10 



precipitate the water wliicli is helu in suspension in tlie form of vapour ; that 

 partial rarefactions are thus jn'oduced in tlie upper part of the atmospliere, 

 in consequence of which currents of air ascend from the earth and others 

 rush in from different parts of the horizon to supply the deficiency at the 

 spot from whence the air ascended. A whiiiwind is thus the exact opposite 

 of a whirlpool ; in the one the neighbouring particles of water rush downwards 

 to a centre ; in the other they rush upward ; and in each case being met 

 obliquely by other currents their motion naturally assumes a rotary or spiral 

 direction. 



Mr. E-edfield, of the Am. Phd. Society, thinks that the whirlwind con- 

 sists of tivo movements : in the lower part of the atmosphere a spiral motion 

 inclining downward, and in the upper region a similar motion rushing upward. 

 This theory rests on the fact that sometimes an upward motion has been 

 observed in the clouds before the storm reaches the earth's sirface. It doe* 

 not appear, however, tha't this theory is founded on auy large induction of 

 facts ; and at present the balance of evidence seems to be in favour of the 

 simpler solution, viz., that the whirlwind ascends from the lower region of 

 the air towards the higher, and is produced by the rarefaction and consequent 

 ascent of the heated air. 



Su- John Herschel thinks that the Gulf Stream is the cause of the para- 

 bolic curves assumed by the path of the storms on the American coast ; and it is 

 certain that they are found very frequently to follow the course of the Stream 

 from W. to N. , while the increased temperature of the Stream may disturb 

 the equilibrium of the atmosphere, and thus carry forward storms which 

 originate in the tropics. 



It is to be noticed that these general laws (as they seem to be) are subject 

 on land to local and temporary causes, such as the direction of mountain 

 chains, ranges of hills, or isolated summits or depressions, which produce 

 looped curves, and sudden and irregular changes in the direction of the wind 

 at the surface of the earth, where of com-se it is first and most completely 

 observed. On water, as these causes do not exist, the course of the storm is 

 found to obey the general laws already explained. 



The whirlwinds of the Atlantic, which are the only phenomena of the 

 kind important to our subject, commence in that part of the ocean which is 

 marked on maps as the region of variable winds, viz., between 10° and 20° N. 

 and 55' and 60° W. They take their comse along the American coast, and 

 thence sometimes follow the Gulf Stream across the ocean. They are felt 

 usually between July and October, August being the month most subject to 

 them. The motion in all the instances recorded in America was N. by W. 

 and S. round to E. This indicates that as the vortex moved N. the air from 

 that quarter rushed in, that the air from the W. and S. followed the vortex, 

 and that as it advanced towards the N.E. the air from the E. met it, while 

 the rotary motion was produced by the meeting at various angles of the parti- 

 cles of air set in motion by all these currents. 



