^{page 10 note 1} Polivanov, E. A. seems to have been the first to point out this correspondence, in Zapiski vostochnago otdeleniya imperatorskago russkago arkheologicheskago obshchestva, 23 (1916), 263–4Google Scholar. See also Conrady, A., “Alte westöstliche Kulturwörter”, Ber. ü. d. Verh. d. Sächs. Ak. d. Wiss. zu Leipzig, Phil.-hist. Kl., 77 (1925)Google Scholar.

^{page 11 note 1} Conrady, op. cit.

^{page 11 note 2} Shafer, R. (Orbis, 12 (1963))Google Scholar has recently published a list of some hundred or so “look-alikes” between Tibeto-Burman and Indo-European on the basis of which he proposes to reconstruct a linguistic superfamily which he calls “Eurasial”. He has not, however, established regular rules of phonetic correspondence nor shown any congruences between the phonological and morphological systems of the two language families.

^{page 11 note 3} “An interpretation of the vowel systems of Old Chinese and Written Burmese”, Asia Major, 10 (1963), 200–222Google Scholar.

^{page 11 note 4} “The Indo-European vowel system and the qualitative ablaut”, Word, 21 (1965), 81–101Google Scholar, and “Close/open ablaut in Sino-Tibetan”, Lingua 14 (1965), 230–240CrossRefGoogle Scholar.

^{page 12 note 1} Childe, V. G., “The first waggons and carts from the Tigris to the Severn,” Proceedings of the Prehistoric Society, 17 (1951), 188–194CrossRefGoogle Scholar.

^{page 12 note 2} On the horse and chariot among the Indo-Europeans and the Aryans in India see Stuart Piggott, Prehistoric India, 1950, Chapter 7.

^{page 13 note 1} The chariot remains recovered from Shang and Chou burials and the literary references to horses and chariots in the Book of Odes have recently been studied by von Dewall, M., Pferd und Wagen im frühen China, Bonn, 1964Google Scholar. Miss von Dewall remarks on the close parallels between chariots in Egypt, Mycenae and Assyria and those in China and accepts as certain that the chariot spread from a single point of origin.

^{page 13 note 2} For example, quite recently, Walter Fairservis, A. in The Origins of Oriental Civilization, New York, 1959Google Scholar.

^{page 13 note 3} This is the view still followed by Gimpera, P. Bosch, Les Indo-européens, problèmes archéologiques, Paris, 1961Google Scholar.

^{page 14 note 1} See Lane, G. S., “The present state of Tocharian research,” Proceedings of the VIII International Congress of Linguists, Oslo, 1958, pp. 252–261Google Scholar.

^{page 14 note 2} See, for example, Kretschmer, P., Anz. d. phil-hist. Kl. d. Ak. d. Wiss. in Wien, 1943, 36 ff.Google Scholar and (in refutation) Mayrhofer, M., Die Sprache, V, 94Google Scholar; VI, 107 ff.

^{page 15 note 1} Gimbutas, M., “The Indo-Europeans: Archaeological Problems,” American Anthropologist, 65 (1963), 815–836CrossRefGoogle Scholar.

^{page 15 note 2} Jettmar, K., “Archäologische Spuren von Indogermanen in Zentralasien,” Paideuma, 5 (1952), 236–254Google Scholar.

^{page 15 note 3} See now Brough, J., BSOAS, 28 (1965), 582–612CrossRefGoogle Scholar.

^{page 16 note 1} Burrow, T., JRAS, 1935, 667 ff.Google Scholar

^{page 16 note 2} Bailey, H. W., “Ttaugara,” BSOS, 8 (1937), 883–921CrossRefGoogle Scholar. The argument that the names Tochari and Throana (Tun-huang) contained phonemes not found in Tocharian is inconclusive on two grounds: (1) the names are only known in foreign transcriptions which do not accurately represent the phonetic values of the native phonemes (does χ represent an aspirated stop or a fricative?—Chinese has a glottal stop in the corresponding place in Ta-yüan); (2) the Tocharian dialect on which the transcriptions were based may have differed from the attested Tocharian languages in these respects—there is a gap of nearly a millennium in date, as well as difference of locality.

