Sec. iv] ABOUT KASHGAR AND YARKAND 85
interval of centuries. Here, as in the case of almost all chance finds of this kind in Turkestan which have not been followed up at the time, the only critical verdict can be non liquet.
Riding about two miles further north over waste ground covered with low scrub, I found numerous ruins of small mud-built houses scattered over an area which, by its clearly traceable irrigation channels, its terraced fields, the almost complete absence of drift sand, and similar indications was plainly marked as having been occupied down to a comparatively recent period. These remains, closely resembling the dwellings of modern villagers in these parts, were found in small detached groups extending for about a mile from north-west to south-east. The greater part of the locality was known to my guides as Koilog h-ata, while they gave the name of Talâri-zemin to its northern end. The survey effected two years later by my assistant Rai Bahadur Lal Singh on his march from Merket to Karghalik showed that a small enclave of actual cultivation near the farms of Kököl and Lai-dang (see Map No. I I) reaches within a few miles of the ruins. These, no doubt, dated from a period when the narrow belt capable of irrigation between the desert edge and the Tiznaf River down to Merket had seen more continuous cultivation than at present. Sir G. Macartney had visited Koilogh-ata earlier in the year, and, stimulated by the find of the Uigur documents in this vicinity, had the rubbish in one or two of these humble dwellings cleared. But the only discovery rewarding his excavation was part of a leather slipper which of course, in the absence of dated relics, could afford no chronological fixing. So there was no inducement for me to spend time over further clearings.
SECTION V.—ALONG THE WESTERNMOST KUN-LUN
After traversing the large and flourishing oasis of Karghalik from north to south on July 6 and 7, I made my way south across the wide gravel glacis of the westernmost Kun-lun to the village of Kok-yâr (Map No. I2) through which a much-frequented caravan-route passes to the headwaters of the Tiznaf and Yarkand Rivers and thus across the Kara-koram to Ladak. The stay of sixteen days which I made at Kök-yar (July 9 to 24), and of which an account has been given in Chapter XII of Desert Cathay, was solely intended to secure me, at the foot of the mountains and thus in relative coolness, the peace needed for the completion of my last tasks in connexion with Ancient Khotan. These kept me so busy throughout that I felt almost glad for the absence of any archaeological distractions in the vicinity of this little submontane oasis. Yet the observations I was able to collect there and on my subsequent move eastwards along the foot of the mountains proved useful as regards both the ethnography and historical geography of this region. I shall offer some supplementary notes on the latter first.
In Ancient Khotan I have already set forth in detail the reasons which induced me to identify Karghalik with the kingdom which Hsüan-tsang calls Chê-chü-chia, and which in the Tang Annals and Sung Yün's itinerary figures under the variously spelt names of Chu-chü po and Chu-chii pan.' Now the special notice of the Tang Annals states explicitly that this kingdom was the same as ` the kingdom of Tzû-ho of the epoch of the Hans. It has annexed and possesses the territory of the four peoples called Hsi yeh, Pu-li, I-nai and Tê jo.' 2 In the Chien Han shu there are separate short notices of Tzû-ho, Hsi-yeh, P`u-li, and I-nai.3 But the statements there made as to the relative positions, etc., of these localities show various discrepancies which cannot be satisfactorily cleared up without access to the original text. Thus we cannot conclude more for the present than
' Cf. Ancient Khotan, i. pp. 89-93 ; for the varying forms of the name cf. also Chavannes, Voyage de Song Yun, p. r9, note 4, and Turcs occid., p. 366.
2 See Turcs occid., p. 123.
3 See Wylie, J. Anlhropol. Ins!., x. pp. 31 sq. Here, as
elsewhere in the case of the important account of the
Western Regions ' furnished by Book xCVI of the CHien Han shu, the need of a new translation by a critical scholar, familiar with the results of modern geographical and antiquarian researches, is sadly felt.