ON THE PASSAGE OF HKAT THROUGH FLUIDS. ] 7 J 



perhaps be fuppofed, that by comparing the rife of tempera- ment cannot be 

 ture in the two thermometers, a concluiion might be drawn made dec 1 five. 

 with refpect to the mode in which the caloric was conveyed to 

 them. If it were intirely conveyed by the iides of the veflel, 

 fince as much ought to be given out by the external as by the 

 internal furface, the water without ihould be heated as much 

 as that within, and confequently the two thermometers fiiould 

 rife equally, and to the fame extent. But if the thermometer 

 in the inner veffel rofe more than that in the outer, it might. 

 be fuppofed, that the interpofed fluid had directly convened 

 to it part of the caloric from the heated ball above, with 

 which it was more nearly in contact. But to eftabliih fuch a 

 conclufion from any differences of this kind, certain circum- 

 ftances are indifpeniiblc, which are abfolutely unattainable. 

 It is requifite, for example, that both thermometers fliouM be 

 precifely on the fame level, at the fame diftance from the fides 

 of the veflel, and covered with the fame quantity of water oc- 

 cupying the fame volume. The fmalleft variation in thefe 

 circumflances would produce a variation in the refult. But 

 if they were even obtained alike, the fituation of the veflbls 

 rnuft ftill remain very different ; the one for example, prefents 

 a much more extenfive furface to the furrounding medium than 

 the other, and the nature of that medium is different, the in- 

 ternal veflel being furrounded with water, the temperature of 

 which is augmenting while the experiment continues, the ex- 

 ternal with atmofpheric air, whofe temperature remains the 

 fame. It is impoflible to eftimate the differences in effect 

 which mutt arife from thefe differences in fituation, and con- 

 fequently no conclufion can be drawn from the comparative 

 alterations of temperature in the refpect ive thermometers, as 

 to the quantity of caloric conveyed to either of them by the 

 iides of the veflel. jj 



The preceding experiments then, or any of a flmilar nature, The fides of the 

 are incapable of determining the queition reflecting the con- vefld Wl1 ' alwa P 

 ducting power of fluids. In all of them a quantity of caloric 

 is conveyed by the fides of the veffel in which the experiment 

 is made, and this quantity we have it not in our power to af- 

 certain, fo as to determine whether it is adequate or not to 

 produce the augmentations of temperature which actually take 

 place, and consequently whether any part of that augmenta- 

 tion is to be afcribed to a conducting power in the fluid. 



4 Every 



