if land tenure relations and/or credit and marketing patterns are

highly exploitative, as was apparent in the KADU project of Ethiopia.
 Even if modern inputs are available and farmers are convinced

(or assisted) by extension agents to adopt them, adoption may not

make economic sense. -A careful analysis of India's Intensive Agri-

culture Districts Program (IADP) in the early 1960's shows that it

was effective in raising farmers' production because they used new

inputs. However, the actual factor productivity (i.e., efficiency

in using resources) of the farmers was unchanged. To increase factor

productivity, the authors of this study believe investments in re-

search systems normally have a higher payoff than investments in

extension systems.2

 In gauging whether or not the new technology is superior,

attention must be paid to the question of. risk avoidance. Highly

productive agricultural technologies often are riskier. They need

the right amount and timing of rainfall./irrigation, and are more

vulnerable to insects and diseases. Because of the high cost of

inputs, their profitability is more dependent on market prices and

access, which may fluctuate. For subsistence peasants on the edge

of survival, risks of these types are extremely serious. A large

drop in production may force sale of all assets, acceptance of virtual

slavery to obtain food, dissolution of the family, and eventual

starvation. Thus the marginal peasant is quite rational when he

rejects risky technology. The definition of technical superiority

should include this question of risk avoidance.
 John Cohen, "Effects of Green Revolution Strategies on Tenants and
 Small-scale Landowners in the Chilalo Region of Ethiopia," The Journal of
 Developing Areas 9 (April 1975), p. 335-358.

 Rakesh Mohan and Robert Evenson, "The Intensive Agricultural
 Districts Programme in India: A New Evaluation," Journal of Development
 Studies 11:3 (April 1975), pp. 148-150.