^{page 326 note 1} As in the eighth century Irish, Dream of Oenghus, translated by Jackson, K., A Celtic miscellany, 1951, No. 39Google Scholar. Here the chief swan-maiden is wooed at the lakeside, though not by seizure of her feather-dress, for she has none. Like her lover it seems she was a swan by pure transformation.

^{page 326 note 2} See the Far Eastern, north-east Siberian, and extreme North American variants quoted in sections 2 and 3 of this article.

³ Such as that of ‘Hassan of Basra’ and ‘Janshāh’ in the Thousand-and-one nights (Burton, VIII, 7 ff., and v, 329 ff.)Google Scholar and similar tales in Persian, (Bricteau, , Contes persanes, 1910, 277)Google Scholar, Kurdish, (Prym, and Socin, , Kurdische Sammlungen, II, 1890, No. 11)Google Scholar, Turkish, (Kunos, , Türkische Vollcsmärchen, 1905, 11)Google Scholar, Avar, (Awarische Texte, ed. Schiefner, , 1873, Mém. de l'Acad. de St. Petersbourg, Sér. VII, Tom. XIX, No. 6)Google Scholar, Russian, (Coxwell, , Siberian and other folk-tales, 1925, p. 687)Google Scholar, Polish (Coxwell, , op. cit., p. 953Google Scholar), German, , since the fourteenth century ‘Friedrich von Schwaben’ (Jellinek, , Deutsche Texte des Mittelalters, I, 1904)Google Scholar, and variously in Western European tales and their descendants overseas.

^{page 326 note 4} See Thompson, Stith, The folk-tale, 1951, 21 ffGoogle Scholar.

^{page 326 note 5} cf. Waley, A., Ballads and stories from Tun-huang, 1960, 259Google Scholar.

^{page 326 note 6} See pp. 330 ff. and 336, below.

^{page 326 note 7} See section 3, below.

^{page 326 note 8} See section 2, below.

^{page 327 note 1} See section 6, below.

^{page 327 note 2} See p. 334, n. 3.

^{page 327 note 3} Kanbū, 'Ināyat Allāh, Bahār i dānish, transl. by Scott, J., II, 1790, 213 ff.Google Scholar, ‘The merchant's son and the Peries’. The maidens appear as four doves, and in other respects the beginnings of sophistication are apparent in this fundamentally simple version.

^{page 327 note 4} See below, p. 347.

^{page 327 note 5} See section 2, below.

^{page 327 note 6} See section 2, below.

_{page 327 note 7} See section 3, below.

^{page 327 note 8} For an example in which it is hard to say whether a tale is ‘pre-European’ or not, see a Tanehetara Indian tale from Maranhao, north-east Brazil, where Iberian and Afro-Brazilian influences have long been at work. The bird-maiden is a vulture. Cited in the Standard dictionary of folk-lore, 1949, 1092. See further, p. 350, n. 2.

^{page 327 note 9} The bird-maidens bathe in a lake in a walled garden. Radloff, , Proben der Volksliteratur der türkischen Stämme Süd-Siberiens, 1886, VI, 122Google Scholar.

^{page 327 note 10} See p. 331, n. 6.

^{page 327 note 11} Quoted by Tylor, , Researches into the early history of mankind, 2nd ed., 1870, 354Google Scholar. The bird-maidens are white doves.

^{page 327 note 12} Codrington, , The Melanesians, 1891, 172Google Scholar. The bird-maidens are ‘Web-wings’ or ‘Dove-skins’.

^{page 327 note 13} See section 4, below.

^{page 327 note 14} See section 5, below.

^{page 328 note 1} Holmström, K., Studier ōver svanjungfrumotivet i Volundarlcvida och annorstädes, 1919, 109Google Scholar. Following the methods of the ‘Finnish’ school, Holmstrom methodically lists the variants of the several types and sub-types of Swan Maiden story according to their geographical distribution. His book is therefore an essential aid to any study of the Swan Maiden story. N. Penzer was convinced of the Indian origin of the story, see p. 349, below, as were supporters of the ‘Aryan’ theory of the origin of fairy tales inspired by Th. Benfey, see p. 349 below.

