where e is an error term assumed to be in- dependent and normally distributed, and the two X variables are population density and the illiteracy rate. The above model was estimated using multiple regressions. The popular semilog functional form could not be used on this data because the vis- itation rate from many cantons was zero. Two specifications were estimated; with and without illiteracy rates. The results are presented in Table 1. Note that all the coefficients of the mod- el have the expected sign. The coefficient on price is negative and statistically sig- nificant. Higher population densities result in more trips, which is expected since people living in less dense cantons prob- ably have nearby rain forests to visit. The higher the illiteracy rate the lower the vis- itation rate, which indicates that visitation is positively correlated with education, and probably permanent income