32 FOOTFALLS OF INDIAN HISTORY
From these examples and from what we can see to have been their inevitableness, we might expect
that any important city of the Buddhistic period would be likely to occur in connection with a
monastic centre some few miles distant. Now it is possible to determine the positions of a great
many such cities on grounds entirely a priori. It is clear, for instance, that whatever geographical considerations might make Benares great would also act at the same time to distinguish Allahabad. By a similar induction, Mathura on the Jumna and Hardwar on the Ganges might also be expected to furnish proof of ancient greatness. Now outside Prayag we have to the present time, as a
haunt of sadhus, the spot known as Nirvanikal. And in the vicinity of Hardwar, is there not Hrishikesh? The caves of Ellora have near them the town of Roza. But this we must regard as a sort of Mohammedan priory, "in as much as its population consists mainly of religious beggars (of course
not celibate) living about the tomb of Aurangzebe. The neighbouring capital that supported the youth
of Ellora was probably at Deogiri, now called Daulatabad.
It is the broken links in the chain, however, that fascinate us most in the light of this historical generalisation. What was the city, and what the state, that made Ajanta possible? What was the city that corresponded to the dharmsala at Sanchi? What was the city, and what the abbey, in the case of Amravati?
