rainy season (December to March) should have more rain than the south, or lee, side. Yet the fact is that there is about the same rainfall on both coasts at that time. During the southeast monsoon the south (windward) side has a much heavier rainfall than the north (leeward), which is normal. On Celebes, where, according to Woeikof, no such deforestation has taken place, the windward and leeward sides have their normal values of rainfall, the former having a notably larger amount. The case is obviously a very striking one. In reply to a letter from the writer, asking whether newer data from Java tended to strengthen or weaken his previous opinion regarding this case, Dr. Woeikof said:
I have not modified my views on forests and rainfall. . . . It seems to me that in later years at Tjilatjap, on the south coast of Java, which I cited as a station surrounded by forests, the rainfall is smaller than before. This would confirm my views, as in this formerly very little settled part of the island, forests are rapidly disappearing.
The Java case remains, then, on the authority of one of the best-known meteorologists, a striking example of forest influence on rainfall. So striking, indeed, is it that one is tempted to ask what other possible controlling factors are here active in producing this surprising result.
Recent European Studies
The careful observations which have lately been made in Europe by several investigators (Schubert, Hamberg, Schreiber and others) in western Prussia, Posen, Sweden, Saxony, France and elsewhere, have clearly shown that rain-gauges at forest stations, and above the forest crowns, do generally catch somewhat more rainfall than do the gauges at the parallel stations in open country at the same elevations. The excess varies roughly, we may say, between 1 per cent, and at the most 10 per cent, of the annual mean. But leading European authorities are pretty well agreed that when definite allowance is made for the effects resulting from differences of exposure, due to the better protection of the forest gauges, the apparent excess within the forest is reduced, by the probability of error, to a very narrow margin indeed. In some cases the margin disappears entirely. Schubert, for example, found a summer excess in forested areas of about 6 per cent. Of these 6 per cent., 3 per cent, he believes to be attributable to the better protection of the forest gauge, leaving 3 per cent. And 2 per cent, of these remaining 3 per cent, he thinks still liable to an error. This leaves but 1 per cent.
Conclusion Regarding Rainfall
It appears, therefore, that we have as yet no satisfactory or conclusive evidence that forests, at least in our own latitudes, have a significant effect upon the amount of rain fall, as distinguished from the amount of the rain catch in the gauge. Nor is there direct and unas-
