See d'Arbois, Cours, II, 315 fit. In L, the hero wins sheep on the first day, cattle on the second, a castle and the fairest of women on the third ; the horse he procures by shaking a magic bridle which belongs to his wife. K agrees substantially, but lacks the incident of the bridle ; the horse comes with the woman. C2 is much the same, but the horse is replaced by a hound. In O'F he wins riches, castle, and lady, all in one game, and loses the second game. In Ci he wins " the finest woman on earth, with twelve attendant maidens and thir- teen horses," in the first game and loses the second. In H he wins his wife by the first game (the magician takes him to his castle and gives him his choice of many beauties, but he takes a girl from the kitchen at her own suggestion ; she becomes beautiful while they are riding home) ; cattle by the second (but he loses them by a trick); by the third, cattle that remain; his choice of horses by the fourth (he chooses a poor-looking mare) ; he loses the fifth game. In J he wins money by the first game, the fairest of women by the second, and loses the third ; his wife procures the horse by means of a magic thread (cf. L). Ci, it should be noted, is the second adventure in a long composite. 2r8 G. L. Kittredge. off one table and never to sleep two nights in the same house ^ till he brings him the sword of light and " the knowledge of the cause of the one story about women." ^ Version y, as we have already seen (p. 200), must have contained, the requirement to bring "the cause of the one story"; to this is added, in I, the demand for the Sword of Light, and thus the 1 So KLC20'P, in almost identical words. In O'F, however, the requirement is laid upon the hero on another occasion (p. 216, note). In H the challenger uses a different formula. In Ci the formula is missing : the challenger says simply, " You are to bring me the sword of light and the story of the man who has it." For S, see note 2. Observe that G shows a trace of the formula that is found in KLC20'F : King Arthur swears a great oath " nunquam cibo f ruar donee ea me nosse con- tingat." Some such formula must, therefore, have stood in y. We cannot tell how the requirement came to be laid upon the hero in y (see p. 212); perhaps G, with its undignified kissing in public, is a good representative of y. '^ The words quoted are from O'F (Jios fath an aon sgeil ar na mndibh), but K has almost the same thing (Jios fath an aon sceil, i.e. " the knowledge of the cause of the one story " ; mistranslated by Kennedy " perfect narrative of the unique story " ). The similarity of G, in which King Arthur sets out to discover the " ingenium mensque feminae," is evidence enough that O'F is here close to y, except for the sword of light, which is peculiar to I. L and S both show a corrup- tion. L has " till you bring me the sword of light and the news of the death of Anshgayliacht.'" This strange name (which Larminie, p. 252, interprets as an sjgeeliaxt, "the Story-Telling") obviously contains the Irish word scH ("story") preserved in O'F and K. In H the hero is to " bring the sword of light of the son of the King of the Speckled Peak and the story, who killed the Antichrist (sgeula cia mharbh ant-An-Chriosdaigh)." An-Chriosdaigh (like A nsAgayliacAt in h) is the name of the monster who has stolen the children. H is farther gone in corruption than L. S has " I lay as crosses and charms upon you that water leave not your shoe till you find out how the Great Tuairisgeul was put to death (ciamar a cliaidh an Tuairisgeul Mor a chur gu bas)." Tuairisgeul (which J. G. Campbell glosses " description, report, calumny ") is a compound of sgeul (the Scottish Gaelic form of Ir. sc^l, "story"). Thus LHS support O'FK. LHS form a group by them- selves, since in them the title of the story is made into the name of a person and that person turns out to be the monster that stole the chUdren. LS also agree in the shaking of the bridle, though the circumstances difEer (see p. 217, note 2). For Ci, see note I . C2 differs from all other versions in requiring the hero to find " the birth that has never been born, and that never will be." This comes from contamination with another story, — the tale of a champion who was, like Macduff, " not of woman born." C2 lacks the Sword of Light, as does S (but see p. 220, note 4). For the requirements in J, see p. 268. Arthur and Gorlagon. 219 independent mdrchen of the Quest for that weapon is incorporated in the frame-story of The Werewolf s Tale. I continues as follows : — The hero fulfils his tasks by the aid of the horse which he has won with his (fairy) wife. This horse carries him (across the sea i) to his father-in- law,* who receives him well and tells him what to do. Three times he rides, on three diflferent horses (furnished by his father-in-law), to the residence of the terrible enchanter = who has the sword and knows the story, summoning: him to give up the one and tell him the other. He rides off as swiftly as possible after each summons, pursued by the enchanter. The first time, the enchanter cuts his horse in two ; the second time, he cuts off his horse's hind-legs ; the third, his blow is harmless.* The enchanter is now weary, 1 So in KHJ; through a loch S ; through the sea (a road opening to King Under- the-Wave's realm) Ci ; over three miles of fire, three miles of hill covered with steel thistles (or needles, Larminie, p. 253), and three miles of sea L. C2 lacks the horse (see next note). 2 The father-in-law is manifestly a prince of the Other World. In L he is called King of France (Greece KJ). In H he is King of Speckled Peak in the Eastern World. In Ci he is King Under-the-Wave (a well-known Celtic character). In C2 the hero goes to the " castle of the son of the King of Lochlin " and becomes his retainer. He performs great services for his master (which have nothing to do with our tale) and finally brings back the wife of the king's son from a giant who had abducted her. In return he asks the solution of his problem. The king's son then brings out the old King of Lochlin, who has long been in hiding, and asks him for the answer. The king twice refuses to tell, but yields at last to the persuasion of a hot griddle. His story is a version of our Werewolf's Tale. " Rough Niall of the Speckled Rock L ; the son of the King of the Speckled PeakH; Fiach O'DudaK ; the Young Champion J. He turns out to be the Were- wolf. In KJ he is one of three brothers, the other two being the hero's father-in- law and the person who sends the hero on the quest. In HO'F he is the brother of the hero's (fairy) wife. In Ci he is Balor Beimenach and is a son-in-law of the hero's father-in-law. King Under-the-Wave. In Cj (which has been much changed by contamination) he is the King of Lochlin, the father of the personage to whom the hero is sent ; but C2 has nothing of the Sword of Light. In S he is an " old grey man " who lives on the farther side of a loch ; nothing is said of his relation- ship to the other characters. In L, Anshgayliacht (see p. 218, note 2) is the brother of the gamester (but this must be an error). * So in K. In L the enchanter (i) cuts the horse in two, (2) cuts off half the horse and half the saddle, (3) cuts away the saddle from under him and the clothes from his back. The second stroke in L will not do ; there should be a steady decrease in the damage done. But perhaps the third stroke in L is more nearly right than in K. In Ci the first blow cuts the horse in two behind the saddle ; the 220 G. L. Kittredge. having been on the watch for three days and three nights, and falls asleep.^ The hero returns, creeps into the bedroom, and steals the sword. He then rouses the enchanter and demands the story. The enchanter at first refuses ; but his wife persuades him to tell it to save his head.^ She is present while the story is told.' When the story is finished, the hero returns to his own home with the sword. The conclusion of The Quest for the Sword of Light '\s now utilized to bring I to a fitting end. The hero takes the sword to the person who had sent him on his perilous journey, and tells him the tale.