Evaporation, and Steam-Boiler Explosions. 29 



have, as the surface cohesion along the normal direction, 

 ^ = a, 1 + b. 1 + c, 1 + d, l + &c. But in the general expres- 

 sion we have, by observing the equality of terms, 



cc = a, 1 + 2(6 . l) + 3(c. l) + 4(d.l) +&c. 

 Comparing these values of m% we see that the surface layer 

 coheres to the mass with a very much smaller force than two 

 internal layers cohere against each other. For the second, 

 third, &c, layers, a like discussion applies, and the cohesion 

 gradually increases on penetrating the mass. 



This formula involves no particular hypothesis as to the 

 value or character of the forces acting, only that the aggre- 

 gate is attractive. But as condensation is a spontaneous 

 phenomenon through all that portion of the aggregation al 

 range in which energetic actions are found, we ought to as- 

 sume that all the effective terms are attractive. To present 

 the grounds which seem to me to authorize the conception of 

 that repulsion in all states of aggregation, is only exercised 

 between adjacent molecules, while the attractive actions are 

 the resultants of all the primary constitutional forces, and 

 extend through larger spheres, would involve the exposition 

 of a complete theory of molecular mechanics. I must, there- 

 fore, leave, as an assumption, the conception that in fluids the 

 only repulsion to be taken into account is that between the 

 contiguous layers (a and 1), which prevents their yielding 

 farther to the cohesive forces pressing them together. 



We should observe, that in consequence of the deficiency 

 of cohesion along the fluid surface, a rarefaction would take 

 place, which would again diminish surface cohesion to a con- 

 siderable extent below that value given by the formula. 



To determine the cohesion measured along a surface, as 

 we have done for that along the normal, let the general for- 

 mula be applied to a surface element. Then, instead of the 



I 



I I 



I I I I 



I ! 



I I 



11 I 1 I I 1 1 I I 1 



I I 



I I I 

 normal layers being full layers, they are essentially but half 



