172 



SCIENCE. 



[N. S. Vol. XV. No. 370. 



are of peculiar value in certain lines of 

 physico-chemical investigations; and they 

 may also be profitably studied along the 

 vague boundary which separates organic 

 from inorganic chemistry. What we may 

 call the contact phenomena between any 

 two departments of knowledge are always 

 interesting. 



In the present revival of inorganic chem- 

 istry, a limited number of subjects have re- 

 ceived the most attention. Among them I 

 may name the study of double salts, of the 

 rare earths and of complex acids and 

 bases. All this work is of value; some of 

 it is fundamental; but more urgent, prob- 

 ably, is a revision of the older data con- 

 cerning much simpler bodies. This task is 

 not attractive; it is far from brilliant in 

 character and promises no startling dis- 

 coveries ; but it is none the less essential if 

 we wish to establish the foundations of 

 chemistry more securely. Consider any 

 group of inorganic compounds, as, for ex- 

 ample, the anhydroi;s metallic halides, and 

 we soon find that our knowledge of them is 

 full of gaps, and that the desci'iptions of 

 many presumably well-known substances 

 are wretchedly incomplete and defective. 

 To remedy this condition of affairs is no 

 small matter; there are errors to eliminate 

 and careless work to be done over; but 

 with modern resources a great improve- 

 ment is possible. Now, thanks to physical 

 chemistry, we can determine molecular 

 weights, either by cryoscopic or ebullio- 

 scopic methods ; and in the periodic law we 

 have a basis for scientific classification. 

 With these aids to research the new data 

 should assume a theoretical value which 

 formerly was lacking. For instance, the 

 structural side of inorganic chemistry has 

 been wofully defective; but now, knowing 

 the molecular weights of substances, prob- 

 lems of structure may be attacked to ad- 

 vantage. The conception of valency can 

 thus be tested to the uttermost degree. 



Underlying all work upon compounds, 

 however, is the study of the elements them- 

 selves. We may speculate as to their ulti- 

 mate nature, or we may condemn specula- 

 tion as useless; but we must agree that 

 accurate knowledge of their relations and 

 properties is most desirable, and especially 

 so with respect to physico-chemical re- 

 searches. In order to correlate the proper- 

 ties of compounds with those of their 

 components, we must fii'st determine 

 the latter, and our present knowledge 

 in this direction is exceedingly incom- 

 plete. Not one element is thoroughly 

 known on the physical side, and some, in- 

 deed, have not as yet been definitely iso- 

 lated. What we require is the exact meas- 

 urement of all the physical properties of 

 all the chemical elements at aU available 

 temperatures ; from such data laws are sure 

 to follow. Here again the periodic law can 

 guide us; for in its curves the measured 

 constants are easiest compared. In this 

 scheme, evidently, the accurate determin- 

 ation of atomic weights is an important 

 feature, for with them all else is coordi- 

 nated. We also need to know, more com- 

 pletely than we do at present, the molec- 

 ular weights of the free elements, because 

 the reactions which we really observe are 

 between molecules and not between atoms. 

 Thus, when monatomic mercury unites 

 with octatomic sulphur, the phenomena 

 which occur involve the breaking down of 

 the sulphur molecule. If, instead of mer- 

 cury, we have diatomic oxygen or tetra- 

 tomic arsenic, the reaction with sulphur 

 becomes still more complex, for in each 

 case, before combination, two molecules 

 must be dissociated. The dissociation, of 

 course, implies a loss of energy, of un- 

 known amount ; and in tliermochemical dis- 

 cussions this undetermined factor is the 

 cliief obstacle to progress. If we could 

 study reactions between monatomic mole- 

 cules alone, we should have ideally the 



