ADDT^ESS. 11 



This point is so important tliat it requires full discussion, l).ut in 

 dealing with it, it is necessary to distinguish carefully between the validity 

 of the arguments which support the earlier and more fundamental pro- 

 positions of the theory, and the evidence brought forward to justify mere 

 speculative applications of its doctrines which might be abandoned 

 without discarding the theory itself. The proof of the theory must be 

 carried out step by step. 



The fii-st step is concerned wholly with some of the most general 

 properties of matter, and consists in the proof that those properties are 

 either absolutely unintelligible, or that, in the case of matter of all kinds, 

 we are subject to an illusion similar to that the results of which we 

 admit in the case of Saturn's rings, clouds, smoke, and a number of 

 similar instances. The believer in the atomic theory asserts that matter 

 exists in a particular state , that it consists of parts which are separate 

 and distinct the one from the other, and as such are capable of indepen- 

 dent movements. 



Up to this point no question arises as to whether the separate parts 

 are, like grains of sand, mere fragments of matter ; or whether, though 

 they are the bricks of which matter is built, they have, as individuals, 

 properties difiPerent from those of masses of matter large enough to be 

 directly perceived. If they are mere fragments of ordinary matter, they 

 cannot be used as aids in explaining those qualities of matter which they 

 themselves share. 



We cannot explain things by the things themselves. If it be true 

 that the properties of matter are the product of an underlying machinery, 

 that machinery cannot itself have the properties which it produces, and 

 must, to that extent at all events, differ from matter in bulk as it is 

 directly presented to the senses. 



If, however, we can succeed in showing that if the separate parts have 

 a limited number of properties (different, it may be, from those of matter 

 in bulk), the many and complicated properties of matter can be explained, 

 to a considerable extent, as consequences of the constitution of these 

 separate parts ; we shall have succeeded in establishing, with regard to 

 quantitative properties, a simplification similar to that which the chemist 

 has established with regard to varieties of matter. The many will have 

 been reduced to the few. 



The proofs of the physical reality of the entities discovered by means 

 of the two analyses must necessarily be different. The chemist can 

 actually produce the elementary constituents into which he has resolved 

 a compound mass. No physicist or chemist can produce a single atom 

 separated from all its fellows, and show that it possesses the elementary 

 qualities he assigns to it. The cogency of the evidence for any 

 suggested constitution of atoms must vary with the number of facts 

 which the hypothesis that they possess that constitution explains. 



Let us take, then, two steps in their proper order, and inquire, firat, 



