HURRICANE 83 storms of the Indian ocean. The rate of progress must be carefully distinguished from the velocity of the wind, as the latter has no known relation to and far exceeds the former. The waves and swells produced by the hur- ricane winds are a most important feature; these waves are the largest and most formi- dable known to the mariner. They form with greatest regularity at points directly in advance of the approaching storm centre ; at other points they form a confused mass of crossed sea ; in the neighborhood of the land the con- fusion is increased by the waves reflected from the shores. Such is the equality of the con- test of opposing waves, that near the central region these sometimes lose their progressive movement and become stationary pyramidal waves, simply rising and falling. The smaller waves that are propagated in all directions from the region of severest winds, degenerate into long gentle swells that outrun the storm in its progress, and announce its presence sev- eral hours or a day in advance of its arrival. Besides these waves, it is believed that the extended region of low barometer allows the formation of a peculiar " cyclone wave," which is similar to the tidal wave of mid-ocean. The cyclone wave is coextensive with the area of low barometer; it is highest at the central lowest pressure, where its elevation above the ordinary sea level should be a foot or more for each inch of barometric depression. From the earliest times the months from July to Oc- tober have been known in the West Indies as the " hurricane season." A table published by Poey in 1855 gives the distribution by months of 355 hurricanes recorded on the Atlantic between 1493 and 1855. According to this work, there are recorded in this period in all in January 5, February 7, March 11, April 6, May 5, June 10, July 42, August 96, Septem- ber 80, October 69, November 17, December 1 ; bnt the annual period is probably not very correctly shown by this list, because of the imperfections of the earlier records. More recently Poey has revised his list and added many later hurricanes, and has published in the Paris Comptet Rendv* for Nov. 24, 1873, and Jan. 5, 1874, the results of a comparison between hurricanes and the frequency of solar spots. His results seem to remarkably confirm those of Meldrum, who had previously stud- ied the hurricanes of the Indian ocean from the same point of view. Poey states that in the majority of cases the years of the great- est number of hurricanes are also the years of the greatest sun-spot frequency. The ex- tensive researches of Koppen (1873) have shown that the amount of heat received from the sun varies annually with the sun spots, whence we infer that the variations in solar heat produce a similar variation in the terres- trial evaporation, and an increased tendency to the formation of hurricanes. The actual number of hurricanes visiting any limited re- gion is of course very small. Since the year 1700 the centres of about 25 have been known to pass quite near the coast of Georgia and South Carolina, which is by far the most fre- quently visited portion of the United States. Nearly all those of the Indian ocean pass near to the islands of Mauritius, Rodriguez, &c. Concerning the origin and cause of the hurri- canes of the Atlantic ocean comparatively little is positively known, but it seems by analogy that they may originate wherever the lower stratum of warm moist air is rapidly elevated above the sea level, whether (1) by being pushed up over an elevated plateau or mountain chain, or (2) by the under-running of a layer of cold dry air, or (3) by the conflict of two opposed and nearly balanced currents of warm moist air. In numerous instances one or the other of these cases seems to have oc- curred ; and as these, combined with (4) the radiation of heat into space, are the prevailing causes that determine the origin and growth of storms in general, there seerns no reason in the case of hurricanes to appeal to more forced theories. The immense mechanical power stored up in the heat and vapor of moist air has been abundantly demonstrated by Espy, Peslin, and Reye. Whenever, by the action of either of the four causes just men- tioned, the process of condensation of vapor into cloud, rain, or snow begins, there at once occurs an influx of air from all sides, and from below as well as from above, to fill up the par- tial vacuum thus created; this influx toward a central region is immediately followed, as shown by Ferrel, by the formation of a whirl whose subsequent development is entirely de- pendent on the supply of moist air. The hur- ricanes of the southern Indian ocean are thus generated in the region of calms between the N. W. monsoons and the S. E. trade winds of that ocean. Similarly hurricanes have been known to originate in the neighborhood of Florida when a cold north wind has swept under the warm moist air of the gulf and ocean. Another class originates in a similar manner in the western portion of the gulf of Mexico after a Texas norther has prevailed for a few days. A few begin in the interior qf Texas when a high barometric pressure on the gulf, or a low pressure in the western territo- ries, forces or draws the air of the gulf up over the plains of Texas. But by far the larger class of the Atlantic hurricanes, including those of greatest extent and violence, appear to origi- nate between the Windward islands and the African coast, and generally quite near to the latter; apparently these begin with heavy rains in the region of calms, such as are accompa- nied on the African mainland by the peculiar harmattan and tornadoes of that coast, which may be, so far as we know, either the conse- quence or the determining cause of the heavy rains. The storms that originate here may either move as far west as the American coast before recurving toward Iceland and Norway, or may describe a much shorter route, and
