44 SOCIAL SCIENTIST differences in the size of land held by different caste groups. The major portion of land (62.55 percent) is owned by non-scheduled castes like Brahman, Sadgope, Gandhabanik, Napit and Kolu. Among the non-owners, 90.54 percent arc from scheduled castes and tribes, composed of the ethnic groups like Bagdi, Duley, Muchi and Santal. Significantly, the Duleys are totally landless. A detailed study of the size of land held by various castes shows that about 89.08 percent of cultivable land is under the control of the upper castes, and Sadgopes alone possess 86.64 percent. Among the scheduled^castes, the Bagdis are in possession of 79.82 acres out of 93.86 acres of government distributed land (94.19 percent). Regarding the size of the holdings, most of them hold one acre or less. More than 80 percent of owners belonging to the non-scheduled castes possess from one acre to 10 acres. It is clear that in Sunia there is an internal differentiation in the distribution of land correlated with caste differences. The most significant aspect is the concentration of land in the hands of non-scheduled castes. During the field study it was found that the agrarian categories correspond to caste categories. This was evident both in ownership right and role performance. Table I shows the relationship between caste and agrarian categories in Sunia. It may be noted that the relationship between caste hierarchy and agrarian category is simple in Sunia. The dominant agrarian categories are the non-cultivating landowners and the owner-supervisors belonging to upper castes. At the bottom there is a mass of scheduled castes and tribes who perform most of the arduous agricultural labour. In between them is a broad range of working farmers and tenants of several castes and tribes. These are the five basic categories which in most cases show a mixed character. This mixed character, except on a few occasions, does not allow castes to form well-defined classes. There has been tension between the traditional way of life and new ideas and needs in the village. This highlights the need for legislative and administrative reforms concerning land. An assessment of the present land policy indicates that it has failed to tackle the issue of land distribution and the prevailing practice of subletting and renting. Occasional legal enactments have failed to produce any kind of reaction on the established social structure in rural areas. Whereas rural people tried to develop a new or modified form of production relations in their own native universe, in most cases, this went against the enactments of the governmental policy.