^ He does not deliver up the second, just at the saddle ; the third, with a piece of the saddle : that is, the blows increase in effectiveness. O'F agrees with K and L as to the first stroke ; the second time the horse's tail is cut off (cf. K) ; the third time the hero finds the enchanter asleep and steals the sword. The owner follows him to the house of the hero's father-in-law. H resembles O'F, but there is no damage done to the hero's horse : the enchanter pursues him to the king's house on the first two nights, hut on the third the sword is stolen. Ca of course lacks the incident (see p. 219, note 2). In J dragons guard the castle and the occupant does nothing. 1 So in L (but confused). In K the hero puts the enchanter to sleep with a magic harp, but K is very much elaborated at this point. In H the guards sleep only three nights every seven years ; they are awakened on the first two nights by the shriek which the sword gives (cf. J. F. Campbell's Young King, p. 215, above), but on the third night the sword is in the hero's hand before it cries out. In J the guardian dragons are asleep on the third night. 2 So in Li- In HKJ the enchanter submits without parley. In O'F he requires the presence of his wife and imposes an extraordinary condition (for which cf. another story in Larminie, p. 74). 2 The presence of the wife while the story is told is an important feature, for it is common to I and G (see p. 212). It occurs in LJHO'F (LO'F are extremely racy here), but not in K. It must once have stood in Ci, which should here be com- pared with O'F. In C2 and S it is of course lacking, on account of other changes. * What follows is given according to LCi, in which the hero acts in accordance with the directions of the owner (the Werewolf). In H also the owner of the sword tells the hero what to do. The challenger will take the sword and vrill ask the hero if there is another so beautiful in the world ; the hero is to assent con- ditionally : " It is beautiful, but fo> ." " What means your but for ? " will be the reply ; and the hero is to explain by taking the sword and cutting off the ■challenger's head ; he is then to throw the sword into the air, as in L. In K the hero says " How shall I give you the sword? " and when the challenger rephes " As you like," the hero cuts off his head with it (cf. J. F. Campbell's Young King Arthur and G or la swordji however, but quibbles as in Campbell's Young King 2in6. Maclnnes's Herding of Cruachan. " I promised to bring the sword ; I did not promise to give it to you.'' Then he throws the weapon into the air, and it returns to its owner. 1 We may infer that, if he had done so, he would at once have been slain with it. and Maclnnes's Herding of Cruachan) ; this may be nearer the original. O'F ends with the conclusion of The Werewolf s Tale; the Werewolf says " So now you have the story of the Shining Sword and the knowledge of the cause of the one story about women," and there is nothing further. In J the challenger dies before the hero's return (cf. S, below) ; the hero keeps the sword. C2S lack the sword. The conclusion of S deserves attention. When the challenger " lays crosses and charms " on the hero to discover "how the Great Tuairisgeul was put to death," the latter (as his wife has bidden him) replies : " I lay the same charms upon you that you leave not this hillock till I return." On reaching home with the story, the hero is instructed by his wife to go to the hill and recount it to the challenger. " What is the good of it," he replies, " when the one bone of him does not stick to another to-day ? " But the woman insists and the hero obeys. When the story is finished, the challenger "rises alive and well from the hillock." This cannot he quite right, for the hero should in some manner baffle or discomfit the chal- lenger, as the other versions show. It is therefore fortunate that Mr. J. G. Camp- bell has put on record an additional incident, apparently from another reciter (Scottish Celtic Review, I, 141) : " It is an addition to the tale that the one who imposed upon the Son of the King of Ireland the task of finding out how the great Tuairisgeul was put to death, and over whose place of decay and disap- pearance the King's son — by his wife's instructions — recounted . . . the man- ner of the Giant's death, was himself a son of the Great Tuairisgeul, and that as the story was being told he gradually rose out of the ground. Also, by the wife's instructions, his head was cut off before he got entirely clear of the ground, for then no one could withstand the young Giant's prowess." This may per- haps be taken as evidence that the Sword of Light was once present in S (as in LKJHO'FCi). With S should be compared the beginning and the end of Mac- Cool, Ceadach Og, and the Fish-Hag (Curtin, Hero-Tales, pp. 463 ff.). Here Fin loses a game of chess to the Fish- Hag. She says to him : " I place you under sentence of weighty druidic spells not to eat two meals off the one table, nor to sleep two nights in the one bed, nor to pass out by the door through which you came in, till you bring me the head of the Red Ox, and an account of what took the eye from the Doleful Knight of the Island, and how he lost speech and laughter." Fin then places the hag under spells " to stand on the top of that gable, ... to have a sheaf of oats fixed on the gable beyond you, and to have no earthly food while I 'm gone, except what the wind will blow through the eye of it needle fixed in front of you." When Fin returns, he finds the hag alive. She 222 G. L. Kit tr edge. This ends our discussion of the frame-story in I. We have found that the greater complication of I in this matter is not due to the loss of material in G, but to the inclusion in I of extraneous material which was not in y. VIII. THE DEFENCE OF THE CHILD. One of the main peculiarities of the group GI is its inclusion of an episode (not found in B or M) which we may call The Defence of the Child. This episode must have stood, in some form, in y, but not in X. It is a combination of two distinct tales, both of which exist, in many versions, independently of The Werewolf's Tale: (i) the exem- plary anecdote of The Faithful Dog and (2) the wild narrative of The Hand and the Child. The Faithful Dog is best known to English readers through the Hon. W. R. Spencer's poem, Beth Gilert, or the Grave of the Grey- hound (written in 1800), which localizes the adventure at the Welsh village of Bedd Gelert. This localization, however, cannot much antedate Spencer's poem. The tale occurs in The Seven Sages and asks for the head, which he refuses to give her : " If I was bound to bring it, I was not bound to give it." On hearing this answer "the hag dropped to the earth, and became a few bones." Another version of this same story forms the second adventure in Curtin's Fin MacCumhail, the Seven Brothers, and the King of France (Curtin, Myths and Folk-Lore of Ireland, pp. 270 ff.). Fin's task is to bring " the head of Curucha na Gras and the sword [note this !] that guards his castle." He dooms the hag to fast under conditions similar to those just described. A companion tells Fin how to act. He is not to give the head and the sword to the hag, but only to show them to her. When she opens her mouth with delight, he is to strike her on the breast with the head. This is done and the hag falls dead. The first adventure in the tale has nothing to do with the second, though the two are artificially connected at the end ; it is a version of The Hand and the Child snA will be discussed presently (no. 5, p. 223, below). With the sentence passed on the hag by Fin cf. Curtin, Hero-Tales, p. 493. For the counter-spell imposed by the quester, see p. 255, note 3. Other cases of quibbling as to the fulfilment of conditions may be seen in Larminie, p. 205 ; Hyde-Dottin, An Sgialuidhe Gaedhealach, p. 41. It is a common device in popular fiction. With the gradual rising of the dead man from the ground in S cf. Miss Dempster, Folk-Lore of Sutherlandshire, Folk-Lore Journal, VI, 160-1. Arthur and Gorlagon. 