we say that there is no hope that we can increase our rainfall really appreciably or effectively by any amount of tree-planting. A whole ocean of water can not give rainfall if the general pressures and temperatures and winds are hostile to precipitation.
As was pointed out at the beginning of this paper, forests are of many different kinds. We can not, therefore, reasonably expect all forests to have the same effects. There may be a difference between tropical and temperate forests, as has already been suggested in the case of Java rainfall, for tropical weather types and rainfall conditions are different from our own, just as tropical forests are different from our own. Tropical rainfalls, as over the great forested Amazon valley, are largely thunderstorm rains, and as forests tend to check air movement, and calms are favorable conditions for convectional overturning, it appears as if tropical forests might be expected to influence rainfall more than our own. Furthermore, from the hot and damp tropical forest, and from the leaves of the closely-crowded tropical trees, there must come a large amount of moisture which will increase the vapor content of the ascending air and tend to increase condensation and rainfall. Thus Woeikof, whose emphasis on the case of Java has been referred to, believes that in low latitudes the vast tropical forests do increase the amount of rainfall. Von Hann, the leading authority on climate, holds that we may conclude "with considerable certainty that, at least in the tropics, the forest may increase the amount of rainfall." Hettner, also, in his work in the tropical Cordillera, came to the conclusion that the forests in the Cordillera of Bogotá favor the growth of clouds and the production of rain. While this is an interesting phase of our discussion, we have as yet no thorough study of tropical conditions by means of the parallel station method. There is also another point. In low latitudes, where the dense tropical forests are found, the rainfall is already so heavy that it is of little or no significance whether there is a good deal more, or a good deal less. In exactly those regions, therefore, where, if anywhere, forests may have a really appreciable influence on rainfall, little or no economic importance attaches to the question. Woeikof believes that rain often begins earlier over tropical forests, and in Mauritius, Walter has called attention to the fact that the number of rainy days seems to be greater over forested areas.
It need hardly be pointed out that, if rain is already falling, the opportunity for it to reach the earth's surface must be better if it falls through the somewhat cooler and damper air over a forest or a grass-covered surface than through a hotter and drier stratum of air over a desert. In the latter case the loss by evaporation may be so great that the drops do not reach the surface at all. Obviously, the contrasts between these two conditions are greatest in the case of the tropical forests and tropical deserts. It must, however, be observed that this
