FEUDAL FORMATIONS 125 India at least, where more intensive work needs to be done. Otherwise Karashima's data, which is in any case confined to a small part of South India, may become the source of yet another inclusive debate. We feel tha.t ultimately the value of theory in the analytical reconstruction of historic structures rests on a dialectical relationship between it and the empirical basis and if any of the two becomes a restrictive or binding force, the essence of historical analysis suffers. The book under review effectively sets the stage for the challenge to further fortify the theory that early medieveal Indian society was 'feudal9. ALOKA PARASHER Dept. of History, University of Hyderabad.