223 the Anglo-Latin and Middle English versions of the Gesta Romanorum, is extant in various Oriental forms (in the Kalllah wa Dimnah, the Hitopade^a, the Pancatantra, and elsewhere), and is commonly regarded as of Eastern (perhaps Buddhistic) origin.' It is briefly as follows : A favorite animal (weasel, ichneumon, dog) protects its master's child from the attack of a serpent or wolf and slays the assailant. The master, returning to his house, is met by the faithful creature, which is covered with blood, and, rashly assuming that it has destroyed the child, he kills it on the spot. Entering the chamber, the master finds his child safe and sound and discovers the dead body of the monster. Too late he repents of his hasty act.^ The second story, which I have called The Hand and the Child, is much more elaborate. We may first consider a group of six Celtic versions (nos. 1-6) which ascribe the adventure to Finn and are manifestly variants of a single highly elaborated tale. These are : — (i) MacDougall, Folk and Hero Tales, no. i, pp. i ff. {How Finn Kept his Children for the Big Young Hero of the Ship, and how Bran was Found) ; (2) J. G. Campbell, The Fians, pp. 204 ff. (How Fionn found Bran); (3) M.iiclnnes, Folk and Hero Tales, no. 2, pp. 32 ff. {Feunn Mac Ciiail and the Bent Grey Lad) ' ; (4) Kennedy, Legend- ary Fictions of the Lrish Celts, pp. 227 ff. {Beanriogain na Sciana Breaca^) ; (5) Curtin, Myths and Folk-Lore of Lr eland, pp. 270 ff. 1 See Benfey, Pantschatantra, Einleitung, § 201 ; Baring-Gould, Curious Myths of the Middle Ages, pp. 134 ff. ; Clouston, Popular Tales and Fictions, II, 166 ff. ; id., A Group of Eastern Romances and Stories, pp. 206 ff., 509-510, 513 ff . ; id., Book of Sindibdd, pp. 56 ff., 236 ff. ; Knowles, Folk-Tales of Kashmir, pp. 429- 430; Jacobs, Celtic Fairy Tales, pp. 259 ff. ; D. E. Jenkins, Bedd Gelert, its Facts, Fancies, and Folk-Lore, Portmadoc, 1899, pp. 56 ff. (cf. P. H. Emerson, Welsh Fairy Tales, pp. 19 ff. ; Rhys, Celtic Folklore, II, 567 ; Frazer, Pausanias, V, 421-2). 2 I have used the simpler form of the story, omitting the elaborations found in the Occidental Seven Sages. For further particulars see Additional Note, p. 269. 8 The tale consists of two parts, originally separate stories, which we may call (i) The Bent Grey Lad, and (2) The King's Children. Only the second part concerns us. *"The Queen with the Speckled Dagger"; or, "The Queen of the Many- Colored Bedchamber." From a MS. 224 G. L. Kittredge. {Fin Mac Cumhail, the Seven Brothers and the King of France^ ' ; (6) Curtin, Hero-Tales of Ireland, pp. 438 ff. (Fin Mac Cool, the Three Giants, and the Small Men)? MacDougall's version (no. i)' is here summarized : A Big Young Hero sails to shore and salutes Finn. He has been losing his children, he says, and it has been told him that there is not a man in the world who can keep them for him but Finn. He lays crosses and spells on Finn to be with him before eating, drinking, or sleeping. Thereupon he departs in his ship, leaving Finn ignorant of his abode. Finn walks along the shore and soon falls in with seven skilful companions : a Carpenter, a Tracker, a Gripper, a Climber, a Thief, a Listener, a Marksman. He takes them all into his service. The Carpenter makes a ship by striking an alder-stock thrice with his axe. The Tracker guides Finn across the sea to the house of the Big Young Hero. Finn lets his seven companions sleep and watches with the Hero's wife, who is about to be delivered of her third child. The first two have been taken away as soon as they were born by a great hand that came down the chimney. Finn keeps himself awake by means of a hot bar of iron. About midnight the child is born and the Hand descends. The Gripper seizes the hand, and after a severe tussle pulls it off at the shoulder. " But the big giant outside put in the other hand, and took the child with him in the cap of the hand." At daybreak Finn and his seven comrades give chase in the ship. That night they come to a rock in the sea, on which stands a castle thatched with eelskins. The door is in the top of the castle. The Climber scales the roof and sees a sleeping giant within, having an infant asleep in the cap of his hand. There are two boys playing shinty on the floor. By the fire lies a great deer-hound bitch suckling two pups. The Climber then carries the Thief up to the door. The Thief enters the castle, and hands 1 Curtin 's tale has a second part, — an adventure of Fin with a hag. It has nothing to do with The Hand and the Child, but is attached to it by making the hag the giant's sister. She apparently comes for vengeance on Fin, though this is not brought out, and indeed is contradicted by something in the second part. The continuation is interesting in connection with the episode of Beowulf and Grendel's Mother and its motivation. 2 This consists of two quite independent stories, loosely attached. The first alone is to our present purpose. " MacDougall (p. 259) notes that the tale was known to two Highlanders of his acquaintance besides the one from whose recitation he derived it. Arthur and Gorlagon. 225 out the baby, the two boys, and the pups,^ and escapes without waking the giant. Finn puts to sea. Soon the Listener hears the giant awake, and send the bitch in pursuit. They throw a pup to the bitch, who returns to the rock with it. Soon after the giant himself appears, wading through the sea. Finn puts his finger under his " knowledge-set of teeth " and finds that the giant is " immortal, except in a mole that [is] in the hollow of his palm." This the Marksman hits, and the giant falls dead.^ Now they sail back to the giant's castle, and the Thief steals both pups. Returning to the home of the Big Young Hero, Finn restores the three children to their parents, asking no reward except one of the pups. This grew up to be Finn's dog Bran, so famous in Fenian saga. There is a feast for a year and a day [after which we may infer that Finn returns to Erin]. No. 2 (J. G. Campbell's version) corrects No. 1 in certain details. The sleepiness of the watchers is caused by magical music,' — a familiar feature in Celtic story. The giant* leaves his arm behind (which is not expressly stated in No. i).^ There-is but one visit to 1 He also steals " the silk covering that was over the giant and the satin cover- ing that was under him," — a familiar trick of the Master Thief. 2 Cf. MacDougall, pp. i6o-i6i. ^ So also in 3. In 3 (as in l) it is Feunn alone who keeps awake (by holding a hot poker under his chin : good folk-lore !), and he rouses Firm-Holder at the moment of peril. In 6 nobody is sleepy. In 5 Finn goes to sleep deliberately and the Skilful Companions watch ; so in 4, except that Finn's sleep is druidic. In 2 all are kept awake by one of the Skilful Companions, whose specialty is that he never sleeps : this reminds us of the Old French proverb : " Qui ne dort pas, n'est pas d'ome" (see Lai de Tydorel, Rom., VIII, 67). For soporific music see Child, Ballads, I, 55; II, 137, 139 f., 511 f.; IV, 18 ff. ; V, 220, 293; add Hyde-Dottin, Sgialuidhe Gaedhealach, pp. 188-189. * The robber is a giant in i, 2, 5 ; a hag in 4, 6; 3 is indeterminate. Clearly he (or she) was originally a Water-monster of some kind : in 1 (cf. 2) the giant's castle is on a rock in the sea, is to be entered only at the top, and is thatched with eelskins. In 4 the hag inhabits a whirling castle, which is reached by boat and has its entrance in the top. Compare the subaqueous abode of Grendel and his mother in Biowulf. In 2 and 6 the giant (hag, 6) has but one eye (in the forehead) and is killed by an arrow which pierces this eye. In i the giant is " immortal except in a mole that was in the hollow of his palm." On whirling castles see A.C.L. Brown, Iwain, pp. 80-81, above. 5 In 5 the robber leaves both his arm and the child. This is probably correct (see p. 227). In 2 and 6 the hand is left but the child is taken ; in 4 both hand and child disappear. 226 G. L. Kittredge. the giant's castle. Three pups are taken ; two are thrown to the pursuing bitch, the third is saved. Three must be right ; it makes the number of the pups correspond to that of the children.^ This point may turn out to be of some significance.'