«>42 0N THE TRANSMISSION OF HEAT. 



matter of the veflel muft convey more or lefs of this caloric 

 downwards, and communicate it to the fluid beneath ; and 

 thus caloric may appear to be conducted to a cpnfiderable 

 depth, though the fluid do not actually poflefs a conducling 

 power. In any. ufual mode of performing this experiment, 

 this fource of fallacy cannot be intirely obviated ; nor can its 

 effects be eftimated with fuch accuracy, as to determine pre- 

 cifely how far any obferved rife of temperature in a fluid is to 

 be afcribed to its operation. 

 Error from the It occurred to me, that this error might be completely 

 bv ibrnSng 1 it of av °ided by employing a hollow cylinder of ice to contain the. 

 ice. fluid. Suppofe a thermometer to be placed horizontally with- 



in fuch a veflel; its bulb jaeing in the axis of the cylinder, and 

 covered with a fluid at the temperature of 32°. Let a hot 

 body be brought in contact with the upper furface of the fluid, 

 or even immerfed in it, but ftill fo as to be at fome diftance 

 from the bulb of the thermometer ; if the thermometer rife it 

 may be concluded that the fluid has a conducling power, ftnee 

 the caloric could reach the bulb in no other mode. Caloric 

 does not .pafs through fluids by radiation ; it is eafy to perform 

 the experiment fo as to prevent the. thermometer from being 

 heated by any motion of the fluid, from immexiing the hot fo- 

 lid ; and laftly, no caloric could be tranfmitted by the fides of 

 the veflel, fince ice cannot have its temperature raifed above 

 32°, and cannot therefore communicate any temperature above 

 If the thermo- that to a contained fluid. The conclufion therefore feems un- 

 meterdo rife the deniable, that if in fuch an experiment the thermometer rife, 

 du&or. caloric muft be conveyed to it by the fluid inferpofed between 



it and the heated folic!. 

 Whether the op- It fecmed doubtful, however, whether the converfe of tho 



fn^be'affilmed" P ro P ofiti <™ wou ^ d hoId C( -] ua,1 y g ood - If the thermometer 

 in cafe the ther- were not to rife, would this prove that the fluid is a perfect 

 life 1 ?— N d ° n<>t non - condu & or ? lt is 110t evident a priori, whether a fluid 

 contained in a veflel of ice is capable of being heated above 

 3,2 9 , or at lean could convey caloric above that temperature 

 to any. perceptible diftance, even allowing it to have a con- 

 dueling power. Suppofe that the heated folid fs immerfed in 

 the fluid, the particles in contact with it will be heated, and 

 will form an afcending current which muft flow towards the 

 iides of ^e. veflel. The ice will abftract the excefs of caloric 

 from thefe heated particles, and they will return to the tem- 

 perature 



