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



properties. If we can do this, we have very strong evidence 

 (1) of the existence of an essential connexion between struc- 

 ture and physical properties, and (2) in favour of the hypo- 

 thesis of valency as a foundation on which to build a general 

 scheme of chemical classification. As a matter of fact, che- 

 mists have not yet succeeded in completing the chemical 

 history of any group of isomeric bodies ; but I think that the 

 recent advances made in this direction warrant us in expecting 

 that the history of many groups will be thus completed — that 

 to do this is indeed a possibility, and that meanwhile to accept 

 the hypothesis of valency, as an liypotliesis^ is not to overstep 

 the limits of true scientific method. 



Again, given two or more elementary atoms, and given the 

 valency of these atoms, the existence of certain compounds 

 becomes, theoretically, impossible. Every failure, succeeding 

 honest attempts, to prepare these bodies lends strength to the 

 theory. We have instances of such failures. 



It also becomes possible, on the assumption that the valen- 

 cies of the constituent atoms of a compound are known, to 

 deduce the (theoretical) possibility or impossibility of the oc- 

 currence of certain reactions. It may be possible, we argue, 

 to introduce a fresh atom of a certain valency into this com- 

 pound ; or we say the introduction of an atom of this valency 

 must cause decomposition, — the compound, under the new 

 condition of affairs, can no longer exist. Every experimental 

 confirmation of such deductions is a point in favour of the 

 theory. Are we not continually meeting with experimental 

 confirmation of facts deduced from the theory of valency ? 



In attempting to deduce possible reactions &c. from the 

 theory of valency, and in attempting to confirm these deduc- 

 tions by appeal to experiment and observation, it is essential 

 that we bear in mind how very imperfect our theory as yet is. 

 If we attempt to apply it to solid and liquid bodies, we find 

 ourselves surrounded by contradictions and difficulties innu- 

 merable: we know almost nothing of the circumstances which 

 condition valency. We are ever ready to draw sweeping 

 conclusions utterly unjustified by the theory ; we are prone 

 to set up what might be called theoretical theories, vague, 

 shifting, indefinite ; and then, when the experimental evidence 

 contradicts our deductions, we are liable to fall into one of two 

 errors — either to manufacture a new, more vague, and more 

 elastic theory to suit the results, or to doubt the utility of 

 theory altogether. To use an hyjjothesis without abusing it 

 is one mark of a master mind. 



23. But in endeavouring to discover the relations which we 

 assume to exist between valency and chemical composition, it 