^{page 328 note 2} The orthodox scholars of the ‘Finnish’ school cite as a Swan Maiden story a Mongol tale about a man who helped a dragon king and who was rewarded by the gift of a bitch that changed into a beautiful woman by night (Jülg, , Mongolische Märchensammlung, 1868, 192)Google Scholar.

^{page 328 note 3} Coxwell, , op. cit., No. 9Google Scholar.

^{page 328 note 4} ‘T'ien K'un-lun,’ pp. 149–55.

^{page 328 note 5} DrWaley, comments: ‘“K'un-lun” means “the Negro”; probably a nickname’, op. cit., 259Google Scholar.

^{page 329 note 1} The narrator is not always apt: (i) the youngest had obviously not turned into a crane ike the others; (ii) he forgets to state that K'un-lun had taken the robe of the youngest.

^{page 329 note 2} ‘T'ien K'un-lun.’

^{page 329 note 3} Zong-in-Sobh, , Folk-tales from Korea, 1952, No. 11Google Scholar.

^{page 329 note 4} T'oung Poo, VI, 1895, 65Google Scholar = Brauns, D., Japanische Sagen und Lokalsagen, 1885, 349Google Scholar; T'oung Poo, VI, 66 ff.; Waley, , The nō plays of Japan, 1921, 218 ff.Google Scholar, ‘Hagoromo’, by Seami, (1363–1444)Google Scholar, and also earlier; the same play as the one recounted in Mitford, , Tales of old Japan, 1893, 111Google Scholar. All of these versions have to do with localization at Miho-no-Matsubara (the pine-strand of Miho) in Suruga Bay, from where the heavenly maiden can conveniently ascend to Mt. Fuji. In all but one variant, the marriage between the poor mortal (who is always a Suruga fisherman) and the heavenly maiden, is sublimated into a heavenly dance which she performs for him in return for her feathery or white fleecy garment. In the remaining variant a marriage pact is made, and it is only later that the heavenly person regains possession of her robe and flies away, to be followed by her husband, who becomes a heavenly person with her. In the Korean version, too, the husband becomes a heaven-dweller for a time, and this Korean story is localized at a temple situated on a holy mountain.

^{page 329 note 6} Williams, S. W., ‘Journal of a Mission to Lewchew in 1801’, Journal of the North China Branch of the R Asiatic Society, New Series, VI, 1871, 163 fGoogle Scholar.

^{page 329 note 6} In support of the Tun-huang cranes there is the name of the heavenly maiden's daughter in the Ryū Kyū variant, namely ‘True Crane’ which in a verbal communication Dr. Waley corrects from ‘True Heron’ of Williams's rendering.

^{page 330 note 1} I am indebted to Dr. Waley for this item of information, and also for answering ten other questions of fact at an early stage of this inquiry.

^{page 330 note 2} Waley.

^{page 330 note 3} de Groot, J. J. M., The religious systems of China, IV, 1901, 289Google Scholar.

^{page 330 note 4} ibid., IV, 289.

^{page 330 note 5} Groot, , op. cit., I, 1892, 172, and 226, n. 2Google Scholar. At I, 57, Groot seeks to account for the symbolism of the crane as an emblem of life and prosperity by quoting the phonetic value of its character which resembled that of the character for ‘happiness’; but one does not need to be a sinologist to know that the characteristics of a living animal with which man has habitual contact will always overpower any puns that can be made with its name, unless the second term of the pun really suits. See below, p. 332, for an early Chinese Crane Dance, which probably had to do with spring, mating, and fertility.

^{page 330 note 6} Florenz, K., Japanische Mythohgie, 1901, 305 fGoogle Scholar.

^{page 330 note 7} See p. 345, below.

^{page 331 note 1} I am grateful to MrDaniels, F. J. for providing me with this reference, for discussing with me the difficulties in dating the story in the Afumi fūdolki, and for helping me to present the information on p. 329, n. 4, accuratelyGoogle Scholar.

^{page 331 note 2} Translated by Arthur Waley, Part II, ch. viii, ‘The picture competition’.

_{page 331 note 3} Waley, , Ballads and stories from Tun-huang, 155Google Scholar.

^{page 331 note 4} ibid., 155.