^ Nos. 3-6 make no mention of the bitch and her pups. No. 3 is incomplete, lacking the visit to the giant's castle in the sea. We at once recognize this story as a composite. It has assim- ilated nearly the whole of a widespread mdrchen known as The Skilful Companions, which has been studied by Benfey and other distinguished scholars,^ and which has nothing whatever to do with The Hand and the Child. In The Skilful Companions — Three or more brothers (or comrades) are suitors for the hand of a beautiful girl. While her father is deliberating, the girl disappears. The companions undertake to recover her. One of them, by contemplation (or by keenness of sight), finds that she has been stolen by a demon (or dragon) and taken to his abode on a rock in the sea. Another builds a ship by his magic (or possesses a magic ship) which instantly transports them to the rock. Another, who is a skilful climber, ascends the castle and finds that the monster is asleep with his head in the maiden's lap.* Another, a mas- ter thief, steals the girl without waking her captor. They embark, but are pursued by the monster. One of the companions, an unerring shot, kills the pursuer with an arrow. The girl is restored to her parents. We are not here concerned with the origin or history of The Skil- ful Companions, which, as every one knows, is a corner-stone of ^ In 3 and 5, three children have already been lost, making four in all ; but this can hardly be right. In i , 2, 4, 6, the whole number is three. ^ Cf . pp. 238-9. * See Benfey, Das Mdrchen von den " Menschen mit den wunderbaren Eigen- schaften," Ausland, 1858, pp. 969 ff. {Kleinere Schriften, II, iii, 94 ff.) ; Wes- selofsky, in Giovanni da Prate, II Paradise degli Alberti, 1867, I, ii, 238 ff. ; d'Ancona, Siudj di Critica e Storia Letteraria, 1880, pp. 357-358 ; Kohler-Bolte, Ztsch. des Ver. f. Volkskunde, VI, 77 ; Kbhler, Kleinere Schriften, I, 192 ff., 298 ff., 389-390, 431, 544; II, 591; Cosquin, Contes pop. de Lorraine, I, 23 ff.; Crane, Italian Popular Tales, p. 67 ; Nutt, in Maclnnes, Folk and Hero Tales, pp. 445 ff. ; Laistner, Rdtsel der Sphinx, II, 357 ff. ; Steel, Tales of the Punjab, pp. 42 ff. ; Jurkschat, Litauische Mdrchen, pp. 29 ff. ; etc. * The number and functions of the skilful companions differ considerably in the several versions. The climber, in particular, is by no means a constant quantity. Arthur and Gorlagon. 227 Benfey's theory of Oriental origins. The story is found in the East and, in varying forms, in almost every country in Europe. Its iden- tity with a considerable portion of The Hand and the Child in the Highland versions which we are studying is evident. To reduce The Hand and the Child, therefore, to something that approaches its original condition we must first of all eliminate those incidents which belong to The Skilful Companions. Such an elimination leaves the following plot : A certain king has already lost two children, who have been carried off as soon as they were born. [Apparently no one knows what has become of them, for all the watchers are overcome with sleep.] The queen is expecting a third child.^ A hero of extraordinary strength visits the king [perhaps by invitation], and undertakes to watch. The child is born. The hero resists the soporific magic, to which all others yield, grasps the gigantic hand that descends through the smoke-hole (or window) to seize the child, and tears it off at the shoulder. The monster escapes, leaving behind the child and the arm. The Hand and the Child belongs, obviously enough, to the type of which the adventure of Bdowulf with Grendel is the most famous representative.^ The similarities are striking ; but, before one infers 1 Possibly we should omit the two children previously lost ; but it seems likely that the ravages of the monster had lasted for some time before he was finally checkmated. We have a good parallel in the Beowulf, in which Grendel has carried off and devoured many of HrotSgar's men before Beowulf undertakes the defence of the hall Heorot and pulls off the monster's arm. See also the stories from Cashmere and California and compare the Japanese legend (p. 228, below). * See Herrig's Archiv, CIII, 154, where Professor Cook notes the similarity between Biowulf and Kennedy's version (our no. 4). Zimmer (Haupt's Ztsch., XXXII, 331-332) detects the influence of Beowulf's encounter with Grendel in Cuchulinn's combat at Curoi's fort in the Fled Bricrend (cf. Andler's strange book, Quid ad Fabulas Heroicas Germ. Hiherni contulerint, pp. 75-76), but I find it impossible to agree with him. Fergus's fight with the sea-monster (Senchtis Mor, Anc. Laws of Ireland, I, 74-75) or Cuchulinn's feat of swimming (Siabur- charpat Coinculaind, or Phantom Chariot of Cuchulinn, Hull, Cuchullin Saga, pp. 284-285 ; cf. Haupt's Ztsch., XXXII, 250, 254) would have afforded him an equally striking parallel. Resemblances between Blowulf and the Icelandic Grettissaga (Grettir cuts off a monster's arm, etc., etc.) were observed by Vig- f usson (Sturlunga Saga, Prolegomena, I, xlix ; Icelandic Reader, p. 404 ; Corpus 228 G. L. Kittredge. historical or literary connection between the Celtic tale and the Anglo-Saxon epic, there are several phenomena to be reckoned with. The child-stealing motive is no part of the Bkowulf, nor of a Japanese legend ^ which resembles Bkowulf in the most striking Poeticum Boreale, II, 501 ff.). That distinguished scholar held that the author of the saga knew the Biowulf, and his opinion has met with some favor, but the case is by no means dear (see Gering, Anglia, III, 74 ff.; Garnett, Amer.Jottrn. of Philol., I, 492 ; Bugge, Paul u. Braune's Beitrdge, XII, 57 ff. ; ten Brink, Beowulf p. 185; Symons, in Paul's Grundriss, I, 21 ; 2d ed.. Ill, 649; Laistner, Ratsel der Sphinx, II, 27 ff; Boer, Ztsch.f. deutsche Phil., XXX, i ff. ; Jonsson, Den Oldnorske og Oldisl. Litteraturs Historic, II, 751, note). The story of The Hand and the Child reappears in a modem Icelandic mdrchen (Arnason, Islenzkar pJ6&sogur og ACfintj/ri, II, 471 ff., translated by Poestion, Isldndische Mdrchen, pp. 285 ff.). The Icelandic version is strikingly similar to the tale as it occurs in L (including the pretended leeching of the monster) and is doubtless derived from Irish (or Scottish Gaelic). The watcher resists the soporific magic and cuts off the kidnapper's arm. Laistner, who compares Poes- tion's translation with Biowulf {Rdtsel der Sphinx, II, 26 ff.), has not observed that the part of the tale which coincides with The Skilful Companions must be left out of account. 1 " At the beginning of the eleventh century, when Ichijd the Second was Emperor, lived the hero Yorimitsu. Now it came to pass that in those days the people of Kiy6to were sorely troubled by an evil spirit, which took up its abode near the Rash& gate. One night, as Yorimitsu was making merry with his retainers, he said, ' Who dares go and defy the demon of the Rashd gate, and set up a token that he has been there ? ' ' That dare I,' answered Tsuna, who, having donned his coat of mail, mounted his horse, and rode out through the dark bleak night to the Rashd gate. Having written his name upon the gate, he was about to turn homewards when his horse began to shiver with fear, and a. huge hand coming forth from the gate seized the back of the knight's helmet. Tsuna, nothing daunted, struggled to get free, but in vain, so drawing his sword he cut off the demon's arm, and the spirit with a howl fled into the night. But Tsuna carried home the arm in triumph, and locked it up in a box. One night the demon, having taken the shape of Tsuna's aunt, came to him and said, ' I pray thee show me the arm of the fiend.' Tsuna answered, ' I have shown it to no man, and yet to thee I will show it.' So he brought forth the box and opened it, when suddenly a black cloud shrouded the figure of the supposed aunt, and the demon, having regained its arm, disappeared." Mitford, Tales of Old Japan, ed. of 1890, p. 105. Professor York Powell gives the same story, in outline (from the vulgate version in " the Japanese children's picture-books of this century, and the colour-prints by Hokusai" and others) and compares it with Biowulf: see his note in An English Miscellany presented to Dr. Furnivall, 1901, pp. 395-396. Arthur and Gorlagon. 