. After those two
days, you again come to great mountains and valleys, and extensive
forests, and you continue to travel westward through this kind of country
for 20 days, finding however numerous towns and villages. The people are
Idolaters, and live by agriculture, by cattle-keeping, and by the chase,
for there is much game. And among other kinds, there are the animals that
produce the musk, in great numbers.[NOTE 2]

NOTE 1.--Though the termini of the route, described in these two chapters,
are undoubtedly Si-ngan fu and Ch'eng-tu fu, there are serious
difficulties attending the determination of the line actually followed.

The time according to all the MSS., so far as I know, except those of one
type, is as follows:

In the plain of Kenjanfu . . . . . 3 days.
In the mountains of Cuncun . . . . 20 "
In the plain of Acbalec . . . . . 2 "
In mountains again . . . . . . 20 "
--
45 days.
--

[From Si-ngan fu to Ch'eng-tu (Sze-ch'wan), the Chinese reckon 2300 _li_
(766 miles). (Cf. _Rockhill, Land of the Lamas_, p. 23.) Mr. G.F. Eaton,
writing from Han-chung (_Jour. China Br.R.A.S._ xxviii. p. 29) reckons:
"From Si-ngan Fu S.W. to Ch'eng-tu, via K'i-shan, Fung-sien, Mien,
Kwang-yuan and Chao-hwa, about 30 days, in chairs." He says (p. 24): "From
Ch'eng-tu via Si-ngan to Peking the road does not touch Han-chung, but
20 _li_ west of the city strikes north to Pao-ch'eng. The road from
Han-chung to Ch'eng-tu made by Ts'in Shi Hwang-ti to secure his conquest of
Sze-ch'wan, crosses the Ta-pa-shan."--H.C.]

It seems to me almost impossible to doubt that the Plain of Acbalec
represents some part of the river-valley of the Han, interposed between
the two ranges of mountains called by Richthofen _T'sing-Ling-Shan_ and
_Ta-pa-Shan_. But the time, as just stated, is extravagant for anything
like a direct journey between the two termini.

The distance from Si-ngan fu to Pao-ki is 450 _li_, which could be done in
3 days, but at Polo's rate would probably require 5. The distance by the
mountain road from Pao-ki to the Plain of Han-chung, could never have
occupied 20 days. It is really a 6 or 7 days' march.

But Pauthier's MS. C (and its double, the Bern MS.) has viii. marches
instead of xx., through the mountains of Cuncun. This reduces the time
between Kenjanfu an