Serindia : vol.3 / 82 ページ（白黒高解像度画像）

0082	Serindia : vol.3
セリンディア : vol.3

Irrigation
resources of
Hāmi
limited.

The existence of the Hāmi oasis is due solely to the irrigation facilities derived from its
vicinity to that easternmost portion of the T'ien-shan which, as its name Karlik-tāgh shows, rises high
enough to catch adequate moisture and bear permanent snow-beds. But, as seen in Maps Nos. 72, 76,
the length of this snowy end of the T'ien-shan is not great—only about 25 miles or so on the crest
line—and the total volume of drainage from the range is not sufficient to pass on the surface over
the piedmont gravel slopes at its south foot except on occasion of rare rain-floods. Much of the
water is lost there by evaporation. Hence only the subsoil drainage coming to light in the form
of marshy springs at the lower edge of the gravel glacis, after the fashion of the kara-su of the
Khotan-Keriya region,⁴ is permanently available for the irrigation of the fertile loess belt further
down. The result is that, leaving aside the small patches of arable ground to be found in the
confined valleys of the Karlik-tāgh, cultivation in the Hāmi tract is restricted to the narrow strips
of fertile soil which canals taken from the low-lying spring basins just referred to can command.
Even in the case of the main oasis the greatest width is only about five miles, and a good deal of
the area thus included is not capable of cultivation owing to marshy soil or for other reasons.
There are no rivers sufficiently large to fertilize alluvial fans of corresponding size from the point
of their debouchure, as is the case at Khotan, Yārkand, Kuchā, etc. On this account Hāmi
cultivation, in spite of the fertility of the soil which all Chinese accounts rightly emphasize and
of favourable climatic conditions, could never have supported more than a limited population.⁵
In this respect, too, the analogy to Lou-lan holds good.

Hāmi as
compared
with Lou-
lan.

But if both Lou-lan and Hāmi were thus destined to serve as bridge-heads for the Chinese
during successive periods of advance beyond the deserts bounding Central Asia on the east, the
difference of geographical position between the two territories necessarily made itself felt otherwise.
When discussing Lou-lan and the ancient route leading to it from the Kan-su marches, I had
occasion to point out how well protected this route was against hostile interference from the north.⁶
The broad desert belt of the Kuruk-tāgh with its utterly barren plateaus provided an effective
natural defence against raiding attacks from the north of the T'ien-shan, where Huns as well as
their nomad successors found their grazing grounds. On the other hand, we have also noted how
precarious the use of that bridge-head must have been from the first owing to special physical
difficulties besetting irrigation. These are always inseparable from the maintenance of a distant
terminal oasis in the desert; aided probably by progress of desiccation, they led in the end, as we
have seen, to Lou-lan being abandoned altogether in the fourth century A. D.⁶ᵃ

In the case of Hāmi we find these conditions exactly reversed. There the water-supply
needed for cultivation, limited as it is by nature, could never have been seriously threatened during

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

引用情報

OCR読み取り結果

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

引用情報

OCR読み取り結果

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