248 ON THE TRANSMISSION OF HEAT. 



Variation of AW thefe experiments were frequently performed, they- 



ftances. " nearly agreed, and the above are the average refults. Every 

 precaution was taken to enfure accuracy, in which I was af- 

 filed by feveral friends. . They were likewife occafionally 

 varied. In one inftance the cylinder of ice was made fo wide, 

 that a fmall thermometer was placed entirely within it in a 

 horizontal pofition, and covered with oil, but the remit was 

 nearly the fame. In others a folid brafs ball was fufpendcd in 

 the fluid inftead of the cup with hot water, but without any 

 material difference in the refult. In con fir tiding the veffelit 

 was found necelfary that the ice fhould be frozen hard, that 

 it might contain the fluid without allowing any of it to tran- 

 fude ; and in the experiment with the mercury the external 

 mould of tin was allowed to remain round the ice cylinder 

 to give it ftrength, and enable it to contain the mercury with- 

 out any rifk. This however could make no difference in the 

 experiment, and the refult was found to be the fame when 

 the tin was removed. 

 Df velopement of In all thefe experiments then, a rife of temperature in the 

 the proccfs. thermometer took place, greater or lefs, proportioned to the 

 quantity of fluid interpofed between it and the heated body, 

 From the nature of the experiment this rife could not be ex- 

 peeled to be confiderable, fince the fluid when heated mull 

 quickly have had its caloric abflracled by the cylinder of ice 

 in which it was contained. Hence at a certain flage of the 

 experiment the thermometer became ftationary, though the 

 heated body fufpended over it flill retained a confiderably 

 high temperature, the communication of caloric to the fluid 

 not being equal to its abftraction by the ice ; and when the 

 communication became lefs rapid than the abflraclion, the 

 thermometer began to defcend. In all thefe experiments, 

 however, even in thofe in which the largeft quantity of fluid 

 was interpoled, the rife of the thermometer was unequivocal, 

 and in fome of them, confidering the circumftances of the 

 experiment, very confiderable. 

 Conclusion. The This rife it appears to me impoflible to afcribe to any other 

 fuids -we.edireEl cau f e t ] lan to a p0W er in the fluid to condud caloric. Any 



conduaortofoeat. , , . , , . , , , ' >n 



other that might be imagined can be proved not to exiit. 



For the fides of Thus it is evident, that the fides of the veflei could not 



the veflel could convey to the fluid in contad with the bulb of the thermo. 

 jiot conduct : J ' . 



meter, any part of the calonc it received. Ice in common 



witji 



