________________
199
SEPTEMBER, 1917] A THIRD JOURNEY OF EXPLORATION IN CENTRAL ASIA
The route followed during the first marches acquainted me with a series of old Buddhist cave temples at Ma-ti-ssu, containing sculptures of Sung times, and with other interesting Buddhist remains in the pretty little town of Nan-kou-chêng at the foot of the mountains. The visit did not pass without profit for my collection of antiques, and also helped to make me realize that we were now near a dividing line of distinct geographical interest. For while to the west cultivation, whether in the plain or along the foot of the mountains, requires irrigation, we now came upon loess slopes and big alluvial fans which rainfall alone suffices to make fertile. Our approach to the watershed of the Pacific Ocean was appropriately foreshadowed by this marked change in climate conditions.
Following the route which leads towards Hsi-ning and ascending through the picturesque gorge and the pass of O-po, we reached the broad valley where the easternmost feeders of the river of Kan-chou gather at an elevation of over 11,000 feet. Thence we were making our way westwards over high alpine grazing grounds frequented in the summer by Tangut herdsmen and horse breeders, when I met with a serious riding accident which might well have put an end for ever to all my travelling. My Badakhshi stallion reared suddenly, and over-balancing himself fell backwards upon me, with the result that the muscles of my left thigh were severely injured. For over two weeks I was unable to leave my camp bed or to use the clutches we improvised. But fortunately the arrangements already made allowed me to let Lal Singh proceed for the topographical tasks I had planned. He carried them through with all his wonted devotion and energy, and no time was lost in our programme. Nearly three weeks had passed when, with my leg still feeling the strain severely, I managed to get myself carried down in a litter to Kan-chou.
During a ten days' halt there I experienced much kindness from Father Van Eecke and other Belgian missionaries, and received the first confused news of the great European conflagration. Then I set out by the third week of August for the long-planned journey through the Pei-shan Gobi. It was to take me back to Turkestan for the work of the autumn and winter. Eight long marches brought me to Mao-mei by a new route skirting the hills on the right bank of the river of Kan-chou, and allowed me to view the remains of the late mediaeval "Great Wall" which runs on to and ends near Su-chou. The complete decay into which it has fallen for considerable distances, notwithstanding its relatively recent origin, helped me to appreciate all the more the time-resisting solidity which the methods of construction employed by the engineers of Han times had assured to their Limes wall. I reached Mao-mei exhausted by the effort which it had cost me to do this journey on horseback, because of the severe strain to my leg. But I found there my brave camels safely arrived and was cheered by Lal Singh rejoining me. By exceptional efforts my indefatigable old travel companion had succeeded in extending our Nan-shan surveys eastwards over an area quite as large as that mapped in 1907.
On 2 September 1914 we commenced the journey which was to carry us right across the great desert area occupied by the ranges of the Pei-shan, where its width is greatest, in the direction from south-east to north-west. The routes wa followed for close on 500 miles had never been surveyed, and I knew that only at one point, the cross-roads of Ming-shui, could we expect to touch ground the position of which was known relative to the routes previously visited by Russian travellers. Wherever possible we moved in two parties and by different routes, in order to increase the extent of the area mapped. For this purpose I had secured at Mao-mei the only two guides available, both Chinese. But