who trace, or pretend to trace, their descent to the early nobility of Islam: Sprenger adduces a curious example in the Moslems of Paniput. These are composed of four castes: the descendants of Abu Ayûb (the citizen whose guest the Prophet was on his first arrival at Medina); the descendants of Othmân; Affghans; and converted Rajpoots. The first two do not intermarry with the two last. They carefully maintain their genealogical trees, in which the pedigree is followed up step by step to the founders of the family in the very age of Mahomet; in later days the births and deaths are entered, and sometimes the marriages also, with the dates. The pedigree of the Qthmânite clan is carefully kept in the custody of the Nawab, the head of the house, but Sprenger does not think it really above a hundred years old. For the last seventeen or eighteen generations, that is, up to the time of Alauddeen Shah, when the family first entered India, the details may be founded more or less on fact. Beyond that, the descent runs through kings of Herat, Sheraz, Kafaristan, Balkh, etc., and is pure fabrication. The same is the experience of Sprenger with all the other pedigrees he has met in India. "Life in the East," he says, "is all too insecure, and under too arbitrary a government, to look for archives extending over several centuries. In the deserts of Arabia such documents are altogether unknown; and it would be childish to imagine that the minute ramifications of any tribe could be retained in the mere memory for a long series of years." ¹ 

It seems probable that registers of lineage, like the Paniput ones, were known at a very early period, and that the practice of keeping them soon became common.² These would be first compiled by their respective families or partisans, for the more distinguished heroes connected with the rise of Islam; and thus it may be concluded that when, say in the second century, the