January 31, 1902.] 



SCIENCE. 



169 



resolve itself into four parts, two statical 

 and two dynamical, not one of which can 

 logically be neglected. First, there are the 

 substances which enter into the reaction; 

 secondly, the physical stimulus, thermal, 

 electrical or actinic, which starts the re- 

 action ; thirdly, the phenomena which occur 

 during the reaction; and finally, the sub- 

 stances produced by the reaction. An ini- 

 tial state of equilibrium is disturbed by 

 some application of energy; transforma- 

 tions of energy take place, and in a final 

 state of equilibrium the process comes to 

 an end. Through a mixture of gases hav- 

 ing certain physical properties we pass an 

 electric spark ; they unite to f oi-m a liquid 

 with different physical properties, the pro- 

 cess being attended by a change of volume 

 and great evolution of heat. The fact of 

 union is chemical ; the other phenomena are 

 physical ; and the two sets of considerations 

 are so interlaced that we are compelled to 

 take them together. Intellectually we can 

 discriminate between them, but the line of 

 demarcation is essentially ideal. The 

 chemical composition of matter cannot be 

 studied apart from its physical relations, 

 nor discussed without the aid of physical 

 terminology. 



It is easier to preach than to practice; 

 to say what should be done than to do it. 

 Between the theoretical statement of a 

 problem and the practical method by Avhich 

 it may be solved there is a profound gulf, 

 over which a direct passage is perhaps im- 

 possible. No reaction has yet been ex- 

 haustively studied on the lines which I have 

 laid down, and possibly none ever will be, 

 for the difficulties in the way of such a re- 

 search are almost insuperable. Of all the 

 snares which nature sets before our unwary 

 feet, that of apparent simplicity is the most 

 deceptive. Honest complexity, evident at 

 sight, we may hope to overcome; it is the 

 unseen obstacle which baffles us. In the 

 present instance a prime difficulty is the 



definition, the isolation of a reaction by 

 itself, apart from other chemical changes. 

 Nearly every reaction which we can ob- 

 serve is, in reality, a complex of several re- 

 actions—a series of steps, some of which 

 may easily escape our notice. "We measure 

 certain phenomena, only to find at last that 

 our result is an algebraic sum, and that we 

 have more unknown quantities than equa- 

 tions. We cannot solve our problem until 

 these factors have been recognized and 

 separated. 



To study individual reactions, then, ex- 

 cept for the determination of definite, spe- 

 cial phases, is not the best mode of proced- 

 ure; chemistry would advance but slowly 

 were we restricted to such a method. In 

 ordinary chemical research, in the work 

 of the compound-maker, for example, the 

 initial and final stages of a series of reac- 

 tions are investigated, and in that way 

 valuable data are obtained. But the aim 

 of science is not so much to amass facts as 

 to connect them by laws and principles; 

 and the more general the latter become, the 

 greater is their intellectual value. We can 

 not build, of course, until we have the ma- 

 terials, but between brick-making and 

 architecture the difference is great indeed. 



Leaving now the apparently simple, and 

 turning to the visibly complex, let us see 

 whether we cannot attack all reactions col- 

 lectively, and in that way reach a more 

 general statement of our real experimental 

 problems. All reactions display the same 

 fundamental phenomena, namely, changes 

 of composition, changes of properties and 

 transformations of energy; if we can 

 classify our data under these categories, we 

 shall begin to see more clearly the road we 

 are to follow. 



Now, recurring for a moment to the 

 analysis of a single reaction, we may con- 

 sider its two statical terms, the nature of 

 the substances with which we begin and 

 end. In any particular instance these ques- 



