The Theory of Glacier motion. 97 



ultimate factor to which we can reduce any piece of matter, 

 and being so, is absolutely unalterable in shape and in size, 

 we cannot divide it or alter it without constituting a new 

 substance altogether. The different forms assumed by 

 each substance are not due in any way to an alteration in 

 its ultimate factors, namely, its molecules, but to the re- 

 arrangement of these same molecules. We cannot, as Dn 

 Croll does, speak of the melting of a crystalline molecule^ 

 and of its flowing into the interstices between adjoining 

 molecules, nor can we speak of a liquid molecule solidifying 

 and assuming the crystalline shape. This language, and the 

 whole induction based upon it, ignores entirely the real nature 

 of molecules. If we put aside this fundamental contradiction, 

 and understand Dr. Croll as referring not to molecules but 

 to the minute particles which are aggregated together in al 

 masses of ice, each particle consisting of a congeries of 

 molecules, which when it melts are loosened, we are no 

 nearer a rational solution of our difficulty. 



Suppose we grant that, in melting, these particles lose 

 their polarity and arrange themselves so that they get better 

 into the spaces between those forming part of the crystalline 

 substance, and better also among themselves, so that on the 

 whole they occupy less room than before, Mr. Croll seems 

 to argue that in solidifying these particles will find the 

 spaces or interstices in which they solidify too narrow for 

 them, and must consequently squeeze their neighbours 

 asunder, and thus cause the ice to expand. To this Mr. 

 Blake replies : " Is the fragment or particle of ice to which 

 the argument is applied divisible or indivisible under the 

 ordinary forces of nature ? If divisible, why does it not 

 flow into several interstices, and get squeezed into more in 

 the act of solidifying, and make several small crystals 

 instead of one bigger one? If indivisible the main mass 

 would remain in the original cavity, and could only crys- 

 tallise by coming back together, or by the main mass 



