388 A CENTURY OF SCIENCE 



accompanying sketch showing wind directions, the reader 

 would infer that, at this time, Eedfield believed the 

 motion of the air to be very nearly in circles about the 

 storm center. The same idea is conveyed by a later 

 paper (42, 112, 1842). Espy (39, 120, 1840) of Philadel- 

 phia, however, claimed that observation showed rather 

 that the wind blew inwards toward a central point, if the 

 storm were round in shape, or toward a central line, if 

 it were oblong. This view Redfield (42, 112, 1842) con- 

 tested, and brought forth much evidence to prove its 

 falsity. A later statement (1, 1, 1846) of his own theory 

 is as follows: "I have never been able to conceive, that 

 the wind in violent storms moves only in circles. On the 

 contrary, a vortical movement . . . appears to be an 

 essential element of their violent and long continued 

 action, of their increased energy towards the center or 

 axis, and of the accompanying rain. . . . The degree of 

 vorticular inclination in violent storms must be subject, 

 locally, to great variations; but it is not probable that, 

 on an average of the different sides, it ever comes near to 

 forty-five degrees from the tangent of a circle, and 

 that such average inclination ever exceeds two points of 

 the compass, may well be doubted/' A qualitative 

 explanation of the effect of the earth's rotation on the 

 direction of the wind near the storm center had already 

 been given by Tracy (45, 65, 1843), and this was followed 

 some years later by Ferrel's (31, 27, 1861) very thorough 

 quantitative investigation of the dynamics of the 

 atmosphere. 



A number of individuals kept systematic records of 

 meteorological observations, among whom was Loomis, 

 whose storm analyses did much to settle the merits of the 

 rival theories of Redfield and Espy. In studying the 

 storm of 1836 (40, 34, 1841) he had drawn on the map 

 lines through those points in the track of the storm where 

 the barometer, at any given hour, is lowest. While this 

 method revealed the general direction in which the storm 

 was progressing, it failed to give much indication of its 

 size or shape. In discussing the two tornadoes of Feb- 

 ruary, 1842, one of which had already been described 

 in the Journal (43, 278, 1842), he adopted a new and 

 more illuminating graphical method. Instead of connect- 



