398 Mr. C. Barus on the Secular 



that arch, disintegrate, and if the decomposition be of a kind 

 that the debris may be chemically recognized, then, by coordi- 

 nating the viscous and the chemical observations^ I obtain a 

 fairly good notion of the molecular mechanism which has 

 produced the viscous effect. This is the method I applied to 

 prove Maxwell's theory. But partial disintegration or recon- 

 struction of molecules may easily be conceived to occur in 

 such a way as to escape detection altogether, or at least to 

 escape detection as much as does a break-up of configurations 

 of molecular wholes. It may, I think, be reasonably sup- 

 posed that the solid molecule is of the form M„, in which n is 

 variable ; and any given value of n will occur less frequently 

 according as it is proportionately greater or less than the mean 

 or typical value of n for the solid. The effect of strain, of 

 change of strain, or even of long continued secular subsidence, 

 is merely to vary the distribution of molecules M M , so that the 

 general properties of any observably finite part of the solid do 

 not change*. 



I will adduce other points of view in succeeding papers ; 

 but the suggestions stated are sufficient, I think, to show the 

 importance of discerning whether the cause of viscous defor- 

 mation in tempered steel is the type of viscous deformation in 

 general ; or whether, in using the phenomena of annealing 

 steel to test Maxwell's theory, I have merely interpreted the 

 exception to prove the rule. The result of such an investiga- 

 tion, besides throwing light on the structure of solid matter, 

 must ultimately lead to inferences bearing directly on the 

 questions of fusion, solution, and volatilization. 



This inquiry being essentially a comparison of detail, will 

 consume much time before it can be brought to an issue ; 

 and as the individual parts of my work have led to results 

 which are of interestf apart from the purposes to which I 

 hope ultimately to apply them, it is perhaps best to commu- 



* In other words, it is here supposed that the assumption of solid state 

 has not thoroughly eliminated either the liquid or even the gaseous mo- 

 lecule ; and that it has, on the other hand, produced molecules of an 

 ultra-solid complexity. Here I touch the keynote of certain difficulties 

 in my mind against conceiving the solid molecule as a uniformly distinct 

 whole, capable of rotation (Kohlrausch viscosity), or capable of translation 

 from strained to isotropic configuration (Maxwell viscosity). I look upon 

 it as a more cumbersome thing, which may under favourable conditions 

 even lose its identity, and which, when tmdergoing the motions stated, 

 comes into serious conflict with contiguous molecules. Such a view, 

 even if it be mere surmise (excepting the case of hard steel discussed) is a 

 suggestive working hypothesis ; for it seems broad enough to enable me 

 to circumscribe the true phenomenon by clipping the postulate. This is 

 the general endeavour of my present work. 



t It is well known that experiments on the secular changes of metals 

 are being made at Glasgow by Mr. Bottomley. Cf. Brit. Assoc. Report, 

 p. 537 (1888). 



