66 SOCIAL SCIENTIST the major constraints lie elsewhere, can possibly be better understood in a more intensive study which focuses on changes taking place among specific tribal groups, or between different sections within the same tribe. Nair is right when he says that large-scale census surveys often fail to understand the specificity of th^ ^tribal situation. For instance, reciprocal exchange of labour between kin households (an important source of labour in Jhum cultivation) is often either ignored or taken as synonymous with wage labour in many surveys. However, his contention that 'hired labour' reported is Meghalaya is only a modified form of traditional exchange labour and 'is not wage labour per se' (p. 101) needs careful consideration, particularly since he takes this to be a major indication of an undifferentiated tribal peasantry. Though Nair does make a distinction between 'family', 'hired/ and 'exchange labour,' he does not show how he actually differentiates between the latter two kinds of labour. If kinship reciprocities define that which is called exchange labour, in what way has labour hired for wages been characterised by him as a modified form of exchange labour? This is important, as Nair's survev data reveals a significant percentage of hired labour in the labour use pattern not merely in settled holdings (34 per cent) but in Jhum holdings as well (22 per cent). In comparison, exchange labour appears less important even in households engaged in Jhum cultivation (Table 8.5). We are not given the proportion of household income that comes from wage labour. It is hence difficult to see whether some tribal households are more dependent on this source of income than others, particularly in areas where settled cultivation is predominant. Among the sample households there are some which are classified as 'non-cultivating' households. The author docs not elaborate as to where such households derive their income form. \ The data also indicate that settled cultivators derive around twice as much income as Jhum cultivators (p. 121). While all households are shown to market some .proportion of output, households engaged in settled cultivation in each land size category market a higher percentage of what they produce. Further, Table 5.6 shows that white paddy and maize are grown mainly for home consumption, 'non-food' items like fruits, jute, etc., are produced entirely for the market. It is thus more than likely that some tribal households have been able to respond to and take advantage of the growing demand for products such as fruits, arecanut and even potatoes which, as a result, are g^wn primarily for the market. Those who have be^n able to bring additional lands under permanent cultivation have possibly bencfitted inost in this respect. Commercialisation and markets in this manner have probably played a rnore important role in the transition from shifting to settled cultivation than suggested by the study. Nair observes that a major constraint in this process of change is the shortage of labour. Hiring of Wiage labour and thereby bririgjing an