On the 29th of August I left Lachoong and proceeded up the valley. The road ran along a terrace, covered with long grass, and bounded by lofty banks of unstratified gravel and sand, and passed through beautiful groves of green pines, rich in plants. No oak nor chesnut ascends above 9000 feet here or elsewhere in the interior of Sikkim, where they are replaced by a species of hazel (_Corylus_); in the North Himalaya, on the other hand, an oak (_Quercus semecarpifolia,_ see vol. i., chapter viii) is amongst the most alpine trees, and the nut is a different species, more resembling the European. On the outer Sikkim ranges oaks (_Q. annulata?_) ascend to 10,000 feet, and there is no hazel. Above the fork, the valley contracts extremely, and its bed is covered with moraines and landslips, which often bury the larches and pines. Marshes occur here and there, full of the sweet-scented Hierochloe grass, the Scotch _Thalictrum alpinum,_ and an _Eriocaulon,_ which ascends to 10,000 feet. The old moraines were very difficult to cross, and on one I found a barricade, which had been erected to deceive me regarding the frontier, had I chosen this route instead of the Lachen one, in May. Broad flats clothed with rhododendron, alternate with others covered with mud, boulders, and gravel, which had flowed down from the gorges on the west, and which still contained trees, inclined in all directions, and buried up to their branches; some of these debacles were 400 yards across, and sloped at an angle of 2 degrees to 3 degrees, bearing on their surfaces blocks fifteen yards in diameter.* [None were to be compared in size and extent with that at Bex, at the mouth of the Rhone valley.] They seem to subside materially, as I perceived they had left marks many feet higher on the tree-trunks. Such debacles must often bury standing forests in a very favourable material, climate, and position for becoming fossilized.