229 "way, nor of an episode in the Perceval which should also be com- pared.'' Per contra, there is a story from Cashmere which resem- bles The Hand and the Child in the matter of the child-stealing, but in which the ogress, though overpowered, does not lose her arm.'' Finally, the loss of the hand and the stealing of the child occur, 1 The Demon Hand is found in the second continuation of Chretien's Perceval (by Gaucher de Dourdan). Perceval enters a solitary chapel at night. There is no one in the chapel, but a slain knight is lying on the altar. One candle is burning before him. Suddenly a great light (clartf) fills the chapel, and as suddenly disappears. A crash {escrois) follows, as if the chapel were falling to pieces (vv. 34,434-469, ed. Potvin, IV, 133-134). Then Une noire mains jusqu'al couste S'aparut derrifere I'autel; La candoile ki ardoit cler Estaint ensi c'on n'i vit goute (vv. 34,470-473). Perceval leaves the chapel in haste. A lame explanation of these phenomena is given (in the conclusion written by Mennecier) by the Roi Pesceor. The chapel was built by Brangemore of Cornwall, mother of King Pinogr^s. She became a nun and was beheaded therein by her cruel son. She was buried under the altar, and since then not a day has passed without a knight's being killed there by the Black Hand ; more than four thousand have lost their lives (vv. 35,397 ff., IV, 166 ff.). Later Perceval visits the chapel again and has a terrific struggle with the Black Hand, which comes in through a window. He overcomes the devil to whom it belongs, not with the sword, which is powerless against him (cf. Grendel), but by means of the sign of the cross (vv. 39,790 ff.,.IV, 304 ff.). Apparently we have here the story of the Demon Hand worked over in a Christian sense. In view of the wide currency of the incident, it would be venturesome to ascribe this partic- ular example of it to a Celtic source; but, since the incident does occur in Celtic, it would be equally credulous to deny the possibility of such a derivation. Of course nobody will hold that Gaucher drew from Celtic directly. The fight with the hand, we should observe, is Mennecier's contribution. Did he know the whole story, left incomplete by Gaucher, or was he simply inventing a dinouement? 2 In The Tale of a Princess, Knowles, Folk-Tales of Kashmir, p. 59: A princess, disguised as a man, entered the service of a merchant. " This merchant had three wives, but no son. The reason of this was, that the night after any of his wives gave birth to a son a ddgin [ogress] appeared and devoured it." A son was born to the merchant. The merchant asked his new servant to watch by the bedroom door and ward off the ogress. The ddgin tried to burst open the door, but the servant prevented her, whereupon she made a dash at him. The servant seized her by the hair and threw her down, but spared her life on her promising to trouble that house no more. 230 G. L. Kittredge. in combination, in a North American Indian tale from California "^ and in the Welsh mabinogi of Pwyll, to which we shall presently return.^ We must put behind us the temptation to genealogize. One fact is clear : the defence of a hall or a hut against the demon that haunts it is a simple theme, to which the theory of " independ- ent origins " must apply if it ever applies to anything. That the defence should result in the demon's losing his arm seems a not unnatural development : at all events, this feature is found in Ireland, in Wales, in England, in Japan, and in California.' The other main element in our story — the kidnapping of the children — is too com- monplace to make any trouble. All manner of uncanny beings are charged with carrying off infants, and everybody knows that the moment of birth, like the moment of death, is a mysterious time and full of strange peril from the darker powers. The genesis of The Hand and the Child, then, is not hard to conjecture. It is an easy combi- nation of two motifs, (i) the Defence of the Hall and (2) the Child- stealing Monster, to which (in the Highland tales summarized above, pp. 223 ff.) other familiar bits of folk-lore (the Skilful Companions, for instance, and the One-eyed Giant *) have associated 1 Curtin, Hero-Tales of Ireland, p. 558, gives part of an Indian tale from Cali- fornia in which a supernatural hag is in the habit of stealing children. She reaches down through the smoke-hole to take one ; five or six men seize her arm and try to pull her down, but in vain. " One man chopped her arm right off with a flint knife, and threw it out ; she fell to the ground where her arm was, she picked it up, and ran home." 2 ggg pp^ 240 ff. , 8 We may compare also the cutting off of the ghoul's leg in Swynnerton, Indian Nights' Entertainment, pp. 358-359. The house-haunting goblin m/ataka, ii, 155 (Cowell, II, 12), is subdued in a more recondite manner. So is the hand that rises from the sea and steals men in the Peregrinaggio di tre Giovani, Figliuoli del Re di Serendifpo, ed. Gassner, Erlanger Beitrdge, X, 23-24, 28 ff.; cf. the parallels cited by Huth, Zt. f. vergl. Litteraturgesch., N.F., Ill, 313-314. Cf. also the Demon Hand in Miss Burne, Shropshire Folk-Lore, p. 113. In a Greek mdrchen (Hahn, Griech. u. alban. Mdrchen, II, 50) a Hand robs the king's apple-tree ; the prince shoots into a cloud and draws blood (cf. Cosquin, Contes pop. de Lorraine, I, 12). * It is an easy process to derive from the Odyssey all monocular giants who meet the fate of Polyphemus ; but such hand-to-mouth methods are more danger- ous than they seem. See Laistner's interesting chapter on Polyphemus (Rdtsel der Sphinx, II, i ff.). Arthur and Gorlagon. 231 themselves. The whole, in a highly elaborated form, has become a part of the Finn cycle, and is used to explain how Finn procured his famous dog Bran. The story of The Hand and the Child is doubtless quite independent of The Faithful Dog. Indeed, the tales differ from each other in almost every respect ; they show but one element in common : the successful defence of an infant. In The Faithful Dog, however, the assailant is not a hobgoblin, but a natural creature (wolf or serpent) ; the defender is not a hero, but an animal (ichneumon, weasel, dog), which fights with the beast and kills it in accordance with common sense and everyday experience ; the danger is unforeseen (not watched for, as in The Demon Hand). Finally, the central point of The Faithful Dog — the fatal mistake, the overhasty judgment which prompts the master to strike down his friend and benefactor — is necessarily wanting in The Hand and the Child. The Faithful Dog is an exemplum, enforcing the danger of precipitate judgments ; its motto might well be King Lear's " Woe that too late repents ! " The Hand and the Child has no moral and is hardly susceptible of one, even at the hands of the melancholy Jaques. Yet nothing was easier than for these two stories to come together. Their common element — the defence of the baby in the cradle against some hideous danger — was almost certain to unite them sooner or later.^ Accordingly they do, in fact, combine to produce an incident somewhat different from either, yet preserving plain traces of both. In this incident an animal defends thebaby from the giant that seeks to steal it, biting off the hand which he stretches into the room ; the animal is accused of killing the baby, but is exonerated.