RURAL NON-AGRICULTURAL ACTIVITIES IN INDIA 15 Two more issues relating to definition need some explanation: (a) the definition of rural and urban areas, and (b) the sectoral coverage of non-agricultural activities. These issues are of practical importance because these terms (rural-urban and sectoral coverage, etc.) are used very frequently but are interpreted differently in various censuses and labour force surveys. A discussion of these issues is in order to identify the limitations of the existing terminology and to explain the interpretation of various terms adopted in this paper. Rural-Urban Boundaries Censuses classify individual settlements as urban on the basis of the size of population or local administrative boundaries. In this paper we have adopted the definition of rural areas as a residual category referring to non-urban areas^ The distinction between rural*and urban areas is important from another point of view. Over time various places are classified as urban areas and»some former towns are declassified also. The boundaries of cities also change to encompass surrounding rural areas. Such increases in Mrten* areas, are likely to lead to an apparent decrease in (he magnitude otrufal non-agricultural activities. It is difficult to assess the extent to which such reclassification affects the quantum of rural non^agrtcultural activities. That the impact can be significant is borne out by the fact that during 1971-81 the contribution of net reclassification of localities and boundary changes to urbanisation wa$ as high as 29 per cent (Visaria and Kothari, 1985: Table 19, 65). Sectoral Composition Rural households in developing countries are seldom so specialized that the work of all household members throughout the year falls in a single economic sector. One of the most critical limitations of aggregate data is their inability to cover (capture) the range of occupations and types of employment status of individuals, let alone of households as economic units. Employment and labour force surveys point to various types of 'secondary* or supplementary occupations. But they normally 'fail to capture degrees of emphasis in labour allocation patterns through annual cycles from year to year'. Even more difficult is an assessment of the relative contributions to total household income of different activities that may be undertaken with varying intensity or for longer or shorter periods of time (ILO, 1983:9). However, estimates of labour force by economic activity generally provide a classification of workers according to their principal sector of employment or occupation. Such estimates are likely to underestimate the extent of non-agricultural work because it is commonly a secondary source of income on a part-time or seasonal basis in rural areas. Such underestimation is likely to be more in regard to the activities of female workers than for those of male workers.