^{page 331 note 5} It is no valid objection, as bird-maiden stories are told, that in this case the girls would have had their bird-dresses on. Throughout the Northern Hemisphere in such stories there is confusion between the states of bird and maiden—see the Tun-huang version itself, in which it is said that all three sisters turn into cranes, but the elder sisters fly up to heaven clutching their heavenly robes, while the third in human form has to borrow a shirt for her nakedness.

^{page 331 note 6} Coxwell, F., Siberian and other folk-tales, 1925, p. 266Google Scholar, ‘The little old woman with five cows’, a Swan Maiden story embedded in much extraneous material.

^{page 332 note 1} Danses et légendes de la Chine ancienne, I, 1926, 211 ffGoogle Scholar.

^{page 332 note 2} See the ingenious and convincing study by Johansen, K. F., ‘Thesée et la Danse à Délos’, Det kgl. Danske Videnskabernes Selskab. Arkssologisk-Kunsthistoriske Meddelelser, Bd. III, Nr. 3, 1945, 12, 58 ffGoogle Scholar. In this study, the relation is shown between the François Vase in the Museum of Florence and references to the Geranos in Plutarch, Theseus, 9.d, and Lucian, , Salt. 34Google Scholar.

^{page 332 note 3} Modern and popular Chinese versions can be followed up in Eberhard, W., Typen chinesischer Volksmärchen (FF Communications, No. 120), 1937, 55 ffGoogle Scholar. Eberhard cites 25 examples, including the Hsüan-chung chi and Tun-huang versions. He summarizes the story thus: (1) a poor young man finds fairies bathing in a river; (2) he wins one of them for his wife by taking her clothes; (3) after many years she finds them again and flies up to heaven; (4) the man pursues her; (5) God commands eternal separation—they may meet but once a year (an astral conception, cf. the story of the Oxherd and the Weaver Maid: but cf. also Urvaśī's one night in the year together with Purūravas in the Śatapatha-brāhmana). Only in the Hsüan-chung chi and Tun-huang versions does it seem that the ‘fairies’ assume bird-form. In several variants an ox tells the man where the fairies bathe, in one variant a deer tells him (as in several other Eurasian variants and also in a well-authenticated group of variants from Utah, U.S.A.).

^{page 334 note 1} As, for example, in an Aleut-Eskimo tale from Kodiak Island, Golder, , Journal of American Folk-Lore, XVI, 1903, 98 ffGoogle Scholar.

^{page 334 note 2} See the paragraph on the migration of cranes in China, below, p. 339.

^{page 334 note 3} Bogoras, W., ‘Chuckchee tales’, Journal of American Folk-Lore, XLI, 1928, 429 ffGoogle Scholar. ‘Story of a bird-woman’, told in the summer of 1895 = Coxwell, F., op. cit., pp. 82 ff.Google Scholar, translating an earlier Russian version by Bogoras. The motif of the gift of feathers recurs in an East Cheremiss version in which the swan-wife succeeds in flying away after the third night of enforced marriage. See Genetz, A., Journal de la Société Finno-Ougrienne, VII, 1889, pp. 125 f., No. 17Google Scholar.

^{page 334 note 4} Bogoras, , op. cit., 431 fGoogle Scholar.

^{page 334 note 5} Swanton, , ‘Tlingit myths and texts’, Bulletin of the Bureau of American Ethnology, 39, 1909, No. 24, p. 55Google Scholar, ‘The brant wives’.

^{page 335 note 1} Coxwell, , op. cit., p. 266Google Scholar.

^{page 335 note 2} Journal of American Folh-Lore, LXVI, 1953, 72Google Scholar.

^{page 335 note 3} The variant from Kodiak Island, see p. 334, n. 1.

^{page 335 note 4} Boas, F., ‘The Central Eskimo’, Sixth Annual Report of the American Bureau of Ethnology, 1888, 615Google Scholar.

^{page 335 note 5} See below, p. 345.

^{page 335 note 6} As in the variant from Kodiak Island, see p. 334, n. 1. The wife's sisters-in-law call her ‘goose-mouth’.

