IN SEARCH OF THE LIMES TO SU-CHOU
SECTION I.-THE LIMES LINE NORTH OF THE SU-LO-HO
MY short stay at An-hsi, from the mth to the 17th of April, was wholly absorbed by pre- Prepara-
parations for the journey which was to take me through the north-western marches of Kan-su, and tiollS at
by multifarious writing work. My previous stays had already enabled me to familiarize myself with the geographical features of the oasis of An-hsi, corresponding to the ancient Kua-thou A 3f1; these have invested it with importance in spite of its modest local resources, ever since in the first century A.D. the desert route leading from it north-westward to Hami was opened as the main line of communication between China and Central Asia. I have fully discussed in Serindia the part which, by reason of its position at one of the great cross-roads of Asia, it has played in the history of China's relations with the ` Western Regions '. I have dealt in the same work with its local limitations and scanty remains.' Hence I may turn at once to the, programme which I planned for my explorations eastwards, and to the preparations it implied.
The task I had set myself for the spring was to trace and explore the line of the Limes of Planned
Han Wu-ti from the vicinity of An-hsi to the north-east of Su-chou and thus to solve a problem explorations
that my rapid surveys of 1907 had raised but not allowed me to clear up. I intended subsequently Etsin-gol.
to move down the -Etsin-gol, which carries the united waters of the river of Su-chou and Kan-chou, to near its terminal basin ; this was ground which, in view of its physical features and a recent discovery made there by Colonel Kozlov, the distinguished Russian explorer, held out promise of combined geographical and archaeological interest. It was important to complete this double task, practically all to be carried out in desert areas, before the great heat set in, and therefore to move quickly.
This consideration, combined with regard for our camels, which had to be spared all needless Impedi-
exertions at this season, made it necessary to limit impedimenta as much as possible. For this naenta
purpose all baggage had to be carefully sifted with a view to leaving behind whatever was not Ya-mên.
needed for the work of the spring and summer. In Igo7 the Ya-mên had proved a suitable place for storing our surplus belongings, and as the local magistrate, now a modest hsien-kuan, and no longer encumbered with the dignity attaching to a prefect, was kindly disposed, I was again able to entrust to its safe keeping all our spare baggage, including whatever antiques we had brought along since leaving Lou-lan. Faithful Ibrahim Beg once more remained behind to mount guard over the Ya-mên store-room and to make sure that its contents were kept dry—a necessary precaution after the experience gained in June and September, 1907, of An-hsi's liability to occasional rain from the mountains.
Before our start from An-hsi my party was strengthened by the arrival of a Mongol interpreter Mongol
whom after many vain endeavours I had managed to secure with the help of Zahid Beg. He came interpreter
from a small camp of Mongol herdsmen who had arrived in the vicinity of Tun-huang from the Kara-shahr side. .1 had foreseen from the first the necessity of having with us a Mongol possessing a knowledge of Turki to facilitate our work during our proposed visit to the Etsin-gol tract on the
1 See Serindia, iii. pp. 1090 sqq.