Mountains serve to deflect the winds blowing against them either to one side or upward, the former frequently producing important changes in direction of the surface air-currents, and the latter, by causing the air to rise, permits of its expansion and consequent cooling, thus favouring precipitation. For this and other reasons precipitation increases with elevation, at least until an altitude of many thousands of feet is reached, and the mountains are cooler and more humid than the adjacent valleys. The air-currents on passing over a mountain range and descending are warmed by compression, and having lost a part and in many observed instances a large percentage of the moisture they previously contained, become warm, drying winds. The chinook winds, as they are termed in America, are marked examples of the influence exerted by mountains on climatic conditions. What are termed above the secondary conditions, tending to modify climate, produce such great changes in the distribution of rainfall, temperature, etc., and in the influence of the planetary winds, that the subdivision of the northern hemisphere into torrid, temperate, and frigid zones, while based on astronomical data, does not serve to represent actual conditions, except in a general way, in reference even to the single element of temperature expressed in these names. A comparison of the isotherms and of the distribution of precipitation as indicated on the preceding maps, with the parallels of latitude, shows at once that these two most important elements of climate are conspicuously independent of distance from the equator. A logical basis for subdividing the continent into _climatic provinces_ must therefore be sought in other directions. CLIMATIC PROVINCES In order to obtain a comprehensive idea of the climate of a continent, it is obviously desirable to subdivide it into areas having more or less similar and distinct atmospheric conditions. The leading difficulty in making such subdivisions is the well-known fact that the climate of any region which may be chosen passes by insensible gradations into that of adjacent regions, and any boundaries that may be drawn are to a considerable extent artificial and arbitrary. While the true basis on which to establish climatic areas or provinces is the resultant of all the weather elements which go to make up the atmospheric conditions recognised under the broader term climate, so many factors have to be considered that it is extremely difficult to arrive at a general acceptable conclusion. The principal elements of the weather to be considered are, approximately, in the order of their importance, beginning with the one which exerts almost universal control--temperature, precipitation, the winds, absolute and relative humidity of the air, and evaporation. [Illustration: PLATE III.--Climate & Life Provinces.]