^{page 336 note 1} I read first the myth and then Harva's illuminating remarks on its ethnological background (see p. 341, n. 6), after I had met it in a mutilated form in Ohlmarks, , Studien zum Problem des Schamanismus, 1939, 212 f.Google Scholar, and then hardly more significantly in Eliade, , Schamanismus und archaische Ekstasetechnik, 1957, 158Google Scholar, quoting Ohlmarks. As an editor of ‘FF Communications’, Harva had naturally referred the reader of this account of the Buryat myth to Eber-hard, W., Typen chinesischer Volksmärchen (FFC, No. 120), 55Google Scholar, that is to various Chinese versions of the Swan Maiden story. At the corresponding point of his account, Ohlmarks, referred to ‘the story of Wayland, etc’ (op. cit., 213)Google Scholar. The present article may be regarded as Ohlmark's ‘etc.’. And at the corresponding part of his work, Eliade refers the reader to Thompson's, Stith famous Motif-index of folk-literature, III, 10, 381Google Scholar, adding: ‘Noch häufiger ist das Motiv von der Vogelfee, die, mit einem Menschen verheiratet, wieder davonfliegt, sobald es ihr gelingt, sich der lange von ihrem Mann gehüteten Federn zu bemächtigen’. I was surprised to find that such excellent hints as Harva's (1938), Ohlmarks's (1939), and Eliade's (French edition, 1951, 150) had, so far as I knew, been ignored.

^{page 336 note 2} Die religiösen Vorstellungen der altaischen Völker (FF Communications, No. 125), 1938Google Scholar.

^{page 335 note 3} Harva, , op. cit., 467 f.Google Scholar, quoting Skazanija burjat, zapisannyja raznymi sobirateljami (ZVSORGO, 1, 2), Irkutsk, 1890Google Scholar. Quoting a piece of the same or of an allied myth, Ohlmarks, , op. cit., 212 f.Google Scholar, cites Poppe, N. N., Opisanie mongolskich ‘šamanskich’ rukopisej (Zap. Inst. Vost.-ved. Ak. Nauk, 1), 1932, 195Google Scholar.

^{page 337 note 1} Harva, , op. cit., 469Google Scholar.

^{page 337 note 2} Harva, , op. cit., 471Google Scholar, quoting Šaškov, S., Šamanstvo v Sibiri (ZRGO 1864, 2), St. PetersburgGoogle Scholar; Potanin, G. N., Očerki severo-zapadnoj Mongolii, IV, 1883, 24Google Scholar. The Ainu, too, sometimes trace descent from a swan-mother (see Batchelor, J., Uwepekere or Ainu fireside stories, 1924, 101 ff.Google Scholar, ‘The divine Swan’; late nineteenth century). The Ainu also knew a story about marriage with a heavenly maiden by stealing her clothes (apparently not feather-robes) while she is bathing (see Etter, C., Ainu folklore, 1949, 155 ff.)Google Scholar. This may well be an offshoot of our ‘celestial’ Far Eastern version (B), via either Japan or Korea.

^{page 337 note 3} Harva, , op. cit., 471Google Scholar, quoting Šaškov and Potanin. Marco Polo records that the Great Khan enjoyed staying at his city of Chagan-nor (‘White Pool’) because of the abundance of swans on its waters. Later he says that when the Khan was at ‘Cachar Modun’, east of Peking, he flew his hawks at the numerous cranes and swans there. In accordance with their different myth of origin, the house of Chingiz were free to kill birds sacred to others.

^{page 337 note 4} Harva, , op. cit., 470Google Scholar, on the authority of Potanin, see [read] n. 2, above.

^{page 337 note 5} Harva, ibid., 470.

^{page 337 note 6} ibid.

^{page 337 note 7} See p. 339, below.

^{page 337 note 8} Harva, , op. cit., 473 fGoogle Scholar.

^{page 338 note 1} I am indebted to my colleague Dr. John Carthy, of the Zoological Department, Queen Mary College, London, for his kindness in spending a morning with me at the Library of the Zoological Society of London, and finding and explaining the books on which the following section is based, and also to Mr. L. G. Ellis of the Library's staff for the kind help he gave us. Dr. Carthy was also good enough to read through what I have written, but I accept responsibility for all its defects, zoological and otherwise.

^{page 338 note 2} Delacour, J., The waterfowl of the world, I, 1954, 51Google Scholar. Thanks to the superb organization of this book, and not least to its maps, it was possible for me to place before the reader precise generalizations in answer to the questions raised by the Swan Maiden story, so far as it concerns swans and geese.

^{page 338 note 3} Delacour, , op. cit., 62 ffGoogle Scholar.

