202 Mr. M. M. Pattison Muir on Chemical Classification. 



about the doings of the body formulated ; and it attempts to 

 connect those properties which it generalizes with the struc- 

 ture or composition of the chemical molecule. Attempts were 

 made from the time of Berzelius to indicate by a formula 

 something of the inner structure of the molecule. It is, how- 

 ever, only since we gained the idea of the molecule as a de- 

 finite atomic system, each part depending upon each other, 

 that these attempts have met with any fair measure of 

 success. 



In paragraph 20 of the present paper I have endeavoured 

 to point out the difficulties in the way of certainly recogni- 

 zing those compounds, the existence of which is possible ac- 

 cording to the theory of valency. These difficulties hold in 

 the present case. Compounds of this or that structure 

 are theoretically possible ; or this or that constitutional for- 

 mula may be assigned to a certain compound ; but how are 

 we to tell what is the structure of the compound before us ? 

 How are we to make sure that the constitutional formula we 

 adopt is the true one ? We cannot tell ; we cannot make sure. 

 Like other scientific questions, this is a question of probability 

 alone. And I think that the advances made by Chemistry, 

 and more especially by the chemistry of the carbon compounds, 

 in recent years, are the best proof we can have in favour of those 

 hypotheses upon which our constitutional formulae are based. 



It is very well to insist upon the dangers of undue theori- 

 zing, and upon the necessity of amassing facts. Facts alone 

 are of little use except as the foundations upon which true 

 know^ledge is raised. Is it not more scientific to venture an 

 hypothesis, once we have accumulated a considerable mass of 

 facts, and to use that hypothesis, than to remain contented 

 with mere empirical knowledge ? But all depends upon hoio 

 we use the hypothesis. If we allow it to carry us away — if, 

 for instance, in framing constitutional formulae, we forget 

 carefully to gather facts concerning reactions, or if we forget 

 the limits within which it is alone allowable at present to 

 make use of the theories of valency and of atom-hnking, even 

 of the theory of molecular structure itself^ — then in place of true 

 constitutional formulae we produce only ^' pictures " which 

 are worse than meaningless, inasmuch as the information 

 which they convey is in all probability false information, while 

 the knowledge assumed on the part of those who manufacture 

 them is certainly not as yet possessed by any human being. 



25. In paragraph 4 of this paper I stated that the various 

 systems of classification which have from time to time been 

 triumphant in Chemistry might be broadly divided into two 

 groups — those in which classification of chemical substances 