^ 1 A Mongolian version of The Faithful Dog (Benjamin Bergmann, Nomadische Sireifereien, I, 102, cited by Benfey, Pant. 1,481) approaches the type of The Hand and the Child in a curious way. A woman has had several children but has lost them all. She is again with child when a polecat (litis) comes to her and promises that she shall lose no more children if she will take him into her service. The mother thinks the talking polecat must have magical powers, and assents. The animal defends the baby from a snake and is killed by the mother. ^ Perhaps there was aversion of The Hand and the Child va which the defence of the child against the Demon Hand was transferred to a dog (a fairy dog, it may be, or a bespelled mortal) before The Hand and the Child came into contact with 232 G. L. Kittredge. In some such form as this, the incident has entered The Werewolf's Tale. It is not found, as we have already seen, in Marie's Bisdavret or in the Lai de Melion, but its presence in I and G proves it for y The Faithful Dog (the " Gelert story"). Such a version, if it ever existed, would easily have become contaminated with The Faithful Dog, and the resultant tale would with equal facility have entered version y of The Werewolf's Tale. These are details that cannot be determined and that do not affect the essentials of our reconstruction. We may note, however, that in an incomparably wild Highland tale a (fairy) dog does actually defend his master, in a cave at night, against a monster that reaches for him through a hole in the roof, and that the monster's arm is bitten off at the wrist. This is the tale of Mac PMC's Black Dog, taken down by J. G. Campbell from recitation in 1863 and published (with an English translation) in the Scottish Celtic Review, pp. 262 ff. A revised translation is printed in the same writer's posthumous work, Superstitions of the Highlands and the Islands of Scotland (Glasgow, 1900), pp. 109 ff. (with four other versions, all from oral tradition). I give a bare outline, which does scant justice to the impressiveness of this extraordinary story. Mac Phie of Colonsay owns a great black dog, presented to him under strange circumstances, which, according to the prophecy of the giver (obviously a fairy man), "will never do service for him but the one day." The dog always skulks when his master calls him to the hunt, and Mac Phie has often been urged to kill him. "Let him alone," is Mac Phie's reply; "the black dog's day will come." One morning, when Mac Phie and other gentlemen are setting out for Jura to hunt, the dog is the first creature in the boat. " The black dog's day is drawing near us," says Mac Phie. On the second night of their excursion, when they are all together in a great cave in Jura, Mac Phie's companions are destroyed by certain ghoulish women [lamiae, or lustful demons, we may be sure : cf. a Suther- land tale communicated by Miss Dempster, Folk-Lore Journal, VI, 162-163^, but the black dog, who lies at his master's feet, springs up when one of the women would approach Mac Phie, and drives them from the cave. Soon a man's hand comes down through a hole in the roof and clutches at Mac Phie. What followed must be given in Mr. Campbell's own words : " The black dog gave one spring, and caught the hand between the shoulder and the elbow, and lay on it. The play began between the hand and the black dog. Before the black dog let go his hold, he chewed the hand till it fell on the floor. The thing that was on the top of the cave went away. . . . Out rushed the black dog after the thing that was outside. This was not [the] time at which Mac Phie felt himself most at ease, when the black dog left him. When the day was dawning, what but that the black dog had returned. He lay down beside Mac Phie. In a few minutes he was dead." Mac Phie took the hand home " that men might see what horror he had met with that night he had been in the cave. No man in Isla[y] or Colonsay had ever seen such a hand, or had ever imagined that such could have existed." A rthur and Gorlagon. 233 (their common original),i which we have seen reason to believe was Irish. '^ The precise form of the episode in y is not easy to determine, but we may come pretty near it by a process of comparison. Let us begin with the condition of the episode in I. In LHO'FS we find practically the whole of The Hand and the Child (as described on the basis of nos. 1-6, pp. 223 ff., above), modified by two features from The Faithful Dog: (i) the substitution of the tame werewolf for the hero, and (2) the suspicion against the animal. We may take L as the basis of our comparison with Gorlagon (G), since, though it is somewhat disordered, it preserves a number of highly significant details. In L the king who befriends the Werewolf had lost eleven = children, all of whom " were stolen the same night they were born." He sets the wolf to watch the twelfth. One night ^ a hand comes down the chimney and seizes the child. The wolf bites off the hand,^ lays it in the cradle with the baby, and falls asleep. In the morning both hand and child are gone. The wolf is covered with blood, and everybody " says the wolf has eaten the baby. But the king refuses to believe this.'' " Loose him," says the king, " and he will get the pursuit himself." 1 See the diagram on p. 175. 2 See p. igS. 2 Plainly an exaggeration of the reciter; two is the correct number, as in HS {and The Hand and the Child in general, see p. 226, note i) ; in O'F it is three. Ci says nothing of the king's previous losses. C2 lacks the whole adventure. * This should properly be the birth night (as in O'F). Nobody knows what has become of the other children (implied in L, expressly stated in H). In H the nurses are put to sleep by a magic song when the third child is stolen (cf. p. 225, above). In S the wolf is present on all three occasions, but apparently he is awake on the third only; the midwives sleep. In KJ there is no Hand; the lady smears her own sleeping child and the wolf with blood and then accuses the wolf. ^ In HO'F he pulls it off (cf. p. 227). In Ci a serpent comes down the chimney ■ and is killed by the wolf. Thus this particular version reverts in part (whether by accident, or by specific modem influence) to the Oriental form of the Gelert story. ^ The specific accusation should come from the Werewolf's wife (so KJHO'FCi). In S the midwives are the accusers (cf. Pwyll, p. 240, below) on the first two occa- sions ; but this version has substituted a cruel stepmother for the unfaithful wife. On the third occasion in S the wolf pursues the monster without delay and there is no opportunity for slander. Ca has been too much changed to be of much use here. ' In H the king credits the accusation, but is undeceived by the discovery of the hand. In Ci the disenchantment comes immediately after the false charge. 234 G. L. Kittredge. [The werewolf's false wife ^ has concealed the child and the hand in a (secret) room.^] The wolf follows the scent of the blood to the door of this room, goes back to the king, takes hold of him, and then, returning to the door, begins to tear at it. The king follows,' and calls for the key. A serv- ant says it is in the room of the stranger woman [i.e., the Werewolf's wife]. She cannot be found, and the king breaks down the door. The wolf runs in and goes to the trunk. The king breaks the lock of the trunk : there lie the child and the hand, side by side, and the child is asleep. Here we must pause a moment to compare the Gorlagon (G). In G the incident has been considerably changed by the general modification which the tale has received at this point : * Instead of defending the child against a giant or hag, the wolf assails the king's steward, who is dishonoring the royal bed, and mangles him frightfully. The queen removes her child to an underground room, and accounts for all the circumstances by alleging that the wolf has devoured it and that the steward has been wounded in opposing the wolf. G, we observe, omits the Demon Hand and inserts an amour between the queen and her steward : the wolf does not defend the baby against an assailant ; he attacks the queen's lover, out of 1 L does not explain how she came to be at the court, but we have already seen that she is really the king's daughter (a point which L has not preserved, but which is assured for I, being found in KJO'FCi) and has returned to her father after betray- ing her husband. Indeed, this relationship (as well as the return) is present also in M, and is thus established f or x (the common source of My) : see p. 1 78. In H the whole scene is laid at the castle of the Werewolf's father, and the stolen chil- dren are the Werewolf's brothers. 2 This is implied in L, and comparison with G establishes the incident for y. O'F, though somewhat confused here, supports L in the main. In HS the giant carries off the child but leaves the hand behind (cf. p. 225). HCi preserve an important link in the story : the lady wakes first in the morning and finds the hand (the serpent and the child Ci) ; thus she is enabled to arrange the details of her plot before the household is stirring. In H she buries the hand in the woods ; in Ci she hides the child in her chamber. Taken together, then, HCi support L, as O'F does, and the course of events in I can be made out perfectly. 