^{page 16 note 3} Maricq, A., “La grande inscription de Kaniska et l'étéotokharien,” Journal asiatique, 246 (1958), 345–441, especially pp. 396–7Google Scholar.

^{page 16 note 4} A. Maricq, op. cit., Henning, W. B., “The Bactrian Inscription,” BSOAS, 23 (1960), pp. 47–55CrossRefGoogle Scholar and Maricq, A., “Bactrien ou étéotokharien,” Journal asiatique, 248 (1960), 161–6Google Scholar.

^{page 17 note 1} Haloun, G., “Zur Üe-tṣï-Frage,” ZDMG, 91 (1937)Google Scholar.

^{page 17 note 2} Pelliot, P., Journal asiatique, 224 (1934), p. 26Google Scholar; Pulleyblank, , Asia Major, 9 (1962), p. 93Google Scholar.

^{page 17 note 3} Shih-chi, 110.0245.1; Han shu, 94A.0596.1.

^{page 17 note 4} Asia Major, 9 (1962), 93Google Scholar.

^{page 17 note 5} Marquart, J., Ērānšahr, 1901, 206Google Scholar.

^{page 18 note 1} Pliny, VI, 13. See Pauly-Wissowa, Reallexikon, article Utidorsi.

^{page 18 note 2} The identity of Aσιοι., Asiani and 'Iτιοι was assumed by Marquart, , also by Herzfeld, E., Archäologische Mitteilungen aus Iran, 4, 26Google Scholar and Konow, S., Corpus Inscriptionum Indicarum, 2/1, LIXGoogle Scholar; but has been denied by Altheim, F., Geschichte der Hunnen, I, 64Google Scholar.

^{page 18 note 3} The question is fully discussed by Pelliot, P., Notes on Marco Polo, I, 418 ff.Google Scholar

^{page 18 note 4} Chin shu, 125.1396.3, noted by Eberhard, W., Das Toba-Reich, 1949, p. 325Google Scholar, who writes incorrectly Wei-ch'ih. To Eberhard belongs the credit of first suggesting the equivalence of the names Yü-ch'ih and Yüeh-chih, though he does so with inadequate phonetic arguments and on historical grounds that are also somewhat dubious. He does not seem to be aware that the name was applied to the kings of Khotan but notes an occurrence of it applied in the T'ang period to a man from T'u-hu-lo = Tocharestan. This would be highly interesting. Unfortunately this man, a well-known painter, is said elsewhere to be from Khotan and probably belonged to the royal family which was given the name Yü-ch'ih, as we have seen. (See Pelliot, op. cit., p. 419). It is tempting, but no doubt rash, to try to see some actual connection between the name Viśa and Yü-ch'ih = Yüeh-chih (dating perhaps to a period of Kushanian rule in Khotan?). But any such conjecture could only be verified by arguments coming from the Iranian side.

^{page 18 note 5} Haloun, G., Asia Major, n.s. I (1949), 119–138Google Scholar.

^{page 18 note 6} Wei shu, 2.1907.3.

^{page 18 note 7} Wei shu, 113.2194.2.

^{page 19 note 1} Hsin T'ang shu, 40.3726.3. See Pelliot, , Histoire ancienne du Tibet, 1961, 141Google Scholar.

^{page 19 note 2} Chin shu, 125. cf. Wei-yüan, Yao, Pei Mao hu-hsing k'ao (1958), 249Google Scholar.

^{page 19 note 3} See Haloun, , ZDMG, 91 (1937), 301 ff.Google Scholar

^{page 19 note 4} Shih-chi, 20.0087.2; Han shu, 17.0347.1. Han shuhas yu instead of jo.

^{page 19 note 5} Han shu, 94.B.0602.3.

^{page 20 note 1} See the articles quoted in note 4, p. 16.

^{page 20 note 2} For the various forms of the name *Argi see Henning, W. B., “Argi and the ‘Tokharians’”, BSOS, 9 (1937–1939), 545–571CrossRefGoogle Scholar.

^{page 20 note 3} Han shu, 94A.0597.2. Shih-chi, 110.0246.2 has Yen-chih. On the pronunciation of the second character, M. cie < *ky-, see Asia Major, 9 (1962), 98–100, 105–7Google Scholar.

