SECTION IV.—THROUGH TALASH AND DIR
I have already in the opening section of this chapter indicated the reasons which on my rapid marches across the fascinating ground of Udyana, during the last days of April, 1go6, precluded any systematic survey of the plentiful ruins I passed en route. As I rode, on April 28, by the broad military road from Chakdara towards the Panjkbra there was no time to revisit the remains of Stûpas and monasteries in the Adinzai plain, nor to examine the ruins of ancient towers and habitations which I knew to dot numerous low spurs projecting into the open fertile valley of Wuch. As the road turning westwards approached the easy Katgala Pass dividing the Swat and Panjkbra drainage, I caught a good view of the picturesque ruins of ancient fortified dwellings rising above the scrub-covered slopes on the south, and reluctantly had to pass them. In the burning afternoon sun they looked, indeed, what their local Pashtu name Sure-ma/pi, derived from the colour of the sandstone material, calls them, ` the red houses '. That these, like the similar ruins seen at so many points on high ground above the Swat River, belonged to the Buddhist period is certain. Only a close survey, however, such as I was able to effect in January, 1912, at similar sites near Palai, south of the Swat Valley range, could furnish the definite evidence.'
But after we had entered the broad valley of Wash (Fig. I I) with its wide vista across the Panjkbra to the snow-covered ranges above Bajaur, and reached the Levy post of Kuz-Sarai, I could not forgo my intention of using what little remained of the day for my first piece of archaeological survey work on this journey. At the hamlet of Gumbat, some two miles to the west-south-west of KuzSarai, I had found the comparatively well-preserved ruins of an ancient Hindu temple, first mentioned by Colonel Deane,2 and closely resembling in plan and style shrines I had seen in the Salt Range. But there had been no time then to effect a proper survey, and now, too, the work had to be done in a hurry,
Soon after turning off south from the main road I found myself in the large and shady grove of Jalal Baba Bukhäri's Ziarat, the much frequented shrine of an orthodox Muhammadan saint whose worship is, as so often elsewhere, manifestly but a survival from the days when the ruined Hindu shrine attracted its pious pilgrims. As I rode up the terraced slopes along the lively little stream which spreads fertility over its alluvial fan, I came unexpectedly, about half a mile south-west of the shrine and on the left bank of the stream, upon a massive wall of Gandhara masonry about fifteen feet high. It had evidently been built to support a terrace of cultivation behind it. On ascending the steep path to the hamlet of Gumbat we passed more walls of similar construction. Some seemed to have belonged to ancient dwellings, but the majority, no doubt, had been intended for terraces. The present Pathan settlers, quite incapable of such solid structures, had been content to profit by them. The remains of ancient dwellings they had long ago quarried away to secure materials for building their huts and enclosures.
From the narrow gullies of the hamlet, where the ponies had to be left behind, I ascended to the ruined temple (see Fig. 3), from which the place derives its name of Gumbat or ' dome'. It occupies a small bit of level ground, just above the last huts, which has been secured in part by cuttings from the rocky slopes to the south and west. The situation recalled to me closely that
' See my Archaeological Survey Report, Frontier Circle, 1911-12, pp. 3 sqq. There is in Raverty, Notes on Afghiinistan, p. 202, an interesting reference, taken from a Ytisufzai chronicle, to ruins of a town in Tâlâsh, once