under my instructions from L.F.' The ancient refuse he had noted on L.I., together with small relics picked up near it, clearly indicated occupation of that Mesa during the historical period. Its position on the plane-table agreed with that north-easterly direction which the position of L.C., L.E. and L.F. relative to the Lou-lan station had led me to conjecture as the probable line of the old Chinese route. So I decided to march straight to it. The experience already gained of the trials confronting our camels made me doubly anxious to avoid any needless moves which would cause loss of time and delay in carrying out our main task. Afrdz-gul's report had shown me that the graves of L.Q. were evidently of the same type as those explored at L.F. but less well preserved, and this made me less reluctant to forgo on this occasion a visit to L.Q., which meant saving a day. Its graves were duly explored a year later.'a
As we moved towards the Mesa the ground became more sandy and dead tamarisks increasingly frequent on low Yârdangs. But the camels in spite of the improved going lagged sadly behind. When the convoy had at last come up at L.I., I learned that one camel had completely broken down and had finally been left behind after removal of its load. Hassan Akh tin, who had had charge of the camels on all my three journeys and whose experience made him more than ever the main prop of my transport arrangements on this difficult desert voyage,8 earnestly urged the need of an immediate halt. He and the other men would require the rest of the day and the night to ` re-sole ' the camels, whose feet had developed cracks and sores on the hard salt-encrusted ground. I f we pitched camp at the foot of the Mesa he would also have the chance, which he eagerly pleaded for, of bringing in the abandoned camel which I had already felt obliged to order to be shot. As the event proved, I was not to regret this enforced halt.
A rapid examination of the three Mesas, lying close together near our Camp C. c and jointly marked L.I. on the map (No. 32. A. 3), produced convincing evidence that they had been occupied at the time when Chinese traffic was passing through Lou-lan, and perhaps even before. The Mesas, rising to about sixty feet in height, showed, indeed, no structural remains. But below the top of the northernmost Mesa abundant refuse of reed-straw mixed with cow-dung was discovered on sheltered portions of the slopes. Similar refuse, together with several pieces of simple reed matting, was also found in a small Nullah dividing the little plateau into two distinct portions. Layers of refuse were likewise found on the top of another of these Mesas lying close by to the southeast, while on its north-west slope we picked up many ragged fragments of a sheepskin lined with a coarse woollen fabric of ` open ' weave. Everything pointed to these Mesas having once served as a camping-place for indigenous herdsmen, perhaps only periodically during summer seasons when the elevation of their tops must have made them welcome places of refuge from the plague of insects that infest the ground near the terminal marshes. We, too, in May 1907 had been glad to seek comparative safety from these pests on similar elevated ground near the point where the westernmost end of the Limes flanks the edge of the terminal basin of the Su-to-ho.9
That such periodical occupation must have extended into historical times and, perhaps, until the final abandonment of Lou-lan, was proved by the glass bead, L.I. 013, which was picked up under my eyes at the foot of the northernmost Mesa, by the side of a well-worked stone arrow-head, L.I. 012 (Pl. XXII),evidently Neolithic. Afràz-gul, when first visiting these Mesas on February 16th with some men, had searched the ground carefully and had then found some bronze fragments, including those of a bronze javelin, L.I. 05 (Pl. XXIII), besides several worked stones, among them the jade celt, L.I. 015 (Pl. XXII), and fragments of coarse decorated pottery, L.I. 01-5 (Pl. XXIII,
features in his character that made his help so valuable for serious desert work, cf., e.g., Desert Cathay, i. pp. 112, 373 sq. a Cf. Desert Cathay, ii. pp. 134, 157 sq.