782 ON THE ANCIENT ROUTE ALONG THE KONCHE-DARYA [Chap. XXI
route that once passed from Ying-p`an to Korla, and a few general remarks may fitly conclude our survey of them. We have seen that the best preserved of these towers show a very close resemblance in all structural features to those of the Limes with which the Emperor Wu-ti had protected the line followed by early Chinese trade and military enterprise towards the Tarim basin as far as the termination of the Su-to-ho. Similarly the disposition of the towers at points and distances which would permit of the communication of fire signals, such as are frequently mentioned in the Han records recovered from the Tun-huang Limes, closely agrees with that on the line of detached watch-stations traced by me that guarded the south-western flank of the Limes on the side of the terminal basin of the Su-to-ho.11 The importance of protecting by similar means that continuation of the great line of communication which lay beyond Lou-lan must have suggested itself with all the more force to those who directed the development of the great Han Emperor's ' forward policy ', because it was just along the western foot of the Kuruk-tagh that the danger of Hun raids was the greatest ; for this region afforded both water and grazing for raiders, while on the other hand missions and caravans could obtain the shelter of permanent settlements only near its northern end.
For the belief that the line of watch-towers from Ying-p`an towards Korla goes back to this early stage of Chinese expansion some direct support may be found in an important contemporary notice preserved in Ssû-ma Ch`ien's history and also embodied in the ' Notes on the Western Regions ' in the Former Han Annals. We are told by SA-ma Chien that after the success obtained by Li Kuang-li's second expedition against Ta-yuan or Farghana (IO2—IOI B. c.) a military governor was established in the Tun-huang region : ' westward from here to the Salt Lake the road at many points was protected by military stations, and in Lun-t`ai there were several hundred soldiers stationed as farmers, the special commissioners in charge of the farms being required to guard the cultivated land and to store the crops of grain for the use of embassies abroad '.12 Ssû-ma Ch`ien's great work was in all probability completed about 99 B. c. or a few years later. This notice of a military agricultural colony at Lun-teai accordingly takes us back to the very time when the Limes west of Tun-huang was being constructed, as proved by the documents there excavated.l3 As Lun-t`ai can safely be located at the oasis of Bugur on the high road west of Korla (Map No. 2I. A. I), the establishment of watch-stations on the line of communication which leads to it may well be assumed to belong to the same years. How long the series of posts then constructed along it continued to be occupied and used for their original purpose we have no means of determining. But that they served as convenient halting-places on a route which was probably frequented by traffic to and from the Lop side long after Lou-lan was abandoned may be safely concluded from the remains of paper documents, the 'rang coin, and other small relics found among their refuse.
There still remains for consideration the question as to the probable continuation of this line of watch and signal towers towards the north. I did not learn of other remains of this kind towards Korla, from which Y. ix, at Suget-bulak, is separated only by a direct distance of about twelve miles. But the plateau that lies to the east of the road connecting Shinega with Korla would have furnished a very convenient position for an intermediate post, and owing to its elevation above the level plain on either side no tower of any height would have been needed there for transmitting signals. The same remark applies also to the final offshoot of the Kuruk-tagh which separates the plain about Korla from the westernmost part of the Baghrash lake basin north-eastwards. A point
11 Cf. Serindia, ii. pp. 633 sq., 641 sqq.
12 See Hirth, `The Story of Chang K`ién', J.A.O.S., xxxvii. p. 116. The corresponding notice in the Former Han Annals
mentions not only Lun-t`ai but also Ch`ü-li as provided with a military colony ; see Wylie, J. Anthrop. Inst., x. p. 22.
13 Cf. Serindia, ii. pp. 728 sqq.