742 



NATURE 



[August ii, 192 i 



of mercurous nitrate. The succeeding chapters 

 describe the estimation in turn of the elements 

 which usually occur. The methods for the direct 

 combustion of carbon are comparatively slow, and 

 it would have been advantageous to add a de- 

 scription of the rapid methods, using small elec- 

 trically heated tubes, which were devised during 

 the war for the enormous number of shell steel 

 samples which had to be analysed in the Admir- 

 alty and other laboratories. In such rapid 

 methods soda lime is used with advantage in 

 place of the more cumbrous potash bulbs. In 

 the Volhard estimation of manganese the simple 

 method of precipitation with zinc oxide and titra- 

 tion without removal of the iron, which is very 

 convenient in the analysis of ferro-manganese, is 

 not mentioned. The estimation of sulphur and 

 phosphorus in steels, about which disputes are 

 most frequent, is treated very thoroughly. 



The analysis of ores, refractories, slags, fuels, 

 and boiler waters is dealt with in later chapters. 

 The section on slags suffers somewhat from its 

 brevity, and many chemists would welcome a 

 fuller account of this important subject. Thus in 

 the analysis of basic slags no mention is made of 

 the distinction between total and available phos- 

 phoric acid, on which the value of the slag so 

 largely depends, and it would also have been well 

 to include some account of the estimation of 

 fluorine in such slags ; the addition of fluorspar 

 in the basic open-hearth process is frequently 

 practised, and its effect is to convert a part of 

 the phosphoric acid into an inert form. 



Mr. Ibbotson's work may be confidently recom- 

 mended to the analyst and student as a trust- 

 worthy guide to the subject by an author of ripe 

 experience in the field in which he has worked so 

 long. 



Relativity and Gravitation. 



Relativity: The Electron Theory and Gravitation. 

 By E. Cunningham. Second edition. (Mono- 

 graphs on Physics.) Pp. vii-l-148. (London: 

 Longmans, Green, and Co., 192 1.) 105. 6d. 

 net. 



THE second edition of Mr. Cunningham's 

 book, like the first, aims at presenting the 

 problems of relativity in a form suitable for the 

 general physicist. More than half the book deals 

 with the special theory, giving the fullest account 

 of the experimental side in any English book. 

 This part is practically unchanged from the first 

 edition — too little changed, for one would have 

 NO. 2702, VOL. 107] 



liked to see the author's views on Majorana's ^ 

 experiments, which are not mentioned. I 



In discussing the general theory, he follows the ! 

 historical order of development, commencing \ 

 with Eotvos's experiment, which showed that the ; 

 weights of two bodies of different constitution in 

 the same gravitational field are proportional to 

 their inertias within 5 parts in lo^. From this he 

 advances by a series of generalisations. First, 

 light has inertia; if Eotvos's result is true for it, 

 it must also have weight. Therefore it cannot 

 travel in straight lines in a gravitational field. 

 Therefore the differential ds, which is intimately 

 related to the behaviour of light in the special 

 theory, must, if it is still to maintain this relation 

 to light, have a form in a gravitational field that 

 takes the field into account. It has also a rela- 

 tion to the motion of a particle in the special 

 theory ; we knew already that it would have to be 

 modified in form to maintain this in a gravita- 

 tional field. 



It is therefore assumed that the same form will 

 still answer both purposes. Previously, again^ 

 the law of gravitation satisfied a condition 

 that was unaltered by any displacement of 

 the origin or rotation of the axes. Sup- 

 pose, then, that the coefficients in the new ds 

 satisfy a condition that is unaltered by any change 

 in the co-ordinates used ; the class of changes 

 admitted is to be as wide as will permit some such 

 condition to be satisfied. This leads at once to 

 the irrelevance of the mesh .system, and appears 

 to the reviewer to be the best reason yet ad- 

 vanced for attributing to this principle any 

 appreciable prior probability. 



The crucial tests of the theory are described, 

 and a chapter is devoted to Weyl's theory of elec- 

 tric and magnetic forces. The book is well 

 arranged and written. Enough does not seem to 

 be made, however, of the crucial tests. For any- 

 thing that any professed exponent of the theory 

 has said, there might be a million other theories, 

 all as probable as Einstein's, which would give 

 the same predictions. It may be pointed out 

 that on p. 114 the assumptions given are 

 not enough to ensure that the coefficient of drdt 

 shall be zero, which is assumed a few lines later; 

 that in the footnote on p. 107 it is implied that 

 a purely imaginary quantity can have a true 

 minimum; and on p. 120 that the mere fact that 

 the resultant velocity of an object is known is not 

 enough to determine its path. But in the main 

 the book is a careful and sound analysis, and 

 can be recommended to all students of the theory. 



H. J.' 



