国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

elaboration of these results will take time, and it is only in the proposed detailed report on my third
expedition that I can hope to present them. To this publication I must accordingly leave the record
of the observations concerning the geography and history of the Turfān region in general that I was
able to make, and the discussion of the many interesting and complex questions to which they give
rise. In the present place I shall limit myself strictly to a record of the modest exploratory work
I found occasion to carry out at a couple of smaller sites, and to the information needed as regards
the origin of certain antiques which were acquired elsewhere.

March from
Pichān to
Lukchun. The hope of being able to study topographical and archaeological facts which might throw light
on the subject of desiccation, as important in Turfān territory as in areas of ancient occupation
within the Tārim Basin, induced me to start my tour at the south-eastern end of the basin. There
Captain Roborovsky's map marked the ruins of Chong-hassār, the 'Big Castle', also locally known
as Hassā[r]-shahrī, not far from the extreme eastern end of the marshy salt-lake bed which forms the
deepest portion of the Turfān depression and gathers whatever is left of its surface drainage (see
Map No. 59. c. 2). The ruins were declared by my Pichān informants to be situated on ground which
was now wholly desert, and my archaeological predilection for the latter helped to decide me to
make a start there. Our march on November 13 led down the barren narrow valley, where the
waters of the Pichān tract disappear between bare clay ridges and drift-sand, and brought me to the
flourishing oasis of Lukchun, watered mainly by the stream coming from Lamjin. Under the name
of Liu-chung 柳 仲 it figures already in Han times as the seat of the 'Chang-chih' of the Western
countries and as a chief foothold of Chinese power.³ Barometrical readings indicated for Lukchun
an elevation of only some 50 feet above sea-level.

Change in
irrigation
methods. For a description of the interesting ground crossed next day on our march to the south-west
I must refer to my Personal Narrative.⁴ Here it will suffice to mention that irrigation from newly
constructed Kārēzes is steadily replacing the precarious cultivation formerly carried on with the
water that the canals of Lukchun could bring down in favourable years to this outlying area. The
immediate cause of this change, which is proceeding in most of the Turfān settlements, is certainly
the increased pressure of population, following the re-establishment of peace and prosperity since
the Chinese re-conquest. But since Kārēz construction is admittedly a modern innovation in the
Turfān region, not dating back further than the end of the eighteenth century at the earliest, we can
account for the far larger population in ancient times, to which a variety of archaeological and
historical indications point, only by assuming that the water-supply available from surface drainage
in former periods was more plentiful than it now is.⁵

Evidence of
desiccation
about Bēsh-
tam. Clear evidence of the desiccation here implied, whatever its cause, period of commencement, or
duration, was forthcoming on the approach to the site. At the farm of Bēsh-tam, some seven miles
from the southern edge of continuous Lukchun cultivation, the last small patches of Kārēz-irrigated
land were left behind. Beyond we passed fields long ago abandoned and overrun by thorny scrub
of the desert, and then crossed a shallow bed in which the water of the Lukchun canals, when left
unused in the winter, endeavours to make its way down to the terminal lake-bed. On account of
evaporation and absorption in the soil, no water could now ever reach so far down during the spring,
when it is most needed in the oasis, or during the terrible heat of the Turfān summer, unless,
perhaps, at the time of quite exceptional rain-floods from the distant mountains. To the south of
this temporary overflow-bed there stretched a wide sandy plain with plentiful thorny scrub and
small tamarisks growing amidst rudimentary dunes. Wherever the ground was left clear of drift-