of Poit'idial Knenjy of Liquid Surfaces. 45 



Conversely, if the free liquid surfaces disappear at the time 

 of the condensation of the vapour, the potential energy of 

 which those surfaces were the seat vanishes, but is found again 

 in the form of heat sensible to the thermometer. To me the 

 preceding theory appears to show perfectly the origin and the 

 nature of the latent heat of vaporization ; indeed, taking for 

 our starting-point the incontestable existence of the potential 

 energy of a free liquid surftice, we have arrived by calculation 

 at this result — that every enlargement of the surface develops 

 cold, as every diminution or suppression of a free surface will 

 give rise to a production of heat. 



2. In the second place, when a solid body is brought to a 

 temperature sufficient to effect its fusion, the surface of con- 

 tact of the already fused layer and the yet solid nucleus pos- 

 sesses negative potential energy ; that is to say, if it is de- 

 stroyed cold is produced. Now, with every fresh layer melted, 

 the primitive contact-surface is replaced by another which is 

 necessarily smaller ; and consequently there will be cooling, 

 if the heat supplied be not sufficient to compensate this effect. 

 It is evident that the fusion will be by so much the more 

 active the more quickly the heat supplied can replace one con- 

 tact-surface by another smaller. 



Conversely, when solidification takes place, there are also 

 formed successive surfaces of contact between the solid and 

 the still liquid material ; but in this case, instead of going on 

 diminishing, the surfaces grow larger and larger, and must 

 consequently produce a progressive heating. It is thus that 

 formula, (5) renders intelligible the origin of the latent heat 

 of fusion. 



III. Let us now suppose that S and t vary simultaneously, 

 as generally happens ; we shall then necessarily have the com- 

 bination of the two preceding effects (I. and II.) ; only, as we 

 may easily convince ourselves, the influence of the surface 

 will far exceed that of the temperature. As examples in con- 

 nexion with the present hypothesis I may cite the evaporation 

 of liquids, the phenomena referred to the spheroidal state, the 

 solution of solids in liquids, &c. 



I shall not dwell upon these different instances, because, 

 after what has been said above, the formula can be applied 

 with great facility ; but I will endeavour to account for a sur- 

 prising fact recently studied by M. Spring*, concerning the 

 variations of the specific heat in the neighbourhood of the 

 maximum of density of certain bodies. 



Rose's alloy, immersed in an oil-bath, is gradually cooled 

 from 118°. The temperature falls rapidly to the point at 

 * See the memoir alreadv cited. 



