But the most decisive, perhaps, as well as practically important evidence of the influence of magnetism, or magneto-electricity, in meteorological phenomena, is derived from the action of storms. My observation has been limited, for my life has been, and must be, a practical one. But, subject to future, and I hope speedy corroboration, or correction, by extensive systematic observation, I think I may venture to divide all storms into four kinds: 1. Those which come to us from the tropics, and constitute the class investigated by Mr. Redfield. That these are of a magneto-electric character is evident. They originate near the line of magnetic intensity, over, or in the vicinity of, the volcanic islands of the tropics; are largely accompanied by electrical phenomena; extend laterally as they progress north; induce and create a change of temperature in advance of them, and do not abate until they pass off over the Atlantic to the E. or N. E., and perhaps not until they reach the Arctic circle. Their extensive and continued action is not owing to any mere _mechanical agency_ of the adjoining passive air, or other supposed currents, originated, no man can tell how, but they concentrate upon themselves the local magnetic currents as they pass over and intersect them, and, by their inductive action upon the surface-atmosphere, in different directions, attract it under them, and within their more active influence. Here the action of the magnetic currents is probably the primary cause, but the power of the storm to concentrate upon itself the new magnetic currents which it intersects as it enters each new, successive field, enables them to maintain and extend their action. The following diagram illustrates the course and gradual enlargement of a mid-autumn tropical storm, which induces a S. E. wind in front, and occasions a thaw. [Illustration: Fig. 20.] 2. Another class originate at the N. W., and extend gradually south easterly on the magnetic meridian. These are most frequent in summer, forming belts of showers, but occur, I believe, at all seasons of the year. They seem to be produced by magnetic waves passing south, and are followed in autumn and winter, and sometimes in summer, by the peculiar N. W. wind and scud, and a term of cooler weather.