^{page 20 note 4} The relevant texts are gathered together and discussed by Laufer, , Sino-Iranica (1919), 324–8Google Scholar. Yen-chih became a general name for cosmetics derived from a number of similar plants.

^{page 20 note 5} Shih-chi, 110.0246.2 and 123.0267.2 (with commentary); Han shu, 94A.0597.2, 96B.0607.2.

^{page 20 note 6} Hou Han shu, 2.0655.1 (commentary); Chiu T'ang shu, 40.3230.2, under Hsi Chou, Liu-chung (Lukchun) Hsien (with in error for ), and again under T'ien-shan Hsien (correctly).

^{page 21 note 1} The name Tun-huang is conventionally read with the ordinary reading of the first character tun, implying M. τwǝn, but according to the commentary of Yen Shih-ku to Han shu, 28B.0425.4 it ought to be read like t'un < M. dwǝn and this is confirmed by the Kuang yün, which lists the character as a homophone of with this reading and quotes the name Tun-huang as an example of its use. The voiced initial is in fact in agreement with the Sogdian form (the Chinese must, however, imply a stop, not a fricative). See Bailey, , BSOS, 8, p. 293Google Scholar. On the other hand a voiced initial in a foreign transcription may be deceptive when the original language does not distinguish voiced and unvoiced stops. Note the frequent use of d- in Chinese transcriptions of Turkish words beginning with t-like M. dât-kân= tarqan and M. dwǝt-khιwât= Türk. See also note 2, p. 16.

^{page 21 note 2} Bailey, H. W., Asia Major, n.s. 2 (1951), 15Google Scholar.

^{page 21 note 3} On this name see Pelliot, , T'oung-Pao, 22 (1923), 126–132CrossRefGoogle Scholar.

^{page 21 note 4} Shui ching chu, 1, p. 22 (Kuo-hsüeh chi-pen ts'ung-shu). The same name occurs in the legendary geography of the Shan hai ching, 3.5b (Ssu-pu peiyao), but we need not assume that the commentary to the Water Classic has been contaminated by this; rather we may suppose that the real name from the western regions has got into the legendary work.

^{page 21 note 6} Han shu, 96B.0609.1 has only Yen , but the passage is quoted in the commentary to Hou Han shu, 77.0802.2 and there has Chü-yen.

^{page 22 note 1} See Asia Major, 9 (1962), pp. 90, 224Google Scholar. It is hardly necessary to refute the old suggestion that Yüan could represent Yavana — Greeks. The initial glottal stop in Chinese corresponding to Greek χ, Sanskrit kh, may represent a consonant in the original language with uvular articulation, cf. Yü-mi M. -iou-mye or -ou-mye, later M. ku-mye, which is no doubt equivalent to Kharoṣṭhi Khema, and other examples of the same kind given in the first reference cited above.

^{page 22 note 2} Hou Han shu, 118.0905.1.

^{page 22 note 3} Han shu, 96A.0607.3.

^{page 22 note 4} Previous attempts at identifying Kuei-shan city are more or less plausible guesses based on the assumption that Ta-yüan = Ferghana. There is a certain superficial attraction about the proposal of Terrien de Lacouperie to identify it with Kāsān, an ancient capital of Ferghana, but the phonetic similarity is very inexact. It may be of interest to note that in T'ang times Kāsān was referred to as K'o-se M. khât-sǝk, probably implying simply *Kas (Chavannes, , Documents sur les Tou-kiue Occidentaux, 148, 273Google Scholar). (On the use of ju-sheng syllables to represent single consonants see my article “The Chinese name for the Turks”, to appear shortly in the Journal of the American Oriental Society.) The earlier discussions may be found in the articles of Shiratori, K., Tōyō Gakuhō, 6 (1916)Google Scholar, Kuwabara, J., Tōsei kōtsūshi ronsū, 1933Google Scholar, T. Fujita, Tōsei kōshōshi no kenkyū, Saiiki hen, etc.