8 In H the Werewolf leads the king to the place where the lady has buried the hand. In Ci he conducts him to the chamber where the child is hidden. O'FS lack the incident. * See p. 185, above. Arthur and Gorlagon. 235 loyalty to his master. These features are peculiar to G, and it is clear that in them G departs from y.^ To continue our analysis of G : The king refuses to believe in the animal's guilt. The wolf touches the king's foot with his paw, seizes the edge of his mantle in his mouth, and nods his head in sign that he wishes the king to follow him. He leads the king to the underground chamber where the child is concealed, and strikes the door with his paw. The queen has hidden the key, but the wolf, impa- tient at the delay, breaks the door down, and, rushing into the room, brings out the child and presents it to the king. He then leads the king to the chamber where the steward lies, and the guilty man confesses the truth. The similarity in detail between L and G is most striking, and is highly significant as to y. We must now return to L : After the rescue of the child, the wolf is its constant companion. One day, the child, then three years old, runs away from home and cannot be found. [It transpires, later, that the Werewolf's wife has him at her house.] When summer comes, the wolf swims back to his own country and hides in his own garden. He sees his wife out walking, and the child with her. Next day, the wolf enters the house and finds the child alone. The boy recognizes his old favorite and begins to kiss him. The magic rod is " in front of the chimney." The wolf jumps at it and knocks it down. The child picks it up. Then the wolf scratches the child, and the boy, in anger, strikes him a light blow with the rod and thus restores him to his human shape. The Werewolf (now a man again) takes the child back to the king in a ship. On the way, he comes to an island, where there is but one habi- tation. Entering, he finds a frightful hag. Her son lies groaning in an inner room. " His hand," says the hag, " was bitten off, twelve years before, in another land." The hero pretends to be a physician, shuts himself up with the hag's son, and burns out his eye (he has but one, in the middle of his forehead)'' with a hot iron, pretending that he wishes to cauterize the corrupt flesh. The deluded hag gives the hero the reward she has promised, — eight lads and three girls, who, she informs him, are 1 This jappears at once from comparison. Positive evidence that the inference is correct will be given later, when we discuss the punishment of the Werewolf's wife in G (see pp. 245 ff). ^ See p. 230 and note 4. 236 G. L. Kit tr edge. the sons and daughters of the king and have all been stolen by her son. The hero takes ship, returns to the king's court, and gives him back the children.^ There is nothing of all this in G. It is peculiar to I, and will be instantly recognized as the concluding adventure in The Hand and the Child (see nos. 1-6, pp. 223 ff.), modified so as to fit it to the exi- gencies of The Werewolf's Tale and, in particular, so as to bring about the disenchantment of the hero. That this adventure was not in y (the common source of G and I) is at once clear. In G the resto- ration of the Werewolf is effected in a very different way, and G is, in this part of the story, in substantial agreement with M and B. Hence we may be sure that the rescue of the king's other children (and probably also the incident of their loss) was not in y, and a fortiori not in x. It was not added to the story until G and I had parted company. It is now easy to reconstruct the episode of The Rescue of the Child in substantially the form which it must have had in y (the common original of G and I) : The scene is laid at the court of a king, the wolf's father-in-law, whither the false wife has fled after the transformation of her husband. She wishes to get rid of the wolf, whom she recognizes and of whom she is very naturally afraid. The wolf defends the king's child and bites off the monster's arm ; the monster flees, leaving his hand behind him, and is heard of no more. The false wife takes advantage of the situation to remove the hand and the child to a secret chamber, and accuses the wolf of devouring the infant. The king refuses to believe the charge, and the wolf leads him to the secret room. Several of the details of the scene may be inferred from the wonderful agreement between G and L : the wolf's seizing the king's robe in his teeth and guiding him to the room ; 1 The different versions of I show considerable variety in details in this part of the story, but LHO'F agree in the main. LH have the pretended medical or surgical treatment of the monster. LO'F show Polyphemus incidents (putting out the eye LO'F ; dressing in goatskins O'F). By a special elaboration, H makes the Werewolf get the Sword of Light in the giant's island. S is much condensed here. KJCi of course lack the rescue of the elder children, since they say nothing of the king's having lost his sons. In LH the rescue follows the Werewolf's disenchantment ; in O'FS the disenchantment follows the rescue. Arthur and Gorlagon. 237 the locked door ; the concealment of the key by the lady ; the breaking down of the door (by the king in L, by the wolf in G). In the room is found a chest (or cradle), in which the child lies sleeping ; the hand is with him. The king is convinced that the wolf is a man under spells, compels his daughter to confess, and reverses the charm. This reconstruction, every detail of which is extant either in G or in I, is a manifest compound, formed, as I have already suggested, by uniting The Hand and the Child and The Faithful Dog. AH the divergences which G shows from the incident as thus reconstructed are accounted for. Four versions of I (LHO'FS) contain, also the second adventure of The Hand and the Child, — the rescue of the king's other chil- dren from the giant's castle. Its presence is easily explained. Some story-teller, familiar with The Hand and the Child in its most developed form (substantially as in nos. i and 2, pp. 224—5), felt that version I of The Werewolf's Tale was incomplete because it did not contain this second adventure, and appended it accord- ingly. Nothing could be more natural. It was simply a case of going on. Version I already contained the Defence of the Child against the demon hand ; the narrator continued with the second adventure, the rescue of the king's other children, which seemed to him a necessary sequel. This addition to I may have been made in very recent times, — even as late as the eighteenth cen- tury. There is no certain evidence on that point. The first inser- tion of The Defence of the Child in The Werewolf ' s Tale is quite another matter. This must have taken place pretty early, since the incident stood in y. There is another Irish story which throws light on version y of The Werewolf's Tale. It is extant as an episode in The Festivities at the House of Conan, a late text edited by O'Kearney from an eighteenth-century manuscript,^ and runs as follows in O'Kearney's translation : 1 Feis Tighe Chonain Chinn-Shleibhe ; or The Festivities at the House of Conan ■of Ceann-Sleibhe, edited by N. O'Kearney from a MS. of Foran of Portlaw (1780), in the Transactions of the Ossianic Society for 1854 (Dublin, 1855), pp. 160-67. 238 G. L. Kittredge. Fionn's rhother's sister, Tuirreann, became the wife of loUann Eachtacli. She became pregnant. loUann's leannan sighe, from jealousy, transformed her into a greyhound and brought her to the house of King Feargus Fionn- liath, presenting her as a present from Fionn. " The wife of Feargus . . . gave birth to an infant the same night that the hound whelped two puppies, a male and a female. It so happened during the previous seven years, that whenever Fergus's wife was confined, a Fomorach used to come that same night, and carry away the infant. However, Eithleann [unknown person] met Fionn at the end of a year, and having arranged a hospitable meeting at the house of Feargus Fionnliath, they delivered Fergus from the plague of the Fomorach." Fionn learned that his aunt was no longer living with loUann and insisted on her being restored to him. loUann required her of his leannan sighe. She went to Feargus's House and got the bitch and restored her to human shape. She then brought her to Fionn and told of the two puppies, giving him his choice to have them as dogs or human beings. He chose the former and these are Bran and Sceolaing.i This is vague and prosaic, but it is plainly a somewhat condensed account of a version of The Hand and the Child.