Despite the existence of an enormous literature dealing with the Meiji Restoration and its origins, it is still surprisingly difficult to acquire precise information about some aspects of Japanese society in the middle of the nineteenth century. One such difficulty is that of obtaining general quantitative data about the great feudal domains (han) which constituted the major political and economic units of the country. This is not to say that details concerning the domains are impossible to find. Many records are readily available, even in print, and some have been used by scholars to support or illustrate general statements. It is commonly accepted, for example, that agrarian productivity increased greatly in Japan between the seventeenth and nineteenth centuries and that land dues were extremely high, especially at the end of the period. It is possible to cite domains as examples for each of these generalisations. On the other hand, it is never very clear whether the examples themselves are typical or merely random, how far they approximate to or differ from the norm. Nor has there been much attempt to discover whether the wide differences which existed between one domain and another in these matters followed any identifiable pattern. It is with these problems that the present article will deal.