^{page 22 note 5} The last contacts between China and Ta-yüan under that name were in the time of Western Chin (3rd century) (Chin shu, 97.1337.1). In the 5th century ambassadors came from the country of P'o-lo-na and presented “blood-sweating horses”: Wei shu, 5, 1915Google Scholar. The Pei-shih, 97.3042.4, probably copying here the original chapter on the Western Regions of the Wei shu, identifies [P'o-]lo-na with the ancient Ta-yüan. Unfortunately the identifications of contemporary place names with those of Han times which were made by the author of this work are extremely unreliable and have caused great confusion. It may be doubted whether he had any reason apart from the “blood-sweating horses” for taking P'o-lo-na to be Ta-yüan. P'o-lo-na is certainly Ferghana. There is another article on it in the Pei-shih under the name Po-han (this time copied from the Sui shu) in which it is identified, not with Ta-yüan, but with Ch'ü-sou, a western people mentioned in the Yü-kung section of the Shu-ching (Pei-shih, 97.3044.1, Sui shu, 83.2535.4). The T'ung tien, 192, on the other hand, apparently following the Sui dynasty work Hsi-yü t'u-chi, identified Ta-yüan with Su-tui-sha-na (Sutrushana). This is followed by the Hsin T'ang shu, 221B.4154.1, where the capital of Sutrushana (near present Ura-tübe) is identified with Erh-shih, the city besieged by Li Kuang-li. In the next passage the Hsin T'ang shu says that the country of Shih (Tashkend) was in the northern part of ancient Ta-yüan. No doubt this identification was what led the T'ang government to give the name Ta-yüan Tu-tu-fu to the city of K'an-chieh in the Tashkend region in 658 (Chavannes, , Documents sur les Tou-kiue Occidentaux, 141, 273Google Scholar). Finally, the article on T'u-huo-lo (Tocharestan) in T'ung-tien, 193 (apparently copying a source of the Sui period) says that T'u-huo-lo produced blood-sweating horses and that its northern borders were the territory of Ta-yüan in the Han period. This is probably correct, though obviously not much reliance can be placed on the statement by itself in view of the other contradictory statements.

^{page 23 note 1} Shih chi, 123.0267.2.

^{page 23 note 2} Herzfeld, E., Archaeologische Mitteilungen aus Iran, 4 (1932), pp. 22 ff.Google Scholar, argues that Chang Ch'ien's Ta-yüan must have been somewhere in the Pamirs. His argument is essentially correct for the time of Chang Ch'ien (though I think he still puts Ta-yüan a little too far east) but it does not agree with the description of Ta-yüan or with the indications of the Han shu. Most other scholars, in so far as they have taken any account of the discrepancy, have tried to reconcile the two sources in some way or other. Tarn, for example, argues that Chang Ch'ien must have got his compass directions wrong. But the most important thing is that he went from Ta-yüan to Yüeh-chih by way of K'ang-chü, which simply does not make sense in terms of the Han dynasty location. In rejecting Herzfeld's suggestion Tarn relies on the description of Ta-yüan given in Chang Ch'ien's account but this ceases to be a valid argument if Ssu-ma Ch'ien interpolated later information as I suggest (and as Tarn himself already suggested in regard to the description of Parthia). See Tarn, W., The Greeks in Bactria and India, pp. 474–7Google Scholar.

The distances given in li in the Han shu are hardly to be relied upon in any precise way. No doubt they come from reports of the Chinese administration of the Western Regions, generals like Li Kuang-li, and ambassadors, which have been conflated by the historian. They have certainly suffered corruption in the course of transmission. Nevertheless over certain stretches one can find a good measure of agreement between distances from one place to the next and the longer distances from each place to Ch'ang-an or to the seat of the Protector-General. This enables one to set up fragmentary itineraries which preserve at least the route followed and the relative distances. Such a portion, going westwards from the Pamirs, is illustrated in the diagram. Over this stretch the distances from the Protectorate-General are largely consistent (assuming a few clerical errors) with the distances between successive places, thus (figures in first column are those given in text; figures in second column are calculated totals):

Over this stretch the distances to Ch'ang-an do not agree at all. On the other hand, the route along the southern side of the Tarim basin to Khotan and Kashgar is fairly consistent in terms of the distances t o Ch'ang-an but not in terms of the distances to the Protector-General. Evidently the sources for these two parts of the Western Regions were quite different.

