38

and Bangladesh, but might be absolutely critical in the environment

of mountainous Nepal.

 Moreover, it is likely that a centralized extension system may

be dominated by urban and bureaucratic political interests, which may

not coincide with farmers' interests. If promotions are determined

at the center, it is almost inevitable that better, more ambitious

agents will gravitate to the center in their orientation, work, and

eventual domicile, and will not work directly with farmers.

 One factor which influences the suitability of a centralized

extension system involves the characteristics of the crops or tech-

nology in question. Centralized systems may be well suited to a crop

for which research is reasonably advanced, which is highly sensitive

to field management, and whose quality is very important for marketing

requirements. Tea and tobacco are major examples, and in different

parts of the world highly centralized, intense, expensive extension

systems are oriented towards such single crops. (In Kenya, tea growers

have one extension agent for 120 farmers, and extension services cost

$18 per farmer. ) Centralization may also be important in commodities

where timeliness in marketing and processing are critical, such as

palm oil and fresh milk.

 A good example of the potential value and problems inherent in

centralized, targeted extension programs can be seen in the experience

of India. From about 1969 through the early 1970's, efforts were made

to target integrated services to small farmers through the Small


 Uma Lele, p. 64-69.