Computation of the relative share of each crop in the minima.* solution is shown in Table 6.7. The difference in yield between wet and dry years is calculated for both crops and compared; for each crop the observed difference is transferred to the other: this transferred difference, divided by the comparative difference, gives indices (1.4 and 0.4) which represent the share of each crop and from which percentages may bc calculated. The resulting proportions of 77:23 could be interpreted in the long run as planting all the village land in maize for seventy-seven years and all in hill-rice for the remaining twenty-three years of each century! In practice, the short. run solution, planting three-quarters of the area to maize and a quarter to hill-rice each year, would clearly be adopted. Gould finds that the proportions derived by this method are in rough accord with the actual land use patterns in the Jantilla area, suggesting that the solution has been reached there by the rugged path of trial and error. Since error in this case must be translated in human terms into starvation, the potential importance of game-theory approaches to substitution problems is clear. Like the determination of location rents in the von ThUnen model, game theory problems of the kind