^{page 25 note 1} op. cit., p. 281.

^{page 25 note 2} Justin, , Epitoma, 42, 2, 1–2Google Scholar.

^{page 25 note 3} Justin, , Epitoma, 42, 1, 1–5Google Scholar.

^{page 26 note 1} Justin, , Prologi historiarum Philippicarum Pompei Trogi, XLI, XLIIGoogle Scholar.

^{page 26 note 2} Ta Tang hsi-yii chi, 1; Hsin Tang shu, 221B.4154.1. See Chavannes, , Documents sur les Tou-kiue Occidentaux, p. 145Google Scholar. The identification with the petty kingdom of Fu-mo, subject to K'ang-chü in Han times, is based on the Wei shu and of no value.

^{page 26 note 3} Barthold, W., Turkestan down to the Mongol Invasion, 95–6Google Scholar.

^{page 26 note 4} Asia Major, 9 (1962), 122Google Scholar.

^{page 26 note 5} O. Maenchen-Helfen wished to find the name Kushan in a number of other names such as Kesh, Kucha, Chü-shih (a tribe living near Turfan in the Han period), etc. In none of these is there more than a vague similarity. Why not also Kashgar and Kashmir? See JAOS, 65 (1945)Google Scholar.

^{page 26 note 6} Prologus, 42. If one accepted the proposal of Haloun to emend Asiani to Cusani the coincidence would be even closer, but this seems unnecessary. See ZDMG, 91 (1937), 244, n. 4Google Scholar.

^{page 26 note 7} Asia Major, 9 (1962), 120, 218Google Scholar.

^{page 27 note 1} Barthold, W., Turkestan, 136Google Scholar.

^{page 27 note 2} See de Lacouperie, T., The Western Origin of the Early Chinese Civilization, 220, 224Google Scholar; Hirth, , JAOS, 37 (1917), 141–2Google Scholar; Chavannes, E., Les mémoires historiques de Se-ma Ts'ien, I, LXXV–LXXVIGoogle Scholar.

^{page 27 note 3} See Henning, W. B., “Argi and the Tocharians”, BSOS, 9 (1937–1939), 545–571CrossRefGoogle Scholar, and “The name of the ‘Tokharian’ language”, Asia Major, n.s. I (1949–1950), 158–162Google Scholar.

^{page 27 note 4} Vorobev-Desyatovskii, V. S., “Pamyatniki Central'no-asiatskoi Pis'mennosti”, Uchenye Zapiski Instituta Vostokovedeniya, 16 (1958), 304 f.Google Scholar, cited by Mayrhofer, M., ZDMG, 37 (1962), 326, n. 6Google Scholar. See also Ivanov, V. V., Tocharskie Yazyki (1959), 153Google Scholar, and Problemy Vostokovedeniya, 1959, 5, 188–190Google Scholar.

^{page 27 note 5} Han shu, 96A.0606.3.

^{page 27 note 6} Ta T'ang Hsi-yü chi, 12.23b (Wen-hsüeh ku-chi k'an-hsingshe reprint, Peking, 1955). cf. Watters, , On Yuan Chwang's travels in India, 2, 304Google Scholar.

^{page 28 note 1} The title yabgu was used in the country of T'u-huo-lo (Tocharestan) in the 6th century when the Chinese set up a short-lived administration there and there is no reason to assume that it had anything to do with the fact that the Turks had ruled there before the Chinese. In describing this administration Chiu Tang shu, 40.3230.4, says “Yüeh-chih tu-tu-fu—established at the capital of the country of T'u-huo-lo, O-huan city (= Warwaliz, Kunduz, cf. Asia Major, 9 (1962), p. 259Google Scholar). With their king and the yabgus () in their tribe were set up 24 prefectures (chou ). The tu-tu had overall command over them”. The next passage concerning the Hephtalites speaks of T'ai-han , rather than yabgu.

^{page 28 note 2} See Asia Major, 9 (1962), 95Google Scholar.

^{page 28 note 3} “Languages of the Saka”, in Handbuch der Orientalistik, I. Der Nahe und der Mittlere Osten, Bd. 4, Iranistik, Abschnitt 1, Linguistik, 136.

