isobars, isotherms, state of the wind and weather, moisture and clouds over a large region of country. The accompanying charts, for April 3d, 8 A.M., and 4th, 8 A.M., 1892, show the general character of such daily weather maps; they will easily be understood by studying the respective legends. On these charts the reader will see the development of a storm that began with an area of low pressure in Colorado and rapidly developed into the great storm centre shown on Chart 2; the latter then passed northeastward over the Lake Region and the Gulf of Saint Lawrence and was followed by an extensive area of clear cool weather on April 15th. The movements and changes of storms and weather will undoubtedly be fully understood only in proportion as we have better knowledge of the facts and of the mechanical and physical laws that govern the atmosphere, but their approximate prediction from day to day is expected and demanded by reason of the many interests that depend upon the wind, temperature, and weather. At present such forecasts are generally based on the evident trend of events, as shown by comparing together the two or three latest weather maps, and in part also on empirical rules or generalizations, based on the study of similar types of maps in preceding years; but in some cases also one may be guided in part by general physical principles that must apply to the case in hand. The generalizations relative to storm movements for the United States, that is to say, the statistics of storms, have been presented in three memoirs by Prof. Elias Loomis, and printed in the Memoirs of the National Academy of Sciences. Similar data for the Northern Hemisphere as a whole were published in 1893 in Bulletin A of the United States Weather Bureau; this compilation is mostly the work of Prof. E. B. Garriott and is based upon ten years of daily maps (1878 to 1887), originally published in the Bulletin of International Simultaneous Observations. In this volume the paths of the storm centres are classified by different types and displayed on charts that show the frequency with which storm centres pass over each square of latitude and longitude.
Charts of storm paths for Europe, Asia, and Japan have been published by Germany, Russia, and Japan respectively, and monthly charts for the United States have been published regularly since January, 1873. By means of these charts one may, in a general way, anticipate the path and velocity of a storm centre when once it has appeared in any part of the Northern Hemisphere. In the Northern Hemisphere such centres move westward when they lie between the equator and the parallels of 25° or 30° N.; they then curve poleward and move northeastward with increasing rapidity toward the parallel of 60° or 70°. The variations from this general rule can best be understood by studying the charts of storm frequency. A similar rule holds good for the Southern Hemisphere, substituting only south for north. But little is known about the tracks of storms within the Arctic Circle. The region of greatest storm frequency extends in a narrow belt east and west from Lake Superior to Newfoundland and its prolongation eastward ends in the interior of Northern Russia. The region of next greatest storm frequency covers the islands of Japan. The north polar region of cold air, whose tendency is to flow outward toward the equator, is inclosed within an oval curve extending from Luzon over Japan, Southern Alaska, British Columbia, the region of the Great Lakes, Newfoundland, the Hebrides, Northern Norway and Sweden, and ending in Siberia at latitude 60° and longitude 90° east of Greenwich. South of this oval the prevailing winds are west and southwest; north of it they are north and east in the stormy season of the year.
The great whirls that we call general storms occur in connection with these polar and equatorial currents, but not necessarily between them. The whirls are explained as partially due to mechanical reactions between the northern and southern currents, but they are not merely hydrodynamic phenomena, since they have also an additional thermodynamic relationship which is quite as important. The warm, moist southerly winds are underrun by the colder and drier northerly winds. This enforced elevation of the southerly winds is accompanied by a corresponding expansion and cooling of the air that is thus elevated, and generally it is soon cooled to its dew point or below. This is followed by condensation of aqueous vapor and the formation of cloud, rain, hail, or snow with a great liberation of latent heat. Consequently the cloudy region will be warmer, but especially will it have a much smaller specific gravity than before.
In very small storms, such as tornadoes, waterspouts, etc., this process gives rise to very rapid uprising currents, a very rapid whirl around the central axis and a very low barometric pressure at the centre, but in extensive storms the vertical current is not so conspicuous, although the buoyancy of the central air tends very strongly to maintain the disturbance. The storm centre undoubtedly has a tendency to move toward the region in which the temperature and buoyancy are most disturbed; but as this region is always moving in advance, the storm centre will remain in the rear and its path will advance somewhat to the left of the direction of the greatest disturbance. But the uplifting of the lower moist air may be greatly intensified if the southerly winds on the eastern half of the storm area are being pushed up over high lands, or it may be almost wholly annulled if these winds must necessarily descend from the high lands to the ocean level. Therefore the relation of the storm's motion to the continents must be carefully worked out.
As regards weather prediction, it is evident at once that the descending winds and those that are coming from the north southward are being warmed up, and therefore in their presence the storm disappears and the weather clears away. For the Atlantic coast of North America rain is to be forecasted only when a south and east wind prevails, and especially when it is blowing on the coast. The actual effect of mountains, plateaus, continents, and the underflow of cold air varies so much on every occasion that the best one can do in forecasting is to familiarize himself thoroughly with the illustrations and exceptions that appear on every daily weather map.
The atmosphere would be at rest on the earth's surface and whirl about with the globe were it not for the sun's heat. All the important meteorological phenomena may be considered as resulting from the interaction of the solar heat, the moisture in the air, the varying temperature,
- Vol. XIII.—25.
