768 EXPLORATION IK CHINESE TURKESTAN. 



be arrived at as to the rate at which tlie woriv of these forces proceeds 

 in various parts of the Taklaniakan. And even then there will ))e lit- 

 tle to guide us as to the corresponding conditions prevailing during 

 earlier historical periods. 



While most of the buildings of this important site had suffered from 

 erosion, there were others where parts at least were still buried under 

 deep sand (pi. v). From some of these m}" excavations brought to light 

 many very interesting ol)je<'ts illustrating the industrial arts of the 

 period. The articles of ornamental wood carving, which include elab- 

 oratel}" worked chairs, small architraves and other achitectural pieces, 

 etc., show decorative motives familiar to us from the relievo sculp- 

 tures of the ruined Buddhist monasteries on the northwest frontier of 

 India, the ancient Gandhara. The date thus indicated fully agrees 

 with the chronological evidence of the Kharoshthi writing on the 

 wooden tablets, apparently memoranda and lists, found scattered in 

 various rooms of the same dwellings. Broken pieces of arms, house- 

 iiold implements, a musical instrvuiient and similar o})jects of domestic 

 use, all of wood, help vividly to liring before our eyes the conditions 

 of everyday life of this distant region in the first centuries of our era. 



It was difficult for me to realize fully that so many centuries had 

 passed since these dwellings were deserted while I traced the plan and 

 arrangement of the orchards and gardens once surrounding them. 

 Rows of fallen poplars, some 50 feet in liMigth, half covered ])y the 

 sand, showed the position of avenues, such as are planted to this day 

 everywhere along the roads and canals of Turkestan oases. The rush 

 fences used then, as now, for the iiiclosures of gardens could bo seen 

 sticking out from the sand. A little digging along them often revealed 

 small heaps of dry leaves that must have accumulated there while the 

 trees, now reduced to blanched and withered trunks, were still thriving. 

 Among these my diggers had no difficulty in distinguishing various 

 fruit trees, such as the peach, plum, apricot, mulberry, etc., Avith the 

 wood of which they are familiar fi'om their own homes. 



The character and conditions of the articles found within the houses 

 plainl}^ showed that they had been cleared by their last inhabitants, 

 or soon after their departure, of everything that possessed value. 

 Luckily, there were left behind the ru])l)ish heaps to reward me with 

 finds of the greatest antiquarian interest. The richest mine of this 

 sort was struck in a small and much-decayed building, one room of 

 which proved to contain a consolidated mass of refuse, lying fully 

 4 feet above the original floor. Among the layers of broken pottery, 

 rags of felt and of woven fabrics, pieces of leather and other rubl)ish, 

 I discovered thereover two hundred documents on wood, of all shapes 

 and sizes. Besides tablets with the Indian Kharoshthi writing, which 

 form the great majority, there came to light numerous narrow pieces 

 of wood ])earing Chinese characters, and two dozen Kharoshthi docu- 



