Whirlwind, a general term applicable to a large class of storms. (See Cyclone, Hurrica'ne, and Water Spout.) In all these storms, except an occasional tornado, the air circulates with considerable regularity around a central region of calms. In some tornadoes this whirling movement is not so well marked as in others, which is explicable on the very plausible hypothesis that the axis of rotation is not invariably vertical, but is sometimes inclined or even horizontal, when the whirling mass of air rolls along like a barrel. In other tornadoes in which the whirling movement seems to be almost entirely wanting, the winds may be considered as due to a direct inrush of cooler or drier and therefore denser air, displacing and uplifting that which had previously lain near the ground. The rapid movements of tornadoes render it very difficult to obtain careful observations made simultaneously by observers on each side of the path, and of many of their features we are still in ignorance. Excellent special studies upon American tornadoes will be found in the works of Espy, Redfield, Hare, Loomis, and Chapellsmith, and the general treatise of Reye, Wirbelsturme (Hanover, 1872). See also the discussions in the Paris Comptes Rendus (1874-'5) between Faye, Peslin, and others, and the annual reports for 1873-5 of the United States army weather bureau.

In the water spout the whirling movement of the air is generally distinctly recognizable, but in the hurricane and typhoon this movement takes place on so grand a scale that it can only be demonstrated by collecting and plotting upon charts the observations made at widely separated stations. In June, 1874, Prof. Ferrel communicated to the philosophical society of Washington the mathematical expression that obtains in all whirlwinds for the relation between the barometric gradient and the velocity and direction of the wind, and his paper has been made the subject of further remarks by Hann in the journal of the Austrian meteorological society. In reference to the origin of hurricanes no more important observations could have been made than those recorded by Blanford in his memoir on the winds of northern India (London "Philosophical Transactions," 1874). This author gives in an appendix the results of his study of several hurricanes, which in substance are as follows: Cyclones are not produced between parallel currents flowing in opposite directions; a calm state of the atmosphere, or one in which the winds are light and variable over the open sea, is a favorable condition, and a second condition is a high or moderately high temperature.

The consequence of this collocation will be the production and ascent of a large quantity of vapor, which will be condensed with the liberation of its latent heat over the place of its production. If this state of things last for some days, the slowly inflowing winds acquire by the influence of the rotation of the earth a whirl, in consequence of which, as Mr. Ferrel has shown, the barometric depression must increase. The last step preceding, and apparently determining the formation of a well defined cyclone in the bay of Bengal is, according to Blanford, the inrush of a saturated stormy current from the southwest or west-southwest. But this last feature may be peculiar to that locality, and those previously enumerated seem to correspond best to the conditions generally observed in the formation of whirlwinds.