AN INTERKSTING PSEUDOSOLID 297 



sidered, the planets can no longer be so regarded, their attrac- 

 tion being in reality perpendicular to their spheroidal surfaces- 

 It seems as if similar considerations must apply also to mole- 

 cules. If a molecule is in fact a space bounded by equipotential 

 surfaces and filled with a swarm of moving corpuscles, the 

 attraction here too must be perpendicular to the equipotential 

 surfaces, and the molecule will be centrobaric onl}' under cer- 

 tain limiting conditions. It thus seems possible to think of an 

 isotropic body as composed of fourteen-sided molecules, not 

 always in their simplest shape but answering to Kelvin's figure 

 after distortion has taken place. 



We found it impossible to produce linear compression of pris- 

 matic masses of foam without a certain amount of permanent 

 set. Reflecting on the nature of the pseudosolid, it appears 

 fairly certain that the bubbles were not all of one size, in spite 

 of all care which might be applied to making the mass fine- 

 grained and homogeneous. Partly on this account also, the 

 orientation of the several pseudomolecules cannot have been 

 uniform. Now, if such a mass is subjected to a linear com- 

 pressive stress, it is clear that some pseudomolecules must 

 be almost in a position of labile equilibrium so that even a small 

 amount of distortion must push some of the bubbles into new 

 positions, the edges of some of the tetrakaidekahedral molecules 

 being forced beyond the corresponding edges of their neighbors 

 in such a way that when the pressure was removed they could 

 not spring back into their original positions. Even the mere 

 lack of uniform orientation of the pseudomolecules aside from 

 tending to set up unstable equilibrium w^ould seem seriously to 

 affect the results of the application of force, since the resistance 

 which they offer must differ somewhat according to the direc- 

 tion of the several faces ; thus a force applied to the plane 

 cubical faces must produce different results from one applied to 

 the undulating octahedral faces of the pseudomolecule. Con- 

 sequently, even if there were no difference in size, some of the 

 pseudomolecules would break or be so distorted as to escape 

 from their original positions of equilibrium before others were 

 similarly affected. 



Do not these facts throw a certain amount of light on the 



