iv. 89. a, we read of a ` lord ' or ` superior of the four castles '. But there is nothing to show that the ` three castles of Nob ' were included among them, and the mention made in M. 1. xiv. 109 of Little Nob side by side with the ` four castles' seems opposed to it.
As the affairs of which the abstracts of the documents mostly tell us seem of a petty nature, it can safely be assumed that the majority of the places they name must be looked for within, or close to, the Lop region. But owing to the strange Tibetan garb in which these local names are presented, and the uncertainty as to the phonetic value intended by apparent transcriptions of non-Tibetan names, conjectural attempts at identification must be hazardous at present. Only in one case is the similarity of forms sufficiently close to justify here a tentative suggestion, viz. in that of the name Cer-cen, M.1. xxviii. 2, which, if really intended for a locality, could obviously be taken to apply to the present Charchan.8 The impossibility of identifying any other of the abundant local names is particularly to be regretted in those unfortunately rare instances where the documents, even in the mere excerpts now available, seem to present us with glimpses of true colour in the local background. Thus in M. i. xxviii. 005 the writer reports ` that he had travelled to Byeu-ling', apparently with a caravan, and ` that they were belated and in extreme want of water '. There is no clue to the identification of Byeu-ling, which is mentioned elsewhere, too, e.g. in M. i. xxvii. 19 ; xxxi. 001. But the situation curiously suggests an experience such as nowadays might befall a caravan which attempted the route through the desert to Tun-huang too late in the season, when the melting of the ice at the salt springs south of the dried-up Lop sea would cause serious trouble about the supply of drinkable water. In another letter, M. i. i. 41, which mentions Shodbrog (` lower Brog '), we find a somewhat similar report from a party evidently in great straits through want of the necessaries of life.
The impossibility of identifying the localities referred to makes itself particularly felt in the case of those records which, like, e. g., M. i. iv. 4, 71 ; vii. 00! ; x. 3 ; xxi. 9; xxiii. 009, treat of military affairs. It is regrettable, too, that we cannot define the frontiers referred to in documents like M. i. xiv. 7, or the posts occupied by the so pa or frontier-guards.9 But even a cursory examination of the abstracts furnished of all these Tibetan records makes it clear that most of them refer to the multifarious petty affairs of the small Tibetan garrisons in the occupied territory and the civil administration of its little oases.
The very limitations which the petty nature of these records implies as regards the functions and resources of the Tibetan officials mentioned in them must warn us not to attach undue importance to the high-sounding titles under which so many of them figure. In all probability these ` ministers ', ` lords ', etc., held positions of small power and extent. At the same time, if the affairs of which the Miran documents give us glimpses were by themselves of very modest consequence, considerable historical interest attaches to the political conditions revealed by the records as a whole. They clearly prove that in the Lop tract the Tibetans maintained not merely military posts, but also an effective civil administration, during the period to which the occupation of the Mirân fort and probably also its construction belong. This was possible only in the period of general Tibetan predominance in the Tarim Basin, of which we know, scanty as our historical sources are, that it extended from the decay and ultimate downfall of Chinese power in the last third of the eighth century A.D. to about the middle of the ninth century.10
Cf. above, p. 323.
° See Dr. Francke's remarks in Appendix G.
10 Cf. Ancient Khotan, i. p. 65 ; M. Grenard's paper in Journal Asiat., 1900, Jan.—Févr. pp. 24 sqq.; below, chap. xXI. Sec. ii.