1885.] theory of viscous solids. 277 



or the disappearance of heat which accompanies a change of state 

 would seem to indicate that the distinction does not depend 

 solely on the mean velocity. The phenomena of diffusion shew 

 that in a liquid which, apart from molecular motions, is at rest 

 the molecules pass, though very slowly, from one part of the mass 

 to another. In the case of a solid there is no evidence of what, 

 when we compare its duration with the time taken by the mole- 

 cule of a gas to describe its mean free path, we may call the 

 secular motion of the molecules; we assume that they oscillate 

 about positions of equilibrium, but of the forces which bind them 

 together we know nothing. 



The question then arises, what are the relations to one another 

 of the molecules of a viscous body ? 



We may say generally that the viscous condition is one of 

 transition between the liquid and solid states. A substance is 

 sensibly viscous only through a certain range of temperature; if 

 it is sufficiently cooled it becomes sensibly solid, if it is sufficiently 

 heated it becomes a more or less limpid liquid. The range of 

 temperature through which a body is sensibly viscous may be 

 considerable, as in the case of pitch or shellac, or it may be very 

 small, as seems to be the case with ice. 



If we consider a portion of a viscous solid, however small, 

 which is nevertheless large enough to be practically examined, 

 it appears to be homogeneous ; it does not follow however that 

 if we were able to magnify it sufficiently to come within a " raea- 

 sureable distance" of being able to see the individual molecules 

 it would still appear to be sensibly homogeneous. If it be allow- 

 able to apply the term " element " to a small parallelepiped whose 

 dimensions are large compared with the dimensions of molecules 

 but small compared with ordinary magnitudes, and which may 

 contain (say) a few hundreds or thousands of molecules, we may 

 say that it is quite conceivable that contiguous " elements " may 

 be very different in their molecular architecture. 



The hypothesis which I wish to suggest is that there is really 

 such a distinction between different small portions of a viscous 

 solid, and that we may describe the whole as consisting of a 

 comparatively solid framework the interstices of which are filled 

 by a more or less liquid mass. We may illustrate this conception 

 by a sponge with its cavities filled with liquid, or better by a 

 mass of gelatine which has imbibed water and passed into the 

 ordinary condition of a jelly; the essential difference being that 

 in these cases the solid framework and the liquid contents are of 

 different chemical composition while in the supposed case of the 

 viscous mass the " solid " and " liquid " parts consist of the same 

 substance in different states of aggregation. 



Of course any such picture of the molecular architecture of a 



