Sec. i] THE LOWER ETSIN-GOL AND ITS TERMINAL BASIN 431
seen, clear proof that the annual supply of water brought down by the river was still sufficient here for vegetation to bind the sandy alluvium deposited by each year's summer flood.
Our route beyond, that day and also on the following day, May loth, lay along the jungle belt that lines the right bank of the easternmost bed of the delta. On the Sai overlooking this three badly decayed towers of clay were passed before we reached the ruined fort known as Bahândurwuljin (` the big fort '), close to the river bank. It consisted of an enclosure, 45 feet square, within walls II feet thick, built of bricks 18" x 9" x 7". It appeared to be of distinctly later date than the small forts near Ta-wan. The belt between this branch of the river and the next one westwards contained abundant vegetation and great stretches of meadow-like ground. It was therefore not surprising to find several Mongol encampments when we crossed to that side in the vicinity of the fort. Though the nights were still refreshingly cool, the marching in the heat of the day had now begun to tell severely on the camels, and those hired from Mao-mei had broken down. So we were doubly glad that the rich grazing which we here came upon gave a chance of securing fresh transport, not indeed without delay and trouble, from the neighbouring camps and from those found near our next camp at Suslun-tora.
The ruined enclosure called Ekki-durwuljin, ` the little fort ', on the right bank lay some distance off our route, and I was unable to visit it. It was stated to be of the same construction as Bah5.n-durwuljin, but smaller. Their position at the two ends of a favourite grazing ground suggested that they were meant as places of refuge for parties encamped in the neighbourhood in case of marauding attacks. Before this a decayed watch-tower built of bricks, 15" x 8" x 3" in size, had been encountered at a point of the route known as Sharakure-sanje. Its position in relation to the two forts suggested that the track we had followed so far marked a route which had been regularly frequented during earlier as well as more recent periods.
After we had covered the first few miles beyond Suslun-tora (Map No. 45. B. 2) on May 21st, the scenery along our route underwent a very distinct change. So far, with the fine groves of Toghraks lining the river-bed, always in rows parallel to its course, and with the luxuriant vegetation extending as far as the eye could reach westwards, the picture was such as I imagined might have met the eyes of travellers by the ancient Chinese high road through Lou-lan, where it passed jungle belts at the head of the Kuruk-daryâ. delta that were still regularly watered by the river. But very soon our route, thereafter lying constantly within a mile or two of the left bank of the easternmost river branch, entered ground very different in character. Instead of soft sandy alluvium, the soil, within a short distance from the river bank, was now everywhere gravel. Vegetation was scanty and often, where convex bends of the bed were skirted, the ground was perfectly bare. Reed-beds were rare and, like the groves of Toghraks met at intervals close to the river-bed, were confined to places where this formed a concave bend with its left bank, thus offering a better chance for the trees to be reached by flood water.
For over forty miles of march on May 21st to 23rd there was no change in this dreary, monotonous scenery. On the right bank of the river branch that we followed, called by our Mongol guides Umne gol, the strips of vegetation appeared quite as confined and narrow. But to add to the desolate look of the landscape, big ridges of dunes appeared on that side between B ôtu-börü and Ulan-sukhe (Map No. 45. c. I), close to the east of the bed, a clear indication that the area beyond is not reached by the floods of the river. As we looked westwards of our route beyond Camp 144, the distant line of Toghraks marking the course of the MIin-gol, the next river-bed on that side, showed for some time across the flat expanse of gravel and then was lost sight of for good. During those marches, until we passed Dzusulun-tsakha on the third day, only two small Mongol camps were encountered, and these too were on the move to grazing grounds farther north. But