^{page 28 note 4} Han shu, 94B.0600.1.

^{page 28 note 5} Han shu, 96B.

^{page 28 note 6} Han shu, 96B.0608.2.

^{page 28 note 7} Han shu, 61.0510.1.

^{page 28 note 8} Asia Major, 9 (1962), 247 ff.Google Scholar

^{page 28 note 9} Two such references to K'ang-chü have been noted: (a) in an address to the throne by the Confucianist Tung Shung-shu which cannot be later than 134 b.c. (Han shu, 56.0496.4), (b) in a proclamation composed by the poet Ssu-ma Hsiang-ju when he was sent on a diplomatic mission to pacify the tribes in Szechwan in 130 b.c. (Shih-chi, 117.0257.2; Han shu, 57B.O5O1.3). Both references imply that a diplomatic mission from K'ang-chü had arrived some time in the reign of the Emperor Wu, who came to the throne in the year 140.

^{page 29 note 1} There has been considerable controversy in the past among Japanese scholars about the original location of the Wu-sun. Haloun concluded that they were probably in the region of Barköl. This is probably as good as one can do. Real certainty is scarcely attainable. See Haloun, , “Zur Üe-tsï-Frage”, ZDMG, 91 (1937), 246Google Scholar.

^{page 29 note 2} op. cit., 252.

^{page 29 note 3} Asia Major, 9 (1962), 136Google Scholar.

^{page 29 note 4} Asia Major, 9 (1962), 227Google Scholar.

^{page 29 note 5} Kwang-chih, Chang, The Archaeology of Ancient China (1963), 235Google Scholar.

^{page 30 note 1} Te-k'un, Cheng, Shang China (1960), 82Google Scholar.

^{page 31 note 1} Kwang-chih, Chang, The Archaeology of Ancient China, 137, 133Google Scholar.

^{page 31 note 2} See Dubs, H. H., History of the Former Han Dynasty, Vol. 2, 132–5Google Scholar.

^{page 31 note 3} Waley, A., “The Heavenly Horses of Ferghana”, History Today, 5/2 (1955), 95–103Google Scholar.

^{page 31 note 4} Meng K'ang (San-kuo Wei), cited in the commentary to Han shu, 96A.0607.3.

^{page 31 note 5} Sui shu, 83.2535.4.

^{page 32 note 1} “Crenelated Mane and Scabbard Slide”, Central Asiatic Journal, 3 (1957–1958), 85–138Google Scholar.

^{page 32 note 2} See especially Izushi, Y. “Temba kō (The horses of the sun in Chinese tradition and western horses of the Han period)”, Tōyō gakuhō, 18 (1930), 346–387Google Scholar.

^{page 32 note 3} Maspero, , La Chine antique, nouv. ed. 1955, p. 134Google Scholar, Karlgren, , Gloss 473 (BMFEA, 16 (1944), p. 58) to Ode 180Google Scholar. Compare also Chou li, 33.2.a (Ssu-pu pei-yao) (Biot, E., Le Tcheou-li, 2, p. 256)Google Scholar, “In the spring he sacrifices to the Horse Ancestor.” The commentary identifies Ma tsu as another name for the constellation Heavenly Four, but, though one rationalistic commentator objects that horses, unlike men, do not have ancestors to be sacrificed to, there is obviously no contradiction in the Horse Ancestor's being both a god and a group of stars, especially when the stars are the object of a cult. A different commentator to Ode 180, not followed by Karlgren, interprets the passage as referring to a sacrifice at the camp rather than one to the Horse Ancestor, and this interpretation is preferred by von Dewall, M., Pferd und Wagen im frühen China (1964), p. 64, n. 44Google Scholar. Miss von Dewall thinks that horses were neither the object of worship nor used as sacrificial animals in Shang and early Chou times. I find it difficult to dismiss the independent evidence of the Mao commentary and the Chou li as having no basis in some genuine tradition, but, in any case, there can be no denying the part that divine horses, or horse-like creatures, played in mythology.

^{page 32 note 4} Kuan-tzu, 20 (p. 27 in ed. of Chu tsu chi ch'eng); Mo-tzu, 33/19/45 (Mo-tzu yin-te).

