320 Prof. Poweirs ^0/^5 on Repulsioji hy Heaty 8fc. 



as to disturb the colours : they continued unaltered for SO 

 seconds more, when the heat was withdrawn. 



A thermometer placed with its bulb on the central black, 

 rose 1° Centigr. in the first 30 seconds of this experiment. 



Another experiment was continued to see how long the 

 central black would remain. After about 4? minutes the lower 

 glass cracked, but not so as to disturb the black ; at 5"^ 30* 

 it disappeared, and the other colours rapidly vanished : in 

 this case the flame was applied closer to the iron plate than 

 in the former. 



(6.) The repulsion between two plates at the small intervals 

 here employed, may be compared with the repulsion between 

 the molecules of the solid body which produces expansion. 



In experiments of this kind we have an interval (t) be- 

 tween the glasses, which is increased to (t + i) ; by a given 

 increase of temperature we can easily compare the ratio of this 

 increase with that of the glass itself by expansion for the same 

 increase of temperature. 



A thermometer having its bulb kept in contact with the 

 lower glass, (the colours being formed close to the edge of a 

 smaller glass laid upon it, and the bulb of the thermometer 

 being close to the same point,) the increase of temperature was 

 noted which reduced the tints from the bright of the first to 

 that of the fourth order. 



In several trials, an increase of only 2° Fahr. was sufficient 

 to produce this effect. 



and i=:QT or t becomes t (1+6). 



From the experiments of Lavoisier and Laplace the dila- 

 tation of plate-glass for a difference of 180° Fahr. is 1-00089, 

 or in this case r^ becomes (1 +'00089) r, consequently for a 

 difference of 2° t becomes (1 + -000098) r. 



^, . . , , i 6-000000 61223 , 



The ratio m the two cases -9- = = — - — nearly. 



(7.) I made some rough attempts towards ascertaining the 

 law by which the repulsive effect of heat between the plates 

 of glass may be regulated, in relation to the interval between 

 the plates. Though these results are hardly of sufficient 

 precision to justify any decisive inferences, yet they may be 

 mentioned, as perhaps the experiments may be thought not 

 unworthy of being repeated. The process consisted in ob- 

 serving the times of the central tint changing to another given 

 tint, commencing from different points in the scale, with a con- 

 stant heat. 