^{page 339 note 1} ibid., I, 71 ff.

^{page 339 note 2} ibid., I, 75 ff.

^{page 339 note 3} ibid., I, 82 ff.

^{page 339 note 4} ibid., I, 141 ff.

^{page 339 note 5} ibid., I, 101 ff.

^{page 339 note 6} ibid., I, 111 ff.

^{page 339 note 7} ibid., I, map 20, p. 156.

^{page 340 note 1} See p. 329, above.

^{page 340 note 2} By Mr. F. J. Daniels.

^{page 340 note 3} My authority for the crane is Makatsch, W., Der Kranich (Die Neue Brehm-Bücherei, Ht. 229), Wittenberg, 1959Google Scholar. It, too, is well organized and has excellent maps.

^{page 340 note 4} Delacour, , op. cit., I, map 11, pp. 84 fGoogle Scholar. See also p. 337, n. 2 (Ainu), above.

^{page 340 note 5} Peters, J. L., Check-list of the birds of the world, IV, 1940, p. 204Google Scholar.

^{page 340 note 6} From a verbal communication by Dr. Carthy.

^{page 341 note 1} Studien zum Problem des Schamanismus, 1939, 1 ff., and see maps 1–3, at the back of the book.

^{page 341 note 2} Ohlmarks, , op. cit., 211Google Scholar.

^{page 341 note 3} Harva, , op. cit., 510Google Scholar, and passim.

^{page 341 note 4} Ohlmarks, , op. cit., 210Google Scholar.

^{page 341 note 5} ibid., 210.

^{page 341 note 6} Harva, , op. cit., 503 ffGoogle Scholar.

^{page 341 note 7} ibid., 503.

^{page 341 note 8} ibid., 508.

^{page 341 note 9} ibid., 509.

^{page 341 note 10} ibid., 524.

^{page 342 note 1} ibid., 525 f.

^{page 342 note 2} ibid., 510.

^{page 342 note 3} ibid., 522 f.

^{page 342 note 4} ibid., 524.

^{page 342 note 5} ibid., 546 f.

^{page 342 note 6} ibid., 519, fig. 84.

^{page 342 note 7} ibid., 549, fig. 104.

^{page 342 note 8} ibid., 555.

^{page 342 note 9} ibid., 559.

^{page 342 note 10} ibid., 475.

^{page 342 note 11} ibid., 465 ff.

^{page 342 note 12} See p. 331, above.

^{page 343 note 1} Eskimo variants: Boas, , Journal of American Folk-Lore, XII, 1899, p. 171, No. 7 (Smith Sound)Google Scholar; Rink, , Tales and traditions of the Eskimo, 1875, 145 (Davis Strait)Google Scholar; Rassmussen, , Nye mennesher, 1905, 181Google Scholar, retold by N. Penzer, The ocean of story, VIII, 1927, 228 f.; and see p. 334, n. 1, and p. 335, n. 4, for other Eskimo versions. For the Chukchi variants of the story see p. 334, n. 3–4, above.

^{page 343 note 2} Among the Buryat, the shaman climbs up a sacral birch tree in his yurt and out through the smoke-vent in order to symbolize his ascent to Heaven. Ohlmarks, , op. cit., 29 fGoogle Scholar.

^{page 343 note 3} Eliade, , Schamanismus und archaische Ekstasetechnik, 385Google Scholar.

^{page 343 note 4} See, for example, Egil's struggle with the ‘Lappish’ wife of King Eirik when he was the latter's prisoner at York, Egil's saga, ch. 59.

^{page 344 note 1} The reference was supplied by DrWaley, . See his The nine songs: a study of shamanism in ancient China, 1955, 13 ff., 47, 52Google Scholar. But see my Addendum, below.

^{page 344 note 2} I am indebted to ProfessorsBasham, A. L. and Brough, J., and to MrWright, J. C., for advising me on what constructions it was permissible to place on the text of Rgveda X.95Google Scholar.

^{page 345 note 1} Geldner, K. F., Der Rig-veda aus dem Sanskrit ins deutsche ubersetzt und mit einem lawfenden Kommentar versehen, III (Harvard Oriental Series, Vol. 35), 1951, pp. 298 ffGoogle Scholar. The alleged names of the Apsarases: p. 301.

