306 PHILOSOPHY OF STORMS. 



If, however, we place our barometer in the lower end of 

 the receiver, and ascertain its fall there, it will give us the 

 actual velocity of the air upwards, free from friction, and 

 from the reaction of the air downwards, as it increases in 

 velocity in its upward motion. And in this case, too, the 

 power of the air to carry up heavy bodies will be the same 

 in every part of the receiver, if it is cylindrical, for the 

 velocity increases in exact proportion to its rarefaction. 

 For example, if the barometer falls an inch on being placed 

 in the receiver at its base, the air will rush up there with a 

 velocity of 240 feet a second, and at the top of the receiver, 

 being only one-third the specific gravity which it had be- 

 low, it will move with three times that velocity, or 720 

 feet per second. Now this velocity would be sufficient to 

 carry up heavy bodies of considerable size, and throw them 

 out at the top of the receiver, masses of ice, for instance, as 

 large as a man's head. 



If, however, we conceive holes of considerable magnitude 

 made in the sides of the receiver, at some distance from its 

 top, it will be manifest, that bodies might be carried up to 

 these holes, which could not be carried up beyond them ; 

 for so much of the air would escape through these .holes, 

 that the velocity of the current above them would be much 

 diminished. And thus these masses might play about near 

 these holes, without ascending much above them, or de- 

 scending much below them. Moreover, if the masses 

 should increase so much in number or size that the current 

 of air even below the holes could no longer sustain them, 

 they would begin to descend ; and if their cooling effect on 

 the air below should be very great, they might, partly by 

 their weight, and partly by their cooling influence, change 

 the current, and cause it to move downwards and outwards 

 below. 



During this operation the barometer at a would stand a 

 little higher than one on the same horizontal level at some 



