470 



4 inches in the short branch, but varying in diameter, were placed 

 in the room near each other and simultaneously observed. 



In a tube of 3 inches uniform diameter, the revolutions were 3*0 

 per minute; in one of 5 inches 9*15, and in one of 6*75 inches 13*15; 

 their respective areas being 7*065, 15*708, and 21*205. In the 

 conical tube on its base, whose area was 14*529, the revolutions 

 were 8*8 per minute. It would seem, then, that the velocity has re- 

 lation rather to the mean area of the tube than to that of the 

 entrance and exit orifices, as the latter were the same in the tube of 

 3 inches uniform diameter, and in the conical tube on its base, while 

 the revolutions of the disc were 3*0 per minute in the former, and 

 8*8 per minute in the latter. When the exit orifice of the tube of 

 6*75 inches diameter was reduced to 3*5 inches, the rapidity of the 

 revolutions was reduced only about 10 per cent. 



The influence of temperature in accelerating or retarding the cur- 

 rents through the tubes next engaged the author's attention ; but 

 before entering into direct experiments, he found by very numerous 

 observations, that on some occasions no appreciable difference could 

 be observed in the temperature of the atmosphere of the room near 

 the floor and the ceiling, while on others there was a mean excess 

 of 0* 1 7 Fahr. near the ceiling without causing any perceptible dif- 

 ference in the velocity of the revolutions of the discs. In forty com- 

 parative observations of the temperature of the external surface of 

 the tube and of the surrounding air, that of the tube was 0*09 higher; 

 in twenty-three it was equal, but in only five was it lower than the 

 surrounding air. 



Of thirty-six comparative observations of the temperature of the 

 air within, and external to the tube, by a delicate mercurial thermo- 

 meter, it was found to be slightly higher within the tube in twenty- 

 seven, and in the remaining nine it was equal, but never lower than 

 that of the external air. The greatest excess was 0*4 Fahr., and 

 the mean excess 0*14 Fahr. 



The accuracy of these results was tested by an extremely delicate 

 differential thermometer, which also indicated a minute excess of 

 temperature within the tube ; the author is led however to infer, 

 that in the thirty-six observations the mean of 0> 14 is rather above 

 the true excess ; taking this however to be the exact amount, and 

 as the atmospheric air is increased only j^ of its volume, for every 



