October 6, 1910] 



NATURE 



425 



chapter a number of " special analyses " are con- 

 sidered ; they relate to methods of testing beetroot 

 seed, to the examination of various materials used in 

 sugar manufacture, and to particular cases of sugar 

 nnalvsis — e.g. the determination of raffinose. 



The chief criticism sugfjested, on looking through 

 the volume, is that a disproportionate amount of space 

 is allotted to preliminary and incidental matters. 

 Apart from an appendix of tabular material and the 

 index, the book contains 294 pages. Of these 

 only 120 are devoted to the real object of the work — 

 beet-sugar making and its chemical control. The 

 rest is taken up with accessory description, much of 

 which is merely general elementary chemistry. This 

 would be much better learned from an ordinary text- 

 book. A " theory of the origin of limestone " (p. 40) ; 

 a description of the metallurgy of iron (p 48), or the 

 chemistry of lead (p. 54) ; a dissertation on the mole- 

 cular structure of the hydrocarbons (pp. 61-2) : all 

 these are rather out of place in a book devoted to 

 sugar ; or, at least, such things should not get the 

 lion's share of the space. Whilst it is legitimate 

 enough to discuss the general chemistry of the sugars, 

 and even perhaps the theory of the polariscope, the 

 rest of the matter in question gives one the impres- 

 sion of being largely "padding." 



This apart, the book deals lucidly with the every- 

 day problems of beet-sugar production, and should 

 prove very useful to those for whom it is written. 



C. S. 



METHODS OF ROCK-AN.^LYSIS. 

 Analyse der Silikat- tmd Karbonatgesteine. By W. F. 

 Hillebrand; translated by E. Wilke-Dorfurt. Zweite 

 Auflage. Pp. xvi + 258. (Leipzig: W. Engelmann, 

 1910.) Price 6 marks. 

 The Analysis of Silicate and Carbonate Rocks. A 

 revision of Bulletin 305. By VV. F. Hillebrand. 

 Bull. 422, U.S. Geol. .Survey. Pp. 239. (Washing- 

 ton : 1910.) 

 AS a consequence of the modern developments of 

 petrology, accurate chemical analyses of rocks, 

 and of the component minerals of rocks, have become 

 more than ever an urgent desideratum; and it is a 

 fortunate coincidence that there has been at the same 

 time a decided revival of mineral chemistry, so long 

 overshadowed by that of the carbon compounds. Not 

 only is the discovery of new rock-types continually 

 providing fresh material, but also it has to be recog- 

 nised that the older rock-analyses, admirable in their 

 own time, no longer sufiice for the requirements of 

 the present day. The best modern analyses have the 

 advantage of greatly improved methods of separation ; 

 and, further, they aim at a much greater thoroughness 

 of treatment, often including estimations of twenty or 

 more constituents, instead of the eight or nine which 

 satisfied Bunsen and Delesse. 



In this work a leading part has been taken by the 

 American chemists, and particularly those of the 

 United States Geological .Survey. In the last thirteen 

 years four bulletins have been issued giving complete 

 analyses of many American rocks, conducted in the 

 NO. 2136, VOL. 84] 



laboratory at Washington. In the first of these, some 

 forty pages were devoted to a discussion of methods of 

 analysis by Dr. Hillebrand. This part, considerably 

 enlarged, was issued as a separate bulletin in 1900, 

 a new edition appearing in 1907 ; and it is this last 

 which has now been translated into German, with 

 some revision and additions by the author. It has 

 been closelv followed by a revised edition of the 

 original, so that the latest advances in this branch 

 of practical chemistry are now accessible equally to 

 German and English readers. 



The methods which are here fully and clearly set 

 forth are, in the main, those which the experience of 

 the author and his colleagues has led him to prefer ; 

 but alternative methods are often given, especially 

 when the first one requires complicated and costly 

 apparatus. Many of the analytical methods described 

 are, of course, familiar to the working chemist, but 

 the author's matured judgment on their relative 

 merits cannot fail to be of use; and, even apart from 

 this critical discussion, it is a great convenience to 

 have the scattered literature of the subject brought 

 together and presented in systematic shape. 



As an illustration of the author's treatment, we 

 may select the estimation of ferrous iron, always a 

 crux in rock-analyses (pp 154-71). First comes a 

 section, added in the present edition, pointing out 

 the important error introduced by oxidation of the 

 material during the process of grinding, and the de- 

 vices by which this error may at least be minimised. 

 This is followed by a comparison between Mitscher- 

 iich's sealed tube method of estimation and the hydro- 

 fluoric acid methods ; the former is in general to be 

 avoided, on account of the reducing action of sulphides 

 present in the rock. Since, however, Mitscherlich's 

 method is probably the best in those cases where it 

 can safely be used, it is described, with important 

 modifications suggested by experience. The general 

 principle of the hydrofluoric acid method is then set 

 forth, with a discussion of the chief sources of error 

 and of the influence of sulphides, vanadium, and car- 

 bonaceous matter on the determination of the ferrous 

 iron. Finally, the author describes the method itself 

 in its various modifications, as advocated by Cooke, 

 Pratt, and Treadwell, respectively. 



Twenty years ago the petrologist who did not per- 

 form his own chemical analyses felt that he was dele- 

 gating part of his legitimate task to another. A 

 more exacting standard has made a division of labour, 

 as regards complete rock-analyses, almost inevitable, 

 and Dr. Hillebrand's manual is accordingly addressed 

 to the chemist rather than the petrologist. On the other 

 hand, few of us are in the advantageous position of 

 the United States Geological Survey, which can com- 

 mand the services of six or eight skilled specialists ; 

 and it is also to be remembered that one complete 

 analysis demands as much time and labour as, per- 

 haps, three of a less ambitious kind. Some petrol- 

 ogists will be of opinion that there is still a place for 

 rock-analyses, conducted according to the best 

 methods, but including estimations of only a moderate 

 number of constituents. 



If the petrologist cannot make his own analyses, he 



