117. CARACHOÇO 161
course the bonito, a staple food of the Maldive Islands, where it is also dried and from where it is exported to various ports of the Indian Ocean; it is the « cobily mash », in Maldivian Kalubili-mds, «black bonito fish » (cf. YULE, Hobson-Jobson2, 223-224; TP, 1933, 416-418; DALGADO, Glossdrio Luso-Asidtico, I, 186-187, « calmelas »; on the word « bonito » itself, cf. YULE, HobsonJobson2, 104-105, and DALGADO, Gloss6rio Luso-Asiritico, I, 137-138). But nowhere can I find that the cachalot has any taste for the bonito. Nor do I expect to find such information, since the oesophagus of the true whale is only two inches wide, and the cachalot, although his oesophagus is not so narrow, mainly feeds on cephalopods. Some confusion seems to have occurred, but I do not know how nor why.
This is Qara-gojo or Qara-hô)o, the ancient Uighur capital, about 17 miles east of Turfan. It was at first mistaken by BENEDETTO (B, CLXIX) for Qara-boto (see « Eçina »), but the error was immediately corrected (Journal des Savants, 1929, 42). The « Carachoco » of Z was altered to « Carachoco » in the late Milan copy (cf. B, 46). It is to be regretted that « Caracoco », although this time correctly identified with Qara-gôjo, should have been retained in B', 73, 440.
In 1912, I devoted a whole paper (JA, 1912, I, 578-603) to the history of the name (for Qoco in Albirûni and Kasyari, cf. Mi, 271), and tried to establish the phonetic evolution r A Kao-ch'ang - Qoco > Huo-chou > Ha-la-huo-chou. As a matter of fact, I had been anticipated in deriving Qoco from Kao-ch'ang by a writer of the Mongol period, Ou-yang Hsüan (1273-1357; GILES, Biogr. Dict. No. 1593), in his « family memorial » of a noble Uighur family ( Yüan wen lei, 70, 10 a). We do not know under what influence the epithet gara, « black », came to be prefixed to the name. Judging from analogous cases, one may be tempted to believe that it was when the importance of Qoco began to be superseded by the growing influence of Turfan. But although Turfan is a much older place than it is generally taken to be (I think I can trace its name as far back as the first half of the 7th cent.), the form Qara-gôjo, at least in Chinese transcription, occurs already in the second half of the 13th cent., and the many mentions of that city in the YS, to the exclusion of Turfan, leave no doubt that it remained the principal city of the Turfan region down to the end of the 14th cent.
The form Qoco is now well attested in mediaeval Turkish texts. The corresponding Chinese transcription was at first RI .Jlj Ho-chou in 913 (Liao shih, I, 3 b; cf. also 36, 10 b), in 1131 (Chin shih, 3, 7 a; « 1130 » is a mistake in Br, I, 222), in works of the early 13th cent. (cf. Br, I, 16, 32), and even in 1279 and 1282 in the YS (10, 8 a; 12, 7 a). Other forms then occur, such as ¶ -Jill Ho-chou in 1281 (YS, 11, 5 b), and A ,'ill Huo-chou which is still to be met much later in Ming shih, 329, 8 a. All these transcriptions are at the same time chosen in such a
IIIF Color HighRes
