National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

archaeologist of the distant future, had my Ho-shang guide not explained that they served to
guide the course of mounted soldiers when practising musketry exercise at the gallop !

Ruined But an observation of true antiquarian interest awaited me when I reached the point on the
temple by Limes agger where the caravan track to Hāmi crosses it, about 350 yards from the north-west
Limes corner of the ch'êng. A row of five small Stūpas was found here standing just outside and parallel
wall. to the low gravel mound, and another of three more ' within the wall '. These little Stūpas, meant
for funeral monuments, would by themselves have sufficed to attest some lingering sanctity in the
spot. But even more definite evidence that local worship still clung to it was provided by a large
ruined shrine, which rises here immediately to the south of the Limes line and close to the east of
the route. A portion of it had been restored recently, as I was told, in memory of those who fell
while defending Ch'iao-wan-ch'êng against the Tungans, and quarters of Tibetan Lamas in an
outer court showed that the shrine was, indeed, still ' in being '.

Survival I have repeatedly had occasion to show that Chinese sentiment has since early times invested
of local with a kind of religious respect all those places where routes passed ' outside the wall ' of the
worship. Empire.⁵ We are therefore fully justified in assuming that here, too, it is the survival of ancient
local worship that accounts for the presence of the modern shrine and Stūpas, just where topo-
graphy enables us to locate a true ' Gate station ' of the Limes. I was unable to trace any ancient
remains at the spot. But such may well lie hidden under the large modern temple. In any case it
deserves to be noted that the point where the Hāmi route crosses the agger lies exactly half-way
between the ruined watch-towers T. XLI. h and T. XLI. i, these standing a mile distant to the west
and east respectively. The distance of one mile is the usual one between the towers located farther
east along this portion of the Limes. It is accordingly very probable that this point, too, once had
its watch-post well placed for guarding the route to Hāmi.

Advantages It only remains for me to add a few remarks about the geographical and quasi-strategic
of route advantages that favoured the selection of the Ch'iao-wan-ch'êng site as the starting-point for an
guarded. alternative route from the Kan-su marches to Hāmi. They also explain the placing of a garrison
here, probably early in the eighteenth century, to guard this and other possible approaches from
that side. The route, as already mentioned, represents the most direct line of communication
between Su-chou and Hāmi practicable for laden animals. It could not compete in importance
of traffic with the An-hsi–Hāmi route, because the greater height and ruggedness of the successive
Pei-shan ranges on the line it follows precludes the use of carts, for which the more westerly route
starting from Hāmi is on the whole well adapted.⁶ On the other hand, it appears to offer greater
facilities for camel grazing than the latter route, and the same advantage holds good also, as we shall
see, of its eastward continuation towards Ying-p'an (Hua-hai-tzŭ) and Su-chou.

Vicinity of The selection of Ch'iao-wan-ch'êng as a base for this desert route during Manchu times was
cultivable probably influenced by the fact that it was the place to which supplies from the cultivated area
ground. south of the Su-lo-ho could be brought most conveniently. As the Map No. 40. B. 4, 5 shows,
a long belt of cultivation extends towards it from the vicinity of the large village of San-tao-kou
on the Yü-mên-hsien–An-hsi high road. In consequence of the devastation attending the Tungan
raids many farms had been abandoned in this belt, and in spite of the irrigation facilities offered
by the large canal-like branches of Su-lo-ho descending here over the river's alluvial fan, much of
the land still remains untilled. Yet, even so, patches of cultivated ground are to be met with to-day
within two or three miles of the abandoned station. At the same time communication with this
ground from the northern bank of the Su-lo-ho is here particularly easy. The river is here confined