than in others. In the latter the dunes are seldom more than 3 m. high, and the bajirs are bigger. The steep slopes are here turned to the west-south-west and the west, suggesting that the course of the prevailing wind forms a curve, being more northerly in the north and more easterly in the south; and this inference is further supported by the situation of the jardangs. In the north-east part of the desert the wind-eroded gullies stretch from north-north-east to south-south-west, sometimes even from north to south; here however they extend from north-east to southwest or even from east-north-east to west-south-west. Yet the sand is present in such preponderance that the jardangs obtrude themselves but little, their edges projecting only here and there; yet they often resemble detached table-like masses. Mounds occur also, though but seldom. The floor of the bajirs was not quite so level as it had been hitherto, but was slightly undulating, and sloped very gently upwards towards the south-west, though this does not of course imply that the desert as a whole rises in the same direction. The rise in each bajir is merely caused by the local undulation of that part of the desert, and no doubt this is also the cause, not only of the formation of the bajirs, but also of the unequal distribution of the sand. The reason the desert is less level here than farther north is that this locality forms a threshold or dividing-ridge between two basins, and has not been under water for a very long period, though because of its varying consistency it has been attacked by the wind with varying degrees of effect. We no longer saw any dry toghrak wood, and the fragments of tamarisk that we passed occasionally were perfectly loose and separate, that is they were never rooted. Curiously enough, we again saw a couple of living tamarisks. Evidently new-corners, they were standing in the very lowest depressions, and if they did possess mounds, the mounds were certainly low. The ground all round them was every bit as barren and arid as elsewhere.
Fig. 143. JARDANGS IN THE SAND.
After that the desert all the way to the northern shore of the newly formed lakes presented the features which I have already described. The general conclusions that admit of being drawn from this crossing of the desert will be discussed in connection with the second crossing farther to the east. This I will now proceed to describe.
The short distance from Altmisch-bulak to the ruins of Lôu-lan was traversed by a different route from that taken before. Leaving the spring of 1st March i 90 I , we directed our steps towards the south-west, and then towards the south, and thus evaded entirely the great eroded torrent of Altmisch-bulak. This latter lay to the east of us, while to the west we had the route from the desert to the springs and the low spur of the mountains which we crossed over before. The surface was hard and broken, and strewn with gravel, and there were small knobs of much weathered green-stone sticking up in several places. For a time we followed an eroded torrent, running
Hedin, ,journey in Central Asia. II. 3o
IIIF Color HighRes
