356 PASSAGE OF HEAT THROUGH FLUIDS. 



cles ; and it is certain, if this conjecture is founded, liquids 

 must necessarily become conductors of heat (though very im- 

 perfect ones) in ail cases where this mobility of their particles 

 is destroyed, as well as in these rare but yet possible cases, 

 where a change of temperature can take place in a liquid with- 

 out giving Its particles any tendency to move, or to be moved 

 out of their places. 

 In all ordinary The unequivocal results of a great many experiments have 



cases they shewn, that in ordinary cases, and perhaps in all cases where 

 transport heat. / _ •' ' ^ 



heat is propagated in considerable masses of fluids, its distrir- 



bution is accomplished precisely in the manner that the nevr 



theory supposes, that is to say, by currents. And it is certaia 



that the knowledge of that fact has enabled us to explain in a 



satisfactory manner several interesting phenomena of nature, 



which before were enveloped in much obscurity. 



It is known When a hot solid body is plunged in a cold liquid, there can 



that an hot ^^ ^^ doubt concerning the existence of the vertical ascending 



Dody heats a " 



fluid upwards 5 Currents which are formed in the liquid, and which convey to 



but can it do j|^g surface the heat which its particles have received ; but 



thisdovvn- ^ 



wards? with respect to the strata of liquid situated under the hot body, 



are they, or are they not heated by this body, by meajjs of a 

 direct communication of heat from above doxinwards, from 

 particle to particle, these particles remaining in their places f 

 This is a question on which philosophers are not yet agreed. As 

 it is a question of great importance, I have long meditated on 

 the means of deciding it ; and after several unsuccessful 

 attempts, I have at last succeeded in making an experiment 

 which I think is decisive. 

 New experi- As the apparatus which I used for this experiment, and 



ni^nt, which I have the honour of laying before the assembly, is 



somewhat complicated ; and as it is indispensably necessary to 

 be intimately acquainted with it, in order to form a judgment 

 concerning the degree of confidence which the results of the 

 experiment may deserve, it is necessary to give a detailed des- 

 cription of this machinery. The annexed figure gives a distinct 

 representation of its principal parts. It is drawn on a scale 

 of a quarter of an inch to the inch, English measure. 



AB, Fig. 2, Plate X. is aboard, of oak, seen in profile; it is 

 J.^ inch thick, 18 inches long, and 11 inches in breadth. It 

 serves to support two square upright pillars, CC, 1S| inches 



k 



