- 3 -


been found to be the most important inputs or resources for measuring productivity for different agro-socioeconomic situations.

     Error: in the choice of resources with which to measure productivity can lead to the generation of cropping systems unacceptable to the farmers for whom they were being designed simply because productivity of the most important factor from the farmer's point of view may actually be less, even though when measured in the technician's terms, productivity may have improved over the traditional system. In the Colombian case (Andrew, 1969), researchers and extension workers were frustrated because farm yields (or productivity) per hectare were not increasing in accordance with known potential even though farmers were adopting ipany of the recommendations being given. Because potato seed was the most limiting resource for the farmers, they were planting in such a manner that the productivity of the seed was increasing significantly, yet because they were planting less than the recommended amount of seed per hectare, productivity per hectare, or yield, was not so great as expected by the scientists based on their erroneous (in this case) concept of productivity.

     In the above example, only potatoes were considered. In multiple cropping systems, when more than one crop is being produced, another type of problem arises. This has to do with the choice of the "product" in which to measure productivity. If one measures the productivity of a maize-bean system only in terms of maize production then increases in bean productivity are forfeited. On the other hand, it does not make sense to sum the production of maize plus beans, because they are two . . completely differen-t, products (especially, for instance, if the beans are harvested as green beans).

     In order to be useful, the unit for measuring
productivity must satisfy several criteria (1Hildebrand, 1976, p. 349). First, it must be common to all the products, a criterion that energy, dry matter or protein would fulfill. Second, it should be relatively easy to measure, particularly in practical or applied situations.as opposed to laboratory conditions. Third, it must reflect quality differences between the products, and fourth it must provide a means of comparing different cropping systems. But the most important criterion of all is that the manner of measuring