Sec. in THE LEGENDARY TRADITIONS OF KHOTAN 165
what he subsequently gathered at Khotan. Still there remains room for suspicion whether the Taxila story, so far as it localized the banished people at Khotan, was not itself directly or indirectly a reflex of the Khotan tradition. There must have been communication between Buddhist Khotan and the Taxila region, whether through Kashmir or Gandhâra, and we know well how far legends and traditions will travel.
A settlement of immigrants from the extreme north-west of India, such as Khotan tradition assumes, would necessarily have left its mark in the racial composition of the population. But two circumstances render it difficult for us to distinguish this mark in present anthropological facts. On the one hand, we have no certain knowledge of the racial characteristics of the population which inhabited the region of Taxila or the neighbouring territories during that early period. On the other hand, the points of anthropological affinities between the Galchas and the present Indian populations nearest to Khotan, the Dards and Kashmiris, are so numerous that, if we assume the supposed immigrants to have resembled the latter, their admixture could scarcely have modified in any very striking way the type of the Homo Alpinus which our inquiry in the preceding chapter has shown to enter so largely into the racial composition of the present Khotanese people. Nevertheless, I may note here that I was frequently struck by a certain curious resemblance in general appearance of features between the Khotanese and the Kashmiris, a resemblance difficult to define yet all the more noteworthy on account of the unmistakable peculiarity of type presented by the Kashmiris 28.
We have seen that the Khotan tradition, as recorded by Hsüan-tsang and the Annals of Li-yul ', is equally explicit about an early immigration from the side of China, which was believed to have given to Khotan its first ruler and half of its original population. We have no means of applying to this part of the tradition the test of historical evidence. But I have had occasion in the preceding chapter to show that certain anthropometrical and philological observations distinctly point to an early ethnic event underlying also this part of the tradition 29. In view of the explanations there given, a mere reference will suffice here to that early infusion of Tibetan or quasi-Mongolian blood which seems to have left so material an impress on the racial character of the Khotan population. We have further seen that the presence of such an element was ' sufficiently marked to make Chinese observers of the fourth or fifth century of our era ascribe to the people of Khotan a certain resemblance to their own race 30. The fact that this element is more strongly represented in the population of the eastern part of Khotan territory (Keriya), is also significant ; for it seems to confirm what the Tibetan version of the tradition relates about the territorial division between the earliest settlements of the two colonies.
The linguistic indications furnished by the non-Indian terms occurring in the Kharosthi documents, and by the ` unknown ' language of certain Brahmi MSS. probably ' containing Khotanese translations of Sanskrit texts, have also been noted above S1. It only remains to call attention to the archaeological evidence which certain artistic remains of ancient Khotan supply. In some terra-cotta figurines from the Yôtkan site, in the painted tablets and frescoes of the ruined shrines of Dandân U iliq, and in some of the relief sculptures of the Rawak Vihâra, we meet again and again with representations of quasi-Mongolian faces strangely contrasting with other features in the treatment of the human figure which are unmistakably derived from Indian models as presented by the style of Gandhâra. These peculiar traits will be duly noticed in the detailed descriptions of the relics of old Khotan art as recovered in the
29 Compare Mr. Joyce's remarks on this resemblance, J. Anthrop. Inst., xxxiii. p. 317 sq. 29 See above, pp. 148 sqq.
30 Compare above, p. 149, for the record of the Northern Wei Annals.
31 See above, pp. 149 sq.