The Contract Responsibility System 59 commune level, maintained out of the collective fund plus a nominal 2 yuan annual contribution per capita. It is not permissible to generalise on the basis of visits to only four communes/What does emerge, looking at these communes as case-studies, is the wide variations — depending on the local, specific circumstances — in the interpretation and implementation of the contract responsibility system. In two cases, the contract was not with the household at all but with the team, the rationale being that the degree of mechanisation was high in one case, while a plantation crop was involved in the other. In the other two communes the contract is with the household; but a varying mix of household and team functioning is observed because the irrigation systems were designed for collective use, and large-scale machinery is wholly owned at a collective level. Machine-ploughing and irrigating were thus observed to be carried out on the former large scale even when all other operations were at a family-labour scale. One of these two communes however had taken a significant step towards privatisation by permitting newly planted orchard land to be henceforth private property up to a certain ceiling (though land already cultivated remained collective property). Whatever the degree of autonomy extended to the household as a producing unit, to date it remains firmly embedded within the collective structure of ownership of land and irrigation assets, collective planning of overall output deliveries, collective payment of taxes, and collective contribution to and utilisation of the accumulation and welfare funds. To characterise the household contractiifg system as representing 'decollec-tivisation' represents at the present stage a highly inaccurate description of what is happening. At the same time, however, the private owenrship of some (non-land) means of production under the new system, as well as the increasing participation in commodity produciton by households, is bound to promote income differentiation. The Chinese planners are well aware of this, for studies on income distribution in rural areas and computation of indices of inequality at successive points of time are being carried out. One reported study indicates, firstly, that initial inequality of personal incomes was extremely small, with a Gini coefficient of below 2.5; and that there has taken place an increase in income inequality during the period of rapid growth, but the Gini coefficient value remains below 3.0. The extremely egalitarian nature of the commune distribution system before 1979 permits some degree of manoeuvreability today with respect to policies which simultaneously promote growth and inequality. There are however certain dangers in the promotion of commodity production. While wide inter-commune variations in income levels have always existed (owing to the reason that some communes enjoyed higher differential rent from better soils, irrigation, location, etc.), this did not generate inter-personal tensions because communes formed virtually