earliest advance into Central Asia. Hence the accounts in the Chinese Annals of the Former Han and succeeding dynasties often mention the territory, first under the name of Lou-lan and subsequently under that of Shan-shan.
Conclusive evidence shows that the small oasis of Charkhlik was already the chief place of Lop when Hsüantsang made his way through it on his return to China about A.D. 645, and that it had been so for several centuries before. Its archaeological remains proved scanty, as might be expected on ground which had remained cultivated for a long time or been reoccupied after abandonment, as it is now again to-day. All the same Charkhlik was a place of importance to me; for it was at this last inhabited locality that I had to make my preparations for my long-planned exploration of the ruined site of ancient Lou-lan, situated in the desert to the north of Lop-nor and first discovered by Dr. Hedin in 190o.
In the next chapters I shall describe the trying but very fruitful and interesting labours accomplished on my second and third expeditions into that forbidding and now utterly waterless desert. But before I do so I may give some account of the discoveries which attended my excavations at the ruins of Miran. I first reached them after setting out from Charkhlik on December 7, 1906, for the Lop desert. Having by some rapid trial digging ascertained their importance, I returned to them by the close of January for thorough excavation. The site proved a very desolate spot, situated about fifty miles to the north-east of Charkhlik at the foot of the absolutely barren gravel glacis which stretches down from the K`un-lun mountains towards the westernmost portion of the Lop-nor marshes. The latter
IIIF Color HighRes
