>9Q 



NATURE 



[May 22, 1913 



student wishing; to follow up a particular subject 

 can find references to what has been already 

 published in relation to it. To facilitate this kind 

 of reference is, in fact, the whole purpose of the 

 book. Some degree of arrangement and classifica- 

 tion of contents is needful to make the index 

 usable at all, and this begins when one science is 

 marked off from another ; but the more minutely 

 the classification is carried out, and the smaller 

 the resulting classes become, the more chance there 

 is of uncertainty as to the class in which a parti- 

 cular paper should be placed. If the question is 

 answered in one way by the compilers of the index 

 and in another way by a person who wishes to 

 use it, the natural consequence is that he does not 

 find the information he wants in the place he 

 turns to first. The fact is that any possible classi- 

 fication is in a great degree arbitrary and con- 

 ventional. The grouping of scientific results that 

 at any time seems most natural and logical inevit- 

 ably reflects not only the then existing state of 

 knowledge, but also the successive stages by which 

 that state of knowledge has been reached. New 

 additions to scientific knowledge are not like bricks 

 added to a building each of which occupies a fixed 

 position and a sharply defined space ; each newly 

 recognised fact sheds light on what was known 

 before and may greatly alter the apparent relative 

 importance of previous acquisitions. 



That such considerations are not irrelevant to 

 the arrangement of this index is shown bv the 

 entry of thirteen references to papers on heat 

 developed on moistening solids, under the 

 general heading 0300 Capillarity, and also under 



Phenomena of Change of State, 1800 General. 



Th^re are, in fact, hosts of phenomena which are 

 essentially related to more than one division of 

 science, and papers dealing with them must neces- 

 sarily \be entered under more than one heading 

 unless 'the index is to be encumbered by a tangle 

 of cross-references. 



We are fully conscious that the Royal Society's 

 committee and the compilers of the index, who 

 have considered the matter as a whole, may have 

 good reasons for deciding on subdivisions and 

 schemes of arrangement the advantages of which 

 are not at once evident to anyone who has only 

 partially examined a part of their work. It is 

 therefore with the greatest diffidence that we 

 venture to raise the question whether the classifica- 

 tion on which the arrangement of the index is 

 founded is not in some cases too minute. Thus 

 the first entry under the heading 2410, Mechanical 

 Equivalent of Heat, gives a reference to Joule's 

 classical paper in the Philosophical Transactions 

 for 1850, and lower down, under the same general 

 heading, we find references to Rowland's deter- 

 NO. 2273, VOL. 91] 



mination and to Reynolds and Moorby's; but 

 Joule's final measurement (Phil. Trans., 1879) is 

 given under a separate sub-heading, Determina- 

 tion of Mechanical Equivalent, under which we 



also find Griffiths (Phil. Trans., 1894), Miculescu 

 (1892), and many others, and, under a sub-sub- 

 heading, "Electrical Method," Joule's determina- 

 tion of 1867. We do not doubt that there are 

 intelligible reasons for the separations and collo- 

 cations of which these are examples, but we con- 

 fess that to us personally they are more bewilder- 

 ing than helpful. For a long time we were not 

 able to find any reference to Schuster and Gannon's 

 measurement by the "electrical method," but at 

 last we discovered it, as well as Rowland's and 

 various other determinations, under " Specific Heat 

 of Water." This is quite an appropriate place, 

 but it is not easy to see why this paper should 

 not also have been entered among determinations 

 of the mechanical equivalent. 



It is no doubt in consequence of our not having 

 mastered the classification adopted by the com- 

 mittee that we have not been able to find refer- 

 ences to such historically important investigations 

 as those of Dulong and Petit into the expansion 

 of mercury, the laws of cooling, and the specific 

 heats of metals. The real difficulty of finding a 

 thoroughly satisfactory system of arrangement 

 arises partly from the enormous mass of material 

 to be dealt with, but still more from the extreme 

 complexity of the material. A strictly alphabetical 

 arrangement offers an alluring simplicity, but a 

 very slight examination of the contents of this 

 volume must convince anyone that it would be 

 hopeless to apply it until the matter has under- 

 gone a preliminary process of arrangement and 

 sifting. The only questions that can arise are as 

 to how this process shall be conducted, and how 

 far it shall be carried ; and probably scarcely any 

 two men would answer these questions in exactly 

 the same way. 



There is no question that the index is a very 

 remarkable and admirable piece of work, on which 

 the Royal Society's committee, the director, Dr. 

 McLeod, and all his colleagues deserve to be 

 heartily congratulated. It will not only be of 

 immense service to those engaged in the study of 

 special questions, but it will help to keep alive a 

 knowledge of the work of the men who laid the 

 foundations of physical science. This work, like 

 foundations generally, is apt to be buried out of 

 sight as the superstructure rises, but it is well 

 that modern builders should cherish the memory 

 of those who made their work possible. 



We find it difficult to close this volume without 

 comparing it with Dr. Thomas Young's "Cata- 

 logue of Works relating to Natural Philosophy and 



