476 THE RUINS OF AK-SIPIL AND RAWAK [Chap. XIV
walls. Over the patches of ground which now lay again clear of sand the small rectangular embankments by which fields are divided at the present day for irrigation could be quite clearly made out. The labourers with me correctly recognized them at once. Just as at the site of ` old Domoko ', the ground formerly occupied by fields showed a relatively firm surface b. Within the area once enclosed by the circumvallation pottery débris was very scanty, perhaps, as M. Dutreuil de Rhins assumed, owing to the accumulation of drift-sand over the eroded ground. But outside at some distance from the line of walls it cropped up on numerous bare patches.
With the exception of a single Chinese coin showing the characters wu-chu (see Plate LX X X IX, 16), which was picked up on the isolated wall portion westwards, no antiques were found during my stay in the immediate vicinity of the ruins. But the latter themselves possibly furnish some indication of their age in the marks which appear on most of the loose-lying bricks, and which M. Dutreuil de Rhins duly noticed. Some of the bricks thus marked, more or less fragmentary, are seen in the photograph reproduced in Plate XV I I I 6. The marks consist of coarse straggling indentures about half an inch broad and a quarter of an inch deep. Several of them seemed to recall Kharosthi characters, especially ka and ga, while others showed a resemblance to the general type of that script. As bricklayers' marks of this kind might have remained in use long after current writing in the script had ceased, it would scarcely be possible to draw a definite chronological conclusion from them, even if their Kharosthi derivation could be established with certainty. But the general impression I gained from the construction of the walls and from the sculptured remains of the neighbouring site of Kighillik, to be discussed presently, favoured the assumption of an early date.
` Ak-sipil ' serves as a popular designation for the whole of the débris areas stretching north of the Yurung-kash and Hanguya. Hence the evidence of the coins and antiques sold to me at Khotan as coming from ` Ak-sipil ' would prove little or nothing for the age of this particular ruin, even if such statements could always be accepted with confidence. Among the thirteen old copper coins alleged to have been brought from Ak-sipil Appendix D shows one small Sino-Kharosthi and one wu-chu piece, four coins of the Chien-yüan period (758-759 A. D.), two Sung coins of the eleventh century A.D., as well as three early Muhammadan coins, one of these (Plate XC, 42) with the name of Muhammad Arslan Khan. The range covered by these coins is thus similar. to that in the case of Yôtkan, but the local attribution is here quite uncertain. The other antiques comprise some terra-cotta grotesques (A. ooi. a, b ; A. 006. c), which closely resemble the Yôtkan type, but show marks of erosion. There is among them also a number of seals in stone and bronze, undoubtedly ancient, but difficult at present to fix chronologically, as there are no legends, and the devices are not sufficiently pronounced in character (see A. ooi. c ; A. 002. a ; A. 004. a, b, c ; A. oo6. a, b, all in Plate L). Among them the stone seal (A. oo6. a) is of interest ; it is engraved on two sides, showing a humped bull on the one and a fire-altar on the other. Yôtkan, as well as the Niya Site, have furnished seals not unlike this series.
My surmise that the high sands to the west and south-west of Ak-sipil might possibly hide more structural ruins was verified when, on April 8, Turdi guided me to the remains of what he had previously talked of as ` the Bat-khana ' or temple of Ak-sipil. They were found to be situated at a distance of about i Z miles to the south-west of the ruined fort, in a small