Masson's efforts were now directed towards getting back to Kabul. His first essay was in company of two brothers of Haji Khan, who vowed to get to Kabul somehow, even if, as Afghans, they had to fight their way through Besud. The party followed up the Topchi valley from Bamian, and crossing by the Shutar Gardan Pass, they reached Karzar. Here again Masson noted extensive ruins _en route_. The road was bad and the difficulties great, "leading over precipices," but they did, nevertheless, succeed in crossing the main divide. Here Masson experienced a very bad time, and to his disgust found that he must retrace his steps to Bamian, owing to counter orders from Haji Khan recalling the escort. There appeared, however, a prospect of getting out of Bamian by the Shibar Pass (an easy pass), leading to the head of the Ghorband valley; and trusting to certain arrangements made by a Paghmani chief, Masson made a fresh attempt, passing eastward the ancient remains of Zohak, and ascending by a fairly easy open track to the valley or plain of Irak. Probably this pass is the one known as Khashka in our maps. The wind was terrific, but the comparative freedom from snow was an unexpected advantage. Passing eastwards from Irak (still on the northern slopes of the Hindu Kush) the party made comparatively easy progress by a valley which Masson calls Bubulak (where he observed tobacco to be growing). They gradually ascended until once again they found themselves in snow, but instead of making direct for the Shibar they inclined to a more northerly pass called Bitchilik, which is separated from the Shibar by a slight kotal (or divide). Here they found the Paghmani chief whom they expected to join, but they found also that the section of Hazaras who held these passes then were determined to bar their passage. Once again Masson had to abandon the attempt (albeit the Shibar route to Kabul would have been a very devious and dangerous one), and returned to Bamian. There are one or two circumstances about this exploration of the western Hindu Kush passes which deserve attention. For once Masson is slightly inaccurate in his geography when he states that the Irak stream drains into the Bamian valley. It joins the Bamian River after it has left the valley and turned northward. So slight an error is only a useful proof of his general accuracy. Another remarkable fact was that he, a Feringhi, was elected by the Afghan gang with which he was temporarily associated as their Khan, or chief! He was a little better dressed than most of them in European chintzes. He found himself utterly unable to restrain their looting propensities, but he made himself quite popular by his civility and his small presents to the wretched Hazaras on whom they were quartered. Incidentally he gives us a most valuable impression of the nature of an important group of Afghan passes, and I doubt if his information has ever been much improved upon.