^ The demon hand has evaporated in the process of condensation. Instead of details, we have a bald general statement r " Having arranged a hospitable meeting at the house of Feargus Fionnliath, they delivered Feargus from the plague of the Fomorach." The bitch and her pups, of which we have already heard in several versions of The Hand and the Child, play an important, if not quite intelligible, part in the present text. One fact comes out clearly : there is a mysterious con- genital relation between the children and the pups. This we have already suspected, on the basis of the other versions and of general folk-lore. Comparing the Feis Tighe Chonain with versions i and 2 of The Hand and the Child (jpp. 224 ff.), we may infer that, in the correct 1- The same story (without the robbery of the children) may be found in Kennedy, Legendary Fictions, pp. 1 74 ff. 2 Like all the versions noted above (nos. 1-6, pp. 223 ff.) that in the Feis Tiglie Chonain has been attached to the Finn cycle, and, like nos. i and 2, it undertakes to explain " how Finn found Bran," his famous dog. Of course there is no occa- sion to suppose that the tale was connected with Bran in the beginning. There are other accounts of Finn's discovery of Bran (see MacDougall, Folk and Hero Tales, pp. 263-264). Arthur and Gorlagon. 239 form of the fully developed story,^ the bitch-hound was not at the giant's castle in the sea, but rather in the chamber where Finn watched ; that the birth of the child and the whelping of the bitch always took place at the same moment ; and that the giant stole both the baby and the whelp. This had already happened twice before, so that when the rescuers visited the giant's castle, they found three children and three dogs. ^ The Feis Tighe Chonain shows a special resemblance to The Were- wolf's Tale which we have not found in other versions of The Hand and the Child: the bitch is a transformed mortal, like the Werewolf. There is even a certain likeness in the cause of transformation. In G and I the lady changes her husband to a wolf because she is in love with another ; in the Feis, the mistress of loUann changes her lover's wife into a bitch in order to keep him for herself. Note also that lollann's mistress is a leannan sighe^ — a fairy mistress ; and that we have seen reason to regard the lady in The Werewolf's Tale as originally z.fie.'' These resemblances need not be pressed. They suffice, however, to show how easy it was for a tale like The Hand and the Child to become inserted in The Werewolf's Tale in Irish story-telling. A remarkable variant of the episode in the Feis Tighe Chonain is thus tantalizingly recorded by O'Kearney in a note : * " It is . . . recorded in tradition that she [read if] was the enchanted hound [i.e., Finn's aunt] that rescued the infant from the grasp of the giant by gnawing off his arm, and that she preserved it until morning. When Feargus and his people found the chamber, in which she kennelled, full of blood, they were on the point of killing her, under the supposition that she had murdered the child ; but they fortunately discovered their mistake in time . . . The sanje authority relates that the hound led Feargus and his people to the giant's cave, where they succeeded in killing him, and also recovered the seven children that had been previously kidnapped by him." 1 That is, the story made by combining The Hand and the Child with The Skilful Companions, — a combination seen in nos. i and 2 (see pp. 224-6). 2 Cf. J. F. Campbell, Pop. Tales of the West Highlands, II, 70. ' See pp. 176-7, 189 ff. < P. 164, note 2. I have corrected an obvious misprint. 240 G. L. Kittredge. This approaches The Werewolf's Tale still more closely, in that it is not a hero in human shape that defends the baby, but an enchanted animal. It affords positive testimony that such a version of The Hand and the Child as that inferred at p. 231 has actually existed, out of combination with The Werewolf's Tale. Good evidence of the antiquity of The Hand and the Child on Celtic soil is furnished by the mabinogi of Pwyll Prince of Dyvet, one of our most precious relics of genuine Welsh tradition. Pwyll preserves the story in a remarkable shape : ^ Rhiannon's child has just been born, and six women are watching. All six fall asleep about midnight, as well as the mother. At dawn the women awake, but the baby has disappeared. Rhiannon is still asleep. There is a bitch hound with her young in the chamber. They kill some of the puppies, smear with blood the face and hands of Rhiannon, and put some of the bones before her. When she wakes and calls for the child, they declare that she has devoured it. The nobles urge Pwyll to divorce his wife. He refuses : " If she has committed a sin, let her do penance." Rhiannon decides to accept penance rather than to dispute the question with the lying nurses. Her penance is, to remain seven years at the court, to take her seat each day beside the horse-block at the entrance, to tell her story to all comers, and to carry them on her back, if they will allow it, from the horse-block to the court." So she passes a part of the first year. There is a lord at Gwent named Teyrnon, who has a very beautiful mare. Every year she drops a foal in the night of the calends of May,* but no one knows what becomes of it. This time Teyrnon resolves to watch. The foal is born, and Teyrnon is admiring its beauty when he hears a great noise. Immediately a claw comes through the window of the house and seizes the foal by the mane. Teyrnon draws his sword and cuts off the monster's arm at the elbow, so that the forearm and the foal remain inside' the window. There is a great noise outside. Teyrnon rushes out and runs in the direction of the noise, but it is so dark that he sees nobody. Returning, he finds just outside the door a little child. Taking it up, Teyrnon goes into the house, shuts the door, and learns that 1 Mabinogion, translated by Lady Guest, III, 60 fE.; by Loth, I, 52 fit. ^ Cf. Rhys, Hibbert Lectures, p. 641 ; id., Arthurian Legend, p. 284. s Cf. Rhys, Celtic Folklore, I, 226. Arthur and Gorlagon. 241 his wife has slept through everything. Teyrnon and his wife adopt the child, and the foal is reserved for him against the time when he shall be able to ride. After a time, Teyrnon hears of what has happened to Rhiannon. He takes the child to the court and all is well. In the episode just summarized, The Hand and the Child has been modified by contamination with a story of a different type, into which it has been worked, — namely, The Calumniated Wife?- In this type " the wife is accused (usually by her mother-in-law or by a rival) of bearing an animal or a monster ^ (or of having devoured her off- spring) ; the child is spirited away (or slain) by the calumniator j the wife is repudiated or subjected to terrible peiknce ; at last the child is restored and the wife vindicated. In the mabinogi the motif of the Hand is utilized to remove the child.* A good old example 1 Cf. Nutt, Scottish Celtic Review, p. 140. 2 The Calumniated Wife has been studied by many scholars. See, for exam- ple, Dunlop-Liebrecht, Prosadichtungen, pp. 265-266 ; Hahn, Griechische u. Albanesische Mdrchen, II, 292 ff . ; D'Ancona, La Rappresentazione di Santa Uliva, Pisa, 1863 ; the same, Sacre Rappresentazioni, III, 235 ff. ; Wesselofsky, Novella delta Figlia del Re di Dacia, Pisa, 1866 ; Todd, Publications of the Modern Lan- guage Assoc, of America, IV, no. 3, pp. ii ff. ; Temple, note in Mrs. Steel, Tales of the Punjab, pp. 364-365; Puymaigre, Folk-Lore, pp. 253 ff., 325-326; Crane, Italian Popular Tales, pp. 1 7 ff . ; Cosquin, Contes pop. de Lorraine, I, Ixiii, 1 90 ; Suchier, CEuvres poetiques de Philippe de Remi, I, xxiii ff. ; Nutt, Celtic Magazine, XII^ 549-550; Milusine, III, 212, 253 ff., 527-528; Clouston, Variants and Analogues of the Tales, in vol. Ill of Sir R. F. Burton's Supplemental [Arabian^ Nights, pp. 617 ff. ; id., Book of Sindibdd, pp. 372 ff.; Kohler-Bolte, Zir