190 Mr. M. M. Pattison Muir on Chemical Classification, 



elements, we attempt to measure the probability in favour of 

 the second assumption ; and from the probability so obtained 

 we attempt further to measure the probability in favour of the 

 first assumption *. 



So far as the first links in this chain are concerned there is 

 no great difficulty ; given fixed valency, the possible combi- 

 nations of a number of atoms are readily deduced ; but when 

 we come to deal with general characteristics and with chemical 

 analogies, all is vague and unsusceptible of quantitative treat- 

 ment. And then there comes the dread, " What if valency 

 is not fixed? what if it is an essentially variable quantity?" 

 And then we seek to allay the uneasy feeling by appeals to che- 

 mical analogies and general characteristics ; and so we get into 

 a circle and become entangled, and perhaps in despair seek 

 refuge in those graphic formulae and strange pictures which, 

 like some religions, demand implicit assent from their victim, 

 but promise him in return rest and quietness of mind. 



I cannot help thinking that chemists sometimes fail to dis- 

 tinguish the kind of evidence upon which constitutional for- 

 mulas are based (but here again we may disagree as to the 

 meanings to be assigned to the words "constitutional for- 

 mulae"), and that they are inclined to repose too much trust 

 in these formulae. 



Some have rushed to the other extreme ; they have, "with 

 that "infinite capability of doubting" which is said to be a 

 characteristic of the " ideal reformer," refused to have deal- 

 ings with the commonly accepted notions which connect 

 together valency, structure, and properties, and have endea- 

 voured to found upon a system of " dynamic equivalents " a 

 more definitely quantitative classification of chemical sub- 

 stances. Their attempts are worthy of all success ; but the 

 task is a long and a laborious one. I can see no reason for 

 not using both methods in the solution of the general problem 

 of classification. However we may succeed in deducing dy- 

 namical equivalents, we cannot overlook the connexion which, 

 it appears, must exist between composition and general action. 

 " Composition," to quote the words of Professor Foster in the 

 article " Classification " in Watts's ' Dictionary ' (vol. i. 

 p. 1022), " Composition is the most fundamental of all che- 

 mical properties ; and the manner in which all other chemical 

 properties depend upon it is the fundamental problem of 

 Chemistry." 



The upholders of the " position theory," it appears to me, 

 recognize this truth, although sometimes the intensity of their 



* No shadow of an attempt at the actual quantitative measurement of 

 any of these probabilities has been, or can be as yet, attempted. 



