724 



SCIENCE, 



[N. S. Vol. VIII. No. 204. 



chemists, as chemists have followed in the 

 footsteps of physicists. 



The advance of the mathematical point 

 of view in chemistry has brought with it 

 an entirely new danger, as well as a new 

 field of power. The accuracy of a result, of 

 course, depends upon the accuracy of the 

 foundation upon which the reasoning rests ; 

 and accurate mathematical processes may 

 lead to wholly erroneous conclusions if they 

 are based upon incomplete data. In chem- 

 istry this cause of error is especially promi- 

 nent, because of the great complexity of 

 most of the phenomena and the fact that 

 they are often modified by subordinate in- 

 fluences. For this reason a physicist, used 

 to simple phenomena and less complex ef- 

 fects, is especially apt, when he deals with 

 problems allied to chemistry, to erect a 

 large superstructure of mathematical rea- 

 soning possessing the semblance of reality 

 upon a paper foundation, and be drearily 

 awakened some fine day by the collapse of 

 his air castle. The only mode of guarding 

 against this subtle cause of disaster is to 

 bring as much skill into every step of the ex- 

 perimentation as into the pure reasoning 

 based upon the supposed facts. Here the 

 physicist is seriously hampered by his lack 

 of knowledge of chemistry, as well as bj' his 

 usual repugnance to dealing with glass ves- 

 sels and liquids ; while, on the other hand, 

 the chemist is equally hampered by his in- 

 bred dislike of brass instruments and his 

 imperfect acquaintance with the manipula- 

 tion of his new sensitive tools. 



In short, one must be an accomplished 

 chemist, physicist and mathematician in 

 in order to attain the highest results in 

 modern theoretical chemistry, and the 

 number of men who have the time or ability 

 to acquire this threefold education can 

 never be large. All honor to van't Hoff, 

 Ostwald, Nernst and the others who come 

 nearest the high ideal ! While it is true, 

 however, that few men can hope to attain 



the highest, it does not follow that the rest 

 of us cannot be of great use. Each man 

 can be of value in his own particular sphere > 

 it is only necessary that he should work 

 faithfully with a single eye to the truth, 

 that he should be as free as possible from -^ 

 prejudice, and that his published work 

 should be as accurate as he can make it. A 

 well conducted organic synthesis, a few 

 carefully determined solubilities, will in the 

 end be more valuable to the progress of 

 science than a false generalization, no mat- 

 ter how ingenious the latter may be. But 

 how great is the responsibility of the collec- 

 tor of facts ! for if his observations are false 

 his work is of less value even than that of 

 the false theorist ; it has not even ingenuity 

 in its favor, and is worse than useless. Ost- 

 wald has more than once pointed to the re- 

 sponsibility attending publication, and we 

 should all do well to heed his warning. 



A comprehensive design which I had once 

 harbored of giving you a resume of the 

 year's work in physico-chemical research 

 throughout the world has been relinquished 

 because of the great number of small p9,pers 

 which could not be treated satisfactorily in 

 the brief space of an evening's talk. The 

 subject of stoichiometry, in Ostwald's rather 

 comprehensive interpretation of this word, 

 has received this year the attention usually 

 accorded to it. Solutions have still occu- 

 pied many able men, without having by 

 any means had their possibilities exhausted. 

 Van't Hoff's admirable little book upon 

 double salts has already begun to exercise 

 an effect upon the chemical world which 

 the scattered and less illuminating papers 

 of his students could not have been ex- 

 pected to exercise. Dr. Gibbs' interesting 

 address upon this important topic will un- 

 doubtedly excite further study on the whole 

 question of the so-called ' molecular com- 

 pounds,' which are so little understood and 

 so hard to reconcile with our only partially 

 satisfactory ideas of quantivalence. The 



