Page:Popular Science Monthly Volume 82.djvu/326

322 process of condensation has been started, it may continue automatically, the liberation of latent heat tending to produce convectional currents.

This perhaps fairly expresses the general view of the average person at this point. However, having seen that the influence of forests upon temperature and upon humidity is so slight, even among the trees, it is unreasonable to expect that the influence upon rainfall over the forest, and especially away from the forest, will be considerable. In the great ascending, damp air masses of a general storm; in the flow of the winds across a mountain barrier; in the active convectional overturning of a summer thunderstorm—what really significant effect can the slightly damper and slightly cooler air of the forest play in the process of producing or determining the amount of the rainfall? We say, "the air over the forest is damper; therefore there will be more rainfall," quite forgetting that the damper air is useless as a source of precipitation unless it is cooled to the dewpoint. Furthermore, this moisture is constantly being carried away by the winds, and distributed through a great mass of air, thereby giving up more and more of whatever rainproducing effectiveness it may have had.

Whatever may be our personal prejudices, and whatever may be the theoretical considerations in favor of an influence of forests upon rainfall, what we really want is the facts, so far as they are at present available. Obviously, in a scientific study of this problem, the historical method of treatment, previously referred to; all theoretical considerations, and all prejudices, must give way before the results obtained by means of actual observations, made under approved conditions, with accurate instruments. There has been great difficulty in securing absolutely trustworthy observations. Many of the older records are clearly unreliable because of the improper exposure of the rain-gauges, the differences in the elevation or exposure of the instruments being enough to account for all the observed differences in their catch. Some excellent series of observations have, however, been carried on during the past twenty-five years or more in several European countries, by the agricultural and the forest experiment stations. A system of parallel or radial stations has been extensively used, these being located within forests and in the surrounding open country. Simultaneous observations extending over as many years as possible are compared, the greatest care being taken to have the best exposures, and to allow for the effect of the wind on the catch in the gauges.

The proper exposure of rain-gauges is one of the most perplexing problems in observational meteorology. Rainfall has long been known to be very "patchy," that is, there are considerable differences within very short distances. Thus it happens that gauges which are near together and under similar conditions of exposure often record quite