The story of the commerce of Central Asia, which centred itself in Herat in the days of Arab supremacy, has a strong claim on the student of Eastern geography, for it is only through the itineraries of these wandering Semetic merchants and travellers that we can arrive at any estimation of the peculiar phase of civilization which existed in Asia in the mediæval centuries of our era; a period at which there is good reason to suppose that civilization was as much advanced in the East as in the West. It is not the professional explorers, nor yet the missionaries (great as are their services to geography), who have opened up to us a knowledge of the world's highways and byways sufficient to lead to general map illustration of its ancient continents, so much as the everlasting pushing out of trade investigations in order to obtain the mastery of the road to wealth. India and its glittering fame has much to answer for, but India (that is to say, the India we know, the peninsula of India) was so much more get-at-able by sea than by land even in the early days of navigation, that we do not learn so much about the passes through the mountains into India as the way of the ships at sea, and the coast ports which they visited. According to certain Arab writers large companies of Arabs settled in the borderland and coasts of India from the very earliest days. Indeed, there are evidences of their existence in Makran long before the days of Alexander; but there is very little evidence of any overland approach to India across the Indus. Hindustan, to the mediæval Arab, commenced at the Hindu Kush, and Kabul and Ghazni were "Indian" frontier towns; and the invasions and conquests of India dating back to Assyrian times include no more than the Indus basin, and were not concerned with anything farther south. The Indus, with its flanking line of waterless desert, was ever a most effectual geographical barrier.