^{page 32 note 5} Huai-nan-tzu, 6 (p. 95 in ed. of Chu tzu chi ch'eng).

^{page 32 note 6} ibid., commentary: “Fei-huang is ch'eng-huang. It comes from the west. It is like a fox in appearance with a horn (or horns) on its back. It lives for 1,000 years.” I Chou shu, 59 (ch. 7.8b) has a passage which may be the source of both the Shan hai ching and the commentary to Huai-nan-tzu. It says that ch'eng-huang resembles the chi'i .

^{page 32 note 7} Han shu, 22.0382.3; Waley, A., “The Heavenly Horses of Ferghana”, History Today (1955), 102Google Scholar.

^{page 32 note 8} There is, naturally, a considerable amount of scholastic lore abou the ch'i-lin. A representative selection is found in Erh-ya shu-chu, 10, under , an alternative writing for .

^{page 33 note 1} Shang chün shu, 18 (Duyvendak, , The Book of Lord Shang (1928), p. 293)Google Scholar; Chan-kuo ts'e, 11.3b, 11.6a (Ssu-pu pei-yao). In these passages ch'i-lin is coupled with lu-erh , the name of one of King Mu of Ch'in's horses. Elsewhere one finds ch'i-chi coupled with lu-erh. But note that the Kuang-yün defines ch'i-lin (written with the horse radical) as a white horse with black spine, (cf. the stripy backs that were a mark of Heavenly Horses in Han times, Waley, op. cit., p. 102.)

^{page 33 note 2} Lun heng, 11.10a, 14.3b (Ssu-pu pei-yao).

^{page 33 note 3} Huai-nan-tzu, 3 (p. 36, ed. of Chu tzu chi ch'eng). The text cannot mean that the unicorn(s) attack and eat the sun and moon as Maspero interpreted it (Journal asiatique, 204 (1924), p. 20Google Scholar) and following him Granet, , Danses et légendes de la Chine ancienne, (1926), p. 374, n. 3Google Scholar.

^{page 33 note 4} Huai-nan-tzu, 4 (p. 65). The pre-eminence of the ch'i-lin among swift-footed animals, like that of the fenghuang (phoenix) among fliers and T'ai shan among mountains and hills, is referred to in Mencius II, A.2.28. cf. also K'ung-tzu chia-yü, 6.4b (Ssu-pu pei-yao).

^{page 33 note 5} Waley, Book of Songs, no. 206. I have slightly modified Waley's translation by rendering (like Karlgren) the exclamation hsü-chieh simply as “Oh!” rather than “Alas”.

^{page 33 note 6} Gloss 18 (BMFEA, 14 (1942), p. 91)Google Scholar.

^{page 33 note 7} Van Huyen, Nguyen, Les chants alternés des garçons et des filles en Annam, Paris, 1934, p. 18Google Scholar.

^{page 34 note 1} See von Dewall, M., Pferd und Wagen im frühen China, pp. 121 ff.Google Scholar, with citations.

^{page 34 note 2} Shuo-wen defines as a “beast of virtue with the body of a fallow deer and the tail of an ox”. It then defines lin as a female ch'i. Lin is defined separately as a “large female deer”. The Erh-ya, on the other hand, has another character which it defines as “like a horse with one horn”. It then adds “One without a horn is a ch'i ”. It defines lin as “having the body of a fallow deer, the tail of an ox and one horn”. It says nothing about the sex of either the ch'i or the lin.

^{page 35 note 1} It is interesting to note that in the animal symbolism attached to the hexagrams of the Book of Changes ch'ien is associated with the horse. See I ching, Shuo-kua, 7.

^{page 36 note 1} Shih-chi, 123.0267.1.

^{page 36 note 2} See Shih-chi chi-chieh in Shih-chi, 130.0281.4. cf. also 12.0042.4.

^{page 37 note 1} Ts'ui Shih , Shih-chi t'an-yüan (1910), 8.15a, b.

^{page 37 note 2} T'oung Pao, 26 (1929), 178, n. 1Google Scholar.

^{page 37 note 3} ZDMG, 91 (1937), 250Google Scholar.

^{page 37 note 4} Shih-chi, 130.0281.3.