June 4, 1908] 



NATURE 



90 



Berthelot, Lord Kayleigh, Leduc, Guye and his 

 co-workers, from the study of gases have been able 

 by a nice combination of exact theory and experiment 

 to bring independent evidence as to the molecular 

 weight of gases ; at the same time, Morley, T. W. 

 Richards, and others at Harvard have considerably 

 increased the accuracy with which the more important 

 atomic, or rather combining, weights are known. 

 Prof. Young's own researches have been closely related 

 to the former investigations. 



The problem in the deduction of accurate atomic and 

 molecular weights from the properties of gases is the 

 precise application of Avogadro's hypothesis; that is, it 

 is necessary to know the relative volumes (at o° and 

 one atmosphere) of the gas under consideration and 

 o.xygen, which contain equal numbers of molecules. 

 When these volumes are known, the weight of the 

 molecule of the gas can be found at once from its 

 density relative to that of oxygen. Berthelot assumes 

 that Avogadro's hypothesis is strictly true when gases 

 are at small pressures; to apply this assumption, pv 

 has been observed at a small pressure and at one 

 atmosphere for several gases, including oxygen. Guye 

 uses the principle of corresponding states; for ex- 

 ample, argon and oxygen have appro.ximately the same 

 critical pressure and temperature, so he assumes that 

 equal volumes of these two gases (at o° and one 

 atmosphere) have the same number of molecules; and 

 finally, the values of a and h in van der Waals's equa- 

 tion have been used directly io find the Avogadro 

 volumes. The atomic weight of nitrogen obtained in 

 this wa}' is i4'oi, as opposed to the formerly accepted 

 i4'o4; there is plenty of evidence that the smaller value 

 is the more accurate one. 



As was to be expected, the treatment of change of 

 state, van der Waals's theory, the vaporisation of 

 mixed liquids, &c. (where so much of the best work is 

 that of Prof. Young himself), is at once clear, precise, 

 and interesting. 



The first chapter, of twelv? pages, on " The Funda- 

 mental Laws of Chemical Combination," which also 

 includes Dalton's atomic theory, seems to us entirely 

 unsatisfactory, and falls much below the standard of 

 the rest of the book. These fundamental subjects re- 

 ceive the usual inadequate treatment which mars so 

 many elementary text-books of chemistry. The defini- 

 tion of an element, as a substance not decomposable at 

 ■will, is artificial, and merely avoids the difficulties 

 raised by the well-verified spontaneous change of 

 radium into helium. Our complaint, however, is with 

 th; presentment of the laws of chemical combination. 

 When a generalisation is raised to the dignitv of being 

 called a law, surely the value of science as a method of 

 thought demands, (i) the definition of the law in clear 

 and precise language, (2) a statement of the observ- 

 ations (with their accuracy) upon which the law is 

 based? None of these is done for the law of definite 

 proportions. It is high time the " law " of multiple 

 proportions was omitted from text-books. The state- 

 ment of it criticised ends with the words " vary in the 

 dift'erent compounds according to very simple numer- 

 ical proportions." Consider C,-„H,,„ and CgHj^ {i.e. 

 CjjHjjj); can the ratio 122 noj be called " very simple 



NO. 2014, VOL. 78] 



numericallv ".-' Few pairs of compounds obey the 

 " law." Though Dalton's atomic theory was only 

 accepted bv chemists generally after half a century of 

 controversy, and is at present rejected by a few, in 

 the two pages devoted to it in " Stoichiometry " the 

 difficulties of the theory are not even mentioned. It is 

 to be hoped that these blemishes may be removed in 

 a future edition from a book which gives such an 

 up-to-date and adequate account of a large part of 

 l>h\sical chemistry, and is one of the volumes most 

 needed of a valuable series. T. H. L. 



NEWTON'S PHILOSOPHY. 



La Philosophie de Neivtoii. By Dr. L6on Bloch. 

 Pp. 642. (Paris : F. Alcan, 1908.) Price 10 francs. 



THE subject of this book is Newton's " phil- 

 osophy " in the large sense in which Newton 

 himself understood that word, not in the narrower 

 sense which is now usual. The author passes in 

 review practicallv the whole of Newton's contribu- 

 tions to science, giving in each case their antecedents, 

 their method, and their outcome. His historical 

 accounts of previous discoveries, with the consequent 

 estimates of Newton's contribution, are usually e.\- 

 cellent, and in his exposition of Newton's ideas he is 

 in general very faithful to his original. 



M. Bloch 's successive chapters, dealing with different 

 parts of Newton's work, have a certain similarity of 

 structure. They generally begin with Descartes, and 

 show the element of arbitrary hypothesis in his views. 

 Then, after some account of intermediate writers, they 

 point out how Newton proceeded by the right inductive 

 methods, collecting his laws and definitions from 

 facts, and verifying their consequences by experi- 

 ments. The hypotheses which he objected to, it is 

 said, were not hypotheses used as such, but hypotheses 

 used as though they were known to be true. It was 

 still customary to object to a new theory, based on 

 observation and experiment, not that it failed to ex- 

 plain the facts, but that it contradicted the maxims 

 of the illustrious So-and-So. This attitude seems 

 strange to us, because it has so completely died out 

 in science. But it still survives in philosophy, where 

 emphatic assertion is one of the accepted methods of 

 proof; and from this analogy we can understand 

 Newton's attitude and the progress it involved. M. 

 Bloch 's scheme involves some unnecessary repetitions, 

 and one gets a little tired of the merits of induction. 

 But substantially what he says on this subject seems 

 just. While attributing to Bacon a great influence 

 in forming Newton's ideas of method, he explains the 

 two respects in which Newton surpassed Bacon's pre- 

 cepts, namely, that his methods were quantitative, 

 and that he realised the part which deduction plays 

 in induction. 



\\'here the book is least satisfactory is on the side 

 of logical analysis. Thus in regard to fluxions, he 

 points out, very justly, how Newton conceived a 

 fluxion physically, and how he often inferred the ex- 

 istence of limits, in problems where his mathematical 

 apparatus was insufficient to prove it, from the fact 

 that he was dealing with physical problems which 



