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

40 miles, but brought its reward in a number of interesting indications, though I was not able to
follow these up until fully five years later. Guided by Ayib Mirāb, an observant old village head-
man in whose orchard I had camped, I rode first through the whole length of the main oasis to the
north-east. On the way there was plenty of evidence of the great-extension which Marāl-bāshi
cultivation had undergone since the Chinese after the reconquest had set about to develop the agri-
cultural resources of this important route junction. At the same time it was easy to see how much
this development was hampered, just as at Ak-su, by the fact that the population consists almost
solely of Dolāns, who until a generation or two ago still lived mainly as semi-nomadic herdsmen.

Beyond the edge of the fields watered from the Töpa-nör we passed into an area covered with
luxuriant scrub and tamarisks which extends right away to the barren hills of the Bel-tāgh and
Ökur-mazār-tāgh. The whole had obviously once formed a huge inundation basin of the Kāshgar
River, and must, in periods when the climate was moister and the river's flood volume larger than it
is now, have been quite impassable for traffic during a considerable portion of the year. In this fact
we may well have one of the reasons why the ancient route from Ak-su to Kāshgar, as already sug-
gested, probably followed a line leading further north along the foot of the outermost Tien-shan.
But this and kindred questions must be left for discussion in connexion with my surveys of 1913.
The growth of vegetation in this old marsh bed was so dense that it would have been difficult to
reach our goal through it in time but for expert guidance. This was supplied by young hunters
picked up at a small isolated holding which Tālib Hāji, an enterprising pioneer cultivator, had
reclaimed in the midst of the jungle.

Following winding tracks, with here and there almost wholly overgrown traces of what my
guides called the 'old road' from Tumshuk, we approached at last the north-western end of the
Ökur-mazār-tāgh, here known as Achal-tāgh. Here I found dried-up pools marking a last remnant
of the northern extension of the great marsh now reclaimed for the most part by Chār-bāgh. What
from the distance had looked like a continuous ridge now proved to be broken by a narrow defile.
Through this passes to the south-east a still clearly recognizable flood-bed which at times of great
inundation was said to have carried water within living memory to the now utterly dry east foot of
the Ökur-mazār-tāgh.

An isolated and precipitous rocky ridge, about a mile long and over 200 feet high in the centre,
separates this gully from the wider defile at the foot of the Bel-tāgh. This ridge (Fig. 332) is known
as Arach and bears at both its ends remains of manifestly ancient defences. That their purpose was
to close both gaps in the hill chain, so as to control any traffic which might pass through them from
the side of Tumshuk or vice versa, was made perfectly clear by the configuration of the ground.
The three hills of Ökur-mazār-tāgh (Achal-tāgh), Arach, and Bel-tāgh for a total distance of over
18 miles form a natural rock rampart which, except for the two gaps flanking Arach, can nowhere
be crossed except on foot and by difficult climbing, as can be seen from Figs. 332, 344. Considering
that the ground at the south-east end of the Ökur-mazār-tāgh is even now very marshy and in quite
recent times was impassable, the importance and natural strength of the defensive line thus furnished
by nature is obvious. The only points needing to be watched were the defiles on either side of
Arach. At the south-eastern end of this hill a low rocky spur was covered on its top for about
60 yards with foundations of walls built of sun-dried bricks, which are likely to mark a watch-station.
Burrowings of 'treasure-seekers' had badly injured them. High up on the same side there rose
a small square watch-tower, still more than 10 feet high, in a position very difficult of access over
precipitous and perfectly smooth rock faces.

Following the Arach ridge towards its north-west end I found its crest occupied by a solidly
built watch-tower, seen in Fig. 332, measuring 24 feet at its base and built of bricks about 14″ × 9″ × 3″