Matter and Force in Theoretical Physics, 473 



handled, which structure is with much probability ascribed to 

 particular arrangements of their discrete atoms, and will, I think, 

 eventually be accounted for by ascertaining the laws according 

 to which the rether (such as I conceive it to be) acts on the 

 individual atoms. But there is no reason from experience to 

 conclude that crystalline structure is the property of masses 

 that are not of comparatively limited extent. It plainly exists 

 only in the solid state, and is therefore dependent on the mole- 

 cular conditions which separate the solid from the fluid state. 

 But these conditions have reference only to the state of aggrega- 

 tion of the atoms very near the confines of the substance, which 

 clearly must be different from the state of aggregation in 

 the interior, where a particular atom is held in its place by 

 forces that arc, at least quam proxime, the same in all directions, 

 whereas the resultant action on an atom at and very near the 

 boundary must needs be much in excess towards the interior. 

 Now TyndalPs very instructive experiment of converting pounded 

 ice into a continuous whole by mere pressure, proves that the 

 superficial conditions of solidity, and the related conditions of 

 crystalline arrangement, may be crushed out by a sufficient 

 amount of pressure. There is therefore no evidence that in the 

 interiors of large masses like the earth or the sun, where the 

 pressure must be enormously great, there can be any atomic 

 arrangement or condition distinguishing the solid from the fluid 

 state ; rather, in the ease of the earth, there is evidence to the 

 contrary. For the fact that the average superficial form of the 

 solid parts of the earth is conformable to the ocean surface is to 

 be accounted for by supposing the condition of solidity to affect 

 only a comparatively thin crust, and to have very little influence 

 on the general figure. Consequently our knowledge of the pro- 

 perties of large masses cannot be said to furnish reasons for 

 attributing to the aether, of which both the extent and the pres- 

 sure are vastly great, a crystalline structure. Still less reason 

 is there for making the peculiar hypothesis of isotropic constitu- 

 tion, which was originally framed to account for the polarization 

 of light. This hypothesis is one of the class which Newton de- 

 clared that he abstained from making, and which for distinction 

 may be called speculative. Although the words speculation and 

 theory have properly the same signification, the former, since it 

 has acquired an unfavourable sense, may be used to distinguish 

 between a theory that rests only on personal and arbitrary con- 

 ceptions, and a theory that is truly such, resting on ideas de- 

 rived from sensational experience and common observation. 

 Where Newton speaks of " somnia/'' he is probably referring to 

 the Cartesian theory of vortices, which, although it was the arbi- 

 trary speculation of an individual, had an extraordinary hold upon 

 the science of that time, as is shown by the circumstance that a 



