PHYSICAL THEORIES OF HEAT. 585 



Laplace and Poisson have maintained the hypo- 

 thesis of molecular intr a -radiation, as the mode 

 in which conduction takes place : that is, they say 

 that the particles of bodies are to be considered as 

 discrete, or as points separated from each other, 

 and acting on each other at a distance; and the 

 conduction of heat from one part to another, is 

 performed by radiation between all neighbouring 

 particles. They hold that, without this hypothesis, 

 the differential equations expressing the conditions 

 of conduction cannot be made homogeneous : but 

 this assertion rests, I conceive, on an errour, as 

 Fourier has shown, by dispensing with the hypo- 

 thesis. The necessity of the hypothesis of discrete 

 molecular action in bodies, is maintained in all 

 cases by M. Poisson ; and he has asserted Laplace's 

 theory of capillary attraction to be defective on 

 this ground, as Laplace asserted Fourier's reasoning 

 respecting heat to be so. In reality, however, this 

 hypothesis of discrete molecules cannot be main- 

 tained as a physical truth ; for the law of molecular 

 action, which is assumed in the reasoning, after 

 answering its purpose in the progress of calculation, 

 vanishes in the result ; the conclusion is the same, 

 whatever law of the intervals of the molecules be 

 assumed. The definite integral, which expresses the 

 whole action, no more proves that this action is 

 actually made up of the differential parts by means 

 of which it was found, than the processes of finding 

 the weight of a body by integration, prove it to be 



