24 ACROSS SWAT AND MR R [Chap. I
of ruins in the vicinity.8 Yet the presence among the old coins, which the Hindu tradesmen of the local Bazar had to show me, of a number of Kusana and later Indo-Scythian copper coins pointed to early occupation.
I was also able to secure two men from Kalam, in the Swat Kôhistân, for the purpose of linguistic and anthropometrical examination. Their speech proved to be identical with Garwi, a language of the Dard group. This they declared to be spoken also by the people of the Dir Kôhistan in the valley of the easternmost branch of the Panjkôra, from Patrak to Tal. Of Dirï', which Dr. (now Sir George) Grierson's survey of the ' Pigaca languages ' mentions as a separate form of speech about Dir,3 I vainly endeavoured to obtain information. And if Garwi prevails also on the Panjkôra headwaters above Dir, it is difficult even to guess where ' Diri ' can be located ; for in the smaller valleys north and west of Dir Pashtu is now alone spoken, though the appearance of the people indicates that their descent is largely from Dard stock. The lively recollection retained at Dir of Kafir raids within the present generation is of interest as affording evidence of the times still recent when both banks of the Kttnar above Asmar formed part of Kafiristan or, as Marco Polo calls it, ' Pashai'.10
That the Kôhistanis of Dir and those further eastward on the Swat River's headwaters are a remnant of the population which held the Panjkôra and Swat Valleys during Buddhist times, and were dispossessed by the invasions of Yûsufzai Pathans, as Colonel Deane first suggested," appears probable. The local traditions, which he and Colonel H. S. Godfrey quote,12 seem to retain a recollection of this origin ; but they evidently do not go back much beyond the conversion to Islam, which is alleged to have taken place here some eight or nine generations ago. If, in the absence of anthropological data, linguistic affinity is taken as a guide, these Kôhistanis of Bashghar or Bashkar, as the valleys at the headwaters of Panjkôra and Swat are collectively known, are certainly to be classed as of Dard stock. And the assumption of the same origin for the inhabitants of Buddhist Udyana would agree well with surviving philological and historical evidence.
8 This negative result is not in contradiction with the statement made in Col. S. H. Godfrey's interesting paper on the Panjkôra Kôhistan, Geogr. Journal, 1912, xl. p. 5o, concerning the existence of ruined houses and forts of early date in the Talash and Dushkhel Valleys of Dir'. The Dushkhel tract adjoins Talash on the south-east and, of course, belongs to Swat, not to Dir, though brought in recent years under the control of the Nawab of Dir.
See Grierson, Pilaw Languages, p. 6.
"o Cf. above, p. r r.
" See Notes on Udyàna, J.R.A.S., 1896, pp. 661 sq.
12 See Colonel Godfrey's paper A summer exploration in the Pan jkora Kohfslan, Geogr. Journal, I912, xl. pp. 5o sqq. I doubt whether much reliance can be placed on the alleged claim of some Dashui Kôhistani communities that their ancestors built the ruined houses and forts noticed in Lower Swat and especially in the Dushkhél tract. It is suggestive of popular etymology'. In any case those ruins must have been deserted and unheeded for long centuries before the assumed emigration from Swat of those Kôhistanis' ancestors.