ON THE PASSAGE OF HEAT THROUGH Ft-UIDS. 1^7 



is changed. It is evident, however, and is admitted by Count 

 Rumford himlelfj that the experiments by which thefe fads are 

 eitabliihcd, coniidered in themfelves, do not prove the pro- 

 poiition, that change of temperature in any fluid is iblely ef- 

 fected by this caufe, for although thefe motions may accele- 

 rate the change, it is ftill poffible, it may likewife in part be 

 owing to a communication of caloric from one part of the fluid 

 to another. 



There is another principle, however, from which Count But it might be 

 Rumford's opinion may be cftabliihed. It is evident that -if it J^jJ^JJ* 

 be juft, a fluid muft be incapable of being heated downwards, fluids cannot he 

 If a hot body be applied to its fuperior furface, the upper ftra- h " ted from 

 turn of fluid will have its temperature raifed, but it will com- 

 municate no part of the caloric it has received to the fluid be- 

 neath, and as it cannot change its place, but muft remain at 

 the furface, the under parts of the mafs muft remain unaltered 

 with refpect to temperature. 



It is by eftablifliing accurately the fact with refpecl to this Many of the ^ 

 point, that the queftion regarding the conducting power of m ents are^?-""* 

 fluids muft be determined. Count Rum-ford has made a num- retted to this 

 ber of experiments to prove the propofition, that caloric isnot pr0 ° J 

 propagated through a fluid downwards. Of thefe the moft particularly that 

 conclufive feems to me to be that in which a heated cylinder cy n n aer did not 

 of iron was fufpended in olive oil, and diftant only two-tenths melt ice placed 

 of an inch from a projetfion of ice fixed at the bottom of the beneath Dlive oi1, 

 veflel, without any of the ice being melted. On the fame 

 principle I employed the apparatus, of which a reprefentation 

 is given Fig. I. 



The thermometer A is bent fo that the leg a, which ifTues New experi- 

 from the bulb is longer than the leg b, to which the fcale is at-™^ Appa "' 

 tached. It is fo far filled with mercury, that at any moderate 

 natural temperature the leg <z (hall remain completely filled. 

 At the bottom of b the fcale commences with 20° of Fahren- 

 heit. This bent thermometer -is attached to a glafs rod, in- 

 ferted in a circular .piece of wood. This -is fixed by means of 

 wax to the bottom of a glafs cylindrical veflel B, three inches 

 in diameter, and nine in height. -With this apparatus, the 

 following experiments w-ere made. 



Experiment 1. Into the veflel R, water was poured till it L He f ed o]]f 



r . . poured upon 



covered the bulb of the thermometer J of an inch ; its tempe- water, heated a 

 rature was 4S e of Fahrenheit, which was like-wife the tempe- thermometer 



r through the wa* 

 rature ter downwards. 



