SOME CONSEQUENCES OF GRAHAM'S WORK 599 



the probability that the elements generally are composite 

 materials may be said to have grown apace and the views 

 that Graham held as to the essential unity of matter have 

 in no wise lost force 



I may now summarise my conclusions, so that they may 

 be clear to those who are not specialists. First, let me again 

 insist that the elements which the chemist attempts to classify 

 are not the actual materials known to the general public 

 but certain abstract conceptions formed by considering the 

 properties and inter-relationships of substances generally. 

 Classification is not usually based on mere matters of fact, 

 so called; interpretation plays a large part and artistic feeling 

 also, as what are termed properties are mostly the expression 

 of reciprocal relationships and not absolute possessions. 1 

 Every one knows that the convenience of a house stands in no 

 very distinct relationship to the materials of which it is built 

 but is mainly determined by its structure and dimensions. 

 So it is in chemistry — structure is the most important factor 

 in determining properties. Thus the great majority of com- 

 pounds of carbon and hydrogen are colourless ; a few of a 

 particular structure are yellow or red ; and by combining 

 oxygen or nitrogen in certain ways with carbon and hydrogen, 

 so that a certain definite character of structure is developed, 

 we can obtain every shade of colour we desire. 



Taken generally, the elements present properties which are 

 those of a set of more or less closely related compounds ; 

 indeed, it is possible to arrange carbon compounds in order 

 of molecular weight in series and groups so as to match very 

 closely, within certain limits, the alterations in properties 

 which are noticeable in passing through the series of elements. 

 The unequal development in various parts of the system is 

 common both to the elements and to the carbon compounds 

 — in the case of these latter, compounds which " theoretically " 

 might be expected to exist have not infrequently proved to 

 be unobtainable. We may apply this argument to account 

 for some, at least, of the gaps in the record of elements. 



The one difference among elements which cannot well be 



1 The present-day tendency to reduce everything to numerical terms and 

 to exaggerate the value of decimals is but the expression of the fact that 

 there are many artificers but few artists. 



