From that break in the hills which gives passage to the Ghorband from the south-west and the Panjshir from the north-east there strikes off one well-known route across the backbone of the Hindu Kush, which is marked near the southern foot of the mountains by the ancient town of Parwan--a commercial site more ancient than that of Kabul--the headquarters of Sabaktagin, the Ghuri conqueror, when he wrested Kabul from the Hindu kings, and of Timur the Tartar in later ages. Consequently, the pass which bears north from that point is often called the Parwan. It was, according to Lord, the chief khafila route from Badakshan (although it may be doubted whether it was ever as popular as the Khawak when the Panjshir route was not closed by tribal hostility), notwithstanding that far less traffic passed that way than by Bamian and the Unai. The head of the pass was known as Sar Alang, so that it figures in geographical records frequently under this name also, whilst the local name acquired for it in the course of surveying in 1883 was Bajgah. To the west of this is the Kaoshan Pass, which is also known _par excellence_ as the pass of "Hindu Kush"; and farther west again is the Gwalian (or Walian), an alternative to the Kaoshan when the latter is in flood. Lord selected the Parwan or Sar Alang Pass, narrow, rocky, and uneven, with a fall of about 200 feet per mile, and was fairly defeated in his attempt to cross, on October 19, by snow. This is about the closing time of the passes generally, the Parwan being only 12,300 feet in altitude, although Lord estimated it at 15,000. It is worth noting here that the Russo-Afghan Boundary Commission party crossed by the Chahardar Pass (a pass to the west again of the Walian) in the same month of October without encountering any insuperable difficulty from snow, although the Chahardar is more than 1000 feet higher than the Parwan. The fact that Lord met a khafila snow-bound near the top of the pass indicates that it was closed rather unexpectedly. Valuable observations were, however, the result of this reconnaissance. It revealed the fact that snow lies lower and deeper on the northern side of the Hindu Kush than on the southern, a fact which is in direct opposition to the general characteristics of the Himalayas. The explanation is, however, simple. In both cases the snow lies lowest on that side which reaches down to low humid plains and much precipitation of moisture. Where the barrier of the mountains breaks the upward sweep of vapour-bearing currents, there snowfall is arrested, and the highlands become desiccated. Lord's observation as a geologist also determined the constitution of these mountains. He noted the rugged uplift (beautiful from the admixture of pure white felspar and glossy black hornblende) of the central granite peaks through the overlying gneiss, schists, and slate, which thus revealed the extension of one of the great primeval folds of Himalayan conformation.