200 Mr. M. M. Pattison Muir en Chemical Classification. 



the constitutioniil formula of a compound is based upon the 

 hypothesis of valency and upon the hypothesis of atom-linking, 

 which is itself an outcome of the former hypothesis. The 

 idea of atom-linking has gradually arisen in chemistry ; it is 

 based more immediately upon the theory of types as inter- 

 preted in the light of the general theory of valency. From 

 the vague ideas entertained in the earlier days of the atomic 

 theory concerning the manner in which the atoms in a com- 

 pound mutually influence one another, we have gradually ad- 

 vanced to clearer conceptions of atomic action. We now 

 regard each atom in a compound molecule as exerting an action 

 of some kind upon that atom, or upon those atoms situated 

 near itself ; we look upon chemical action as an action between 

 atom and atom ; we imagine a regular atomic system of some 

 kind, not a conglomeration of atoms confusedly huddled toge- 

 ther. We have attained to the notion of atomic " chains " 

 and atomic "rings "* &c. 



Most chemists now regard the molecule as an atomic system 

 the members of which act and react upon one another in ac- 

 cordance with definite laws. The system is regarded as bound 

 together in one whole, each part depending immediately upon 

 the part next itself, and each depending mediately upon 

 every other. From this idea, and from the theory of valency, 

 the possibility of the existence of various kinds of atomic 

 structures is deduced. For instance, a series of divalent atoms 

 is capable of combining in two ways, either to form a " chain," 

 such as — — S — — , in which case the molecule is evidently 



unsaturated, or to form a "ring," such as / \ , in which 



case the molecule is saturated. The chain is evidently capable 

 of taking up two monovalent atoms, and so becoming a satu- 

 rated molecule although still maintaining its " chain " cha- 

 racter : thus we know of the compound SO2 Clg; which is pro- 

 bably constructed thus : CI— 0— S— 0— CI. 



Constitutional formulae are, then, based upon the theory of 

 valency and the theory of atom-linking f. In determining 

 these formulae, certain subsidiary assumptions are generally 

 made. Of these, perhaps the most important are the as- 



* Of course we cannot for a moment regard the atoms in a molecule as 

 fixed immovahly in certain positions : we must look on them as in a 

 continual state of motion. But this is not opposed to the idea of the 

 maintenance of definite relative positions by the moving atoms, 



t This statement may be denied by those who use the expression 

 " constitutional formula" in a sense different from that in which I am 

 now employing it. See forward. 



