OVER THE ARKA-TAGH 1017 
at the heels of a troop of khulans behind the hills on 
the right of our line of march. I was afraid he had 
continued his pursuit too far, and lost the trail of the 
caravan, and so had gone quite astray. I sent a man 
back to the place where the dog was seen last ; it was 
where we left a donkey that was completely exhausted. 
But the man came back without having seen any trace of 
the dog, and he left the donkey behind dying. A couple 
of horses, which had also given up, were on the other 
hand brought into camp. 1 he camels stood the fatigues 
of the journey best. 
Once more it was evening. I had my tea, and bread 
and rice pudding, and then smoked a pipe or two, whilst I 
worked out the notes I took on the road during the day. 
When I lay down to sleep, I felt quite lonely without 
Yolldash, for he was my best companion, and ate and slept 
at my side. But at three o’clock in the morning I was 
awakened by something pushing against my bed. There 
was “his lordship,” wagging his tail and licking my face, 
literally beside himself with joy. The poor beast had been 
on foot all the afternoon and night seeking us ; he was 
actually too tired to eat, but with a sigh of intense relief 
curled himself up at once in his usual sleeping-place. 
August 28th. The ground was quite level, and with the 
exception of the two sick horses, all the animals marched 
well. The Southern Mountain Range was now clearly 
distinguishable ; its snowfields were at a less altitude than 
those of the Arka-tagh. This day we crossed another 
low pass. The watershed we passed over the day before 
must therefore have been of secondary importance only, 
especially as the intricately winding stream, which we had 
followed most of the day, curved round it on the north. 
Now however the drainage channels all flowed definitely 
towards the east, converging upon a tiny freshwater lake. 
East of the pass stretched a broad shallow valley, to all 
appearance quite level ; but it had a slope, as was proved 
by a deep dry watercourse. 
There was an abundance of grass everywhere in that 
region, and consequently great numbers of khulans. One 
11.-23 