^{page 346 note 1} As for example in the Irish, OldDream of Oenghus, which is not a ‘perfect’ Swan Maiden story (see p. 326, n. 1)Google Scholar. Oenghus was told to go in human shape to the border of a loch and to call the maiden Caer, who was a swan and a maiden in alternate years. Oenghus did as he was bidden, and Caer came to him as a swan. ‘He cast his arms about her. They fell asleep in the form of two swans, and went round the lake three times.’

^{page 346 note 2} Ed. Jellinek, M. H. in Deutsche Texte des Mittelalters, I, 1904Google Scholar. See especially p. xviii for the occurrence of the name ‘Wieland’ in the MSS.

^{page 346 note 3} These are the standard doves of the Near Eastern and European fairy stories and may be regarded as an urbanization of the motif.

^{page 347 note 1} Stack, ed. Lyall, , The Mikir, 1908, 55 ffGoogle Scholar. The material was collected c. 1884.

^{page 347 note 2} Even in Somadeva's Kathā-sarit-sāgara, see Penzer, , The ocean of story, VIII, 1927, 58Google Scholar (Tawney's translation), the bird element is absent; for the ascetic steals the clothes of a heavenly maiden while she bathes and she, as a result of a curse, becomes his wife. The only connexion with a bird is that the maiden can fly up to Heaven. But so can the ascetic when he has eaten their child. A story of the eleventh century.

^{page 348 note 1} álptarhamir, i.e. ‘swan, body-casings’: cf. fia rhamr elsewhere in the Edda (of gods); and Old Saxon fe erhamo (ninth century) and Old English fe erhama (of angels).

^{page 348 note 2} ProfessorBasham, points out in The wonder that was India, 1954, 317Google Scholar, that ‘sometimes the apsarases appear in the role of valkyries, raising slain heroes from the battlefield and bearing them to heaven to be their lovers’. On pp. 405 ff. he gives a condensed account with some verse translations of the Śatapatha-brāhmaṇa version of ‘Urvaśī and Purūravas’.

^{page 348 note 3} cf. Schneider, H., Germanische Heldensage, II, 2, 1934, 77 ffGoogle Scholar.

^{page 349 note 1} See Benfey, , Pantschatantra, I, 1859, pp. 263 ffGoogle Scholar.

^{page 349 note 2} Without feather-garments; The ocean of story, VIII, 213 ff.

^{page 349 note 3} See p. 347, n. 1.

^{page 350 note 1} Apart from copious references to Swan Maiden stories of all kinds I was able to derive little help for my present purpose from Holmström's, Studier, see p. 328, n. 1Google Scholar

^{page 350 note 2} See Teit, J., Publications of the Jesup North Pacific Expedition, II, 1907, p. 727Google Scholar. The green and red garters give it away that we have to do with the French and then French-Canadian story of ‘La Belle-jarretière-verte’, see Barbeau, , ‘Contes populaires canadiens’, Journal of American Folk-Lore, XXX, 1917, 361 ffGoogle Scholar. Cf. also Farrand, D., ‘Traditions of the Chilcotin Indians’, Publications of the Jesup North Pacific Expedition, II, 1900, p. 26, No. xiGoogle Scholar, where Farrand agrees that some elements are probably French. The part of the pursuit in which the lovers turn into ducks must certainly be so.

^{page 350 note 3} In her sentiments on this score the queen of Candaules, and later queen of Gyges, was old-fashioned. As Herodotus remarks when telling her story: ‘With the Lydians as with most barbarian races, it is thought highly indecent even for a man to be seen naked’ (The histories, bk. I, transl. A. de Selincourt).

^{page 351 note 1} References to Swan Maiden stories were obtained from Holmström's Studier ōver svanjungfrumotivet already referred to, and from Stith Thompson's great works: Motif-index of folk-literature, motifs B 653.1 (marriage with a bird-maiden), F 302.4.2 (a fairy comes into a man's power when he steals her wings; she leaves him when she finds them), K 1335 (seduction or wooing by stealing the clothes of a bathing girl); and Tales of the North American Indians, 1921, 198, 356.

I am grateful to Professor Kenneth Jackson for suggesting how I might bring this inquiry more into line with current practice in folk-tale studies than was originally the case.