390 THE LOU-LAN SITE [Chap. XI
outside surface of the third base was visible behind the brickwork of the second. Whether the dome had contained a central shaft or chamber could no longer be ascertained, owing to extensive havoc wrought in this portion of the Stûpa. The sun-dried bricks used throughout were fairly hard, and showed the same dimensions as observed in L.A. u, viz. i8" x 12" X 4". Their slightly reddish appearance in places can be explàined by a conflagration which may at one time have destroyed structures at the foot of the Stûpa, and by bonfires lit on its top after the shrine had fallen into ruin. The orientation of the sides of the base agreed with that of the walled enclosure and of the ruined dwellings within it. The terms ` south face ', ` north-east corner ', etc., applied above, are therefore not quite exact, and have been used merely for -sake of convenient brevity.
Changes An observation made on the ground adjoining the- Stûpa may find mention here, as it serves to
by surface. illustrate the vicissitudes which the surface of the soil must have undergone since the occupation of
the site ceased. On the north slope of the erosion terrace which the Stûpa occupies I found a thick layer of dead tamarisk brushwood which once had grown up here covering a piece of worked timber undoubtedly fallen from the ruin above. The position occupied by this dead tamarisk growth was six feet below the original level of the ground as marked by the already mentioned foundation of tamarisk fascines beneath the Stûpa base. It is obvious that wind-erosion must, after the abandonment of the site, have lowered the ground level to this depth before a temporary return of moisture permitted tamarisk scrub and the like to grow up again.
Varying In full agreement herewith I found that wherever the tops of Yârdang ridges near the Stûpa
factors affect bore dead tamarisks with the roots still embedded in the soil, their surface lay five to six feet below surface
levels. the original ground level as marked by the Stûpa foundations.5 I noticed the same difference of
level also elsewhere, e.g. south of the enclosing wall of the station, where the )(di-clang tops bearing dead tamarisks showed a level six feet lower than that on which the wall had been built. It clearly follows that the process of denudation and erosion, since the site was abandoned some time in the .fourth century A.D., has not been a continuous one, and further that no safe conclusions can be drawn from the results of any measurement of levels over wind-eroded ground, such as that crossed by Dr. Hedin's line of levels south of the L.A. site, as far as the configuration of the ground during earlier periods is concerned. As the process of erosion is neither constant in its progress nor necessarily uniform over the whole area, only structural remains of an archaeologically datable character can furnish reliable indications as to the levels which the surface of the soil may have occupied at particular historical periods. It has appeared desirable to emphasize this limitation of our knowledge of earlier surface levels in view of a theory which, merely on the basis of a single line of measured levels, has assumed the ruined station L.A. to have stood on the actual shore of ` the ancient Lop-nor '.e
Ruined • The nearest and most conspicuous of the structural remains traced outside the walled station
Stiipa L.A. was the large ruined mound built of sun-dried bricks, L.A. xi (Fig. 98), and situated about 40o yards XI.
east-south-east of the Stûpa. It occupies the top of a wind-eroded terrace ; south of it the soil has been scooped out to a depth, of some fifteen feet, as seen in Fig. 98. Close examination showed that the mound represents the remains of a Stûpa of which the cylindrical or dome portion has been completely destroyed, human agency probably aiding the erosive force of the winds. The base appears to have been orientated like that of the Stûpa within the station, but owing to the far-advanced decay of the outer masonry the sides could not be established with full precision. Judging from the north face, which has suffered less than the rest, the base may have formed a square of about 43-44 feet on the ground. No arrangement in stories could be made out. But
b Such tamarisk growth on a Yardang top is seen in Fig. 6 Cf. Hedin, Central Asia, ii. p. 635 and passim.
98 on the left of the foreground.