OF ARTS AND SCIENCES, 291 



known intense and persistent cause of rotation at a given 

 point in the upper atmosphere, and the effect of that rotation 

 could hardly be propagated downwards through an elastic 

 fluid by a narrow trunk, and without an adequate resisting 

 agency to the centrifugal force. From both observation and 

 theory, Dr. Channing was therefore disposed to regard the ro- 

 tation as incidental to these phenomena, rather than their pri- 

 mary cause. 



" It was necessary to have an axial cause which should con- 

 tinue to operate during the existence of the tornado, and con- 

 fine these energetic phenomena within the limits of the trunk 

 where the most powerful action takes place. The effect of 

 rotation was to produce dispersion rather than this intense 

 axial action so peculiarly restricted. 



" Tornadoes are described usually as preceded or followed 

 by electric phenomena, but rarely as accompanied at the same 

 time with active electrical discharges. In the tornado, how- 

 ever, which crossed Providence River, in 1838, the trunk was 

 seen to descend from the cloud, and the water to be agitated 

 and raised beneath it. Successive flashes of electricity then 

 passed through the trunk, apparently from the water to the 

 cloud above. After each discharge, the agitation of the water 

 appeared immediately to subside for a moment. Here was 

 the common phenomenon of the spark drawn from the prime 

 conductor, and the falling of the pith balls. The disturbance 

 and elevation of the water under the point of the descending 

 trunk, long before the completion of the column or any visible 

 mechanical connection exists, is a fact of common observation 

 in water-spouts. There is, therefore, reason from observation 

 to infer a silent discharge of electricity by means of these 

 trunks during a tornado. Such a discharge is, indeed, always 

 a necessity of the case, the trunk of the tornado serving as a 

 partial conductor between the clouds and earth. The tornado 

 seems to exist, as a general rule, precisely when the moisture 

 of the air or some other cause determines a silent discharge of 

 electricity, instead of a discharge in the common form of the 

 flash. 



