A double exponential function provides a good description of this diffusion process. This function was derived from the data concerning the diffusion process of the two improved varieties ICT A-Ostua and ICT ASan Martin V.B. in Guatemala in the years 1990-1992. The two relevant parameters may be interpreted in the following way: Parameter a1 reflects the marginal rate of substitution between the yields obtained from both, the improved and the farmer' s bean varieties. Parameter a2 considers the relationship of the acquisition costs among the two varieties, which are based on the seed price plus the transaction's costs. This concept allows to discern the difficulties faced by the farmers, such as the lack of information of where and when to acquire the improved seed, the delivery costs and so on. Unfortunately, this function no~ only tends to over-estimate the number of farmers who buy the improved seed during the year of maximum demand but al so the time span of this maximum demando A third degree polynome would be more appropiated, but its inter-dependent parameters are more difficult to interptet, in contrast to the easy interpretation of the double exponential function.