356 TJic American Geologist. 
June, 189!! 
order which is characteristic of the work. Further, it appears 
as if the autlior had not been able to choose conventional col- 
ors for the different oxides. The same violet color, for instance, 
stands sometimes for JNIgO, sometimes for Al-().i, and a^^ain 
for CaO. In the text he indicates curves III, when he is 
speakino^ of curves \', etc. Such negjigence and disorder of 
the editorship of any book ought to be severely blamed. Life 
is short and sufficiently full of trouble even with the reading of 
necessary and useful things. Authors ought not to so loosely 
arrange their thoughts, and so carelessly express them as to 
cause the loss of time in deciphering them. Having said this, 
I will make the further remark that in the first three tables 
the zero of the ordinates is not the same for all the lines, ap- 
,)arently as if the author intended to indicate in the diagrams 
only a series of correlations and not the absolute values of the 
ordinates. 
The conclusions of the first chapter are resumed by the 
author himself in the French svnopsis joined to the book. 
The third chapter is also synopsized in French. Lewinson- 
Lessing says that the purpose of this article was expressed in 
the "Note on the classification and nomenclature of eruptive 
rocks" read to the fourth international congress of geologists 
at St. Petersburg. It will therefore be easy to know what the 
author believes on this subject. The synoptical table of rocks 
was also' presented at St. Petersburg. In this chapter Lewin- 
son-Lessing says the chemical composition of a rock is inde- 
pendent of every other consideration, while its mineralogical 
composition is conditioned. He considers the chemical com- 
position as an independent variable, and the mineralogical 
composition, structure, etc., as functions of that variable. Con- 
sequently, according to the author, the classification of eruptive 
rocks ought to be based at the outset, on the chemical com- 
position, and in accordance therewith should be constructed 
step by step the principal subdivisions. 
Critique by the rcvieiver. It would seem that, before mak- 
ing classifications of rocks based on their chemical composi- 
tion, it is necessary to have many more analyses. Instead of 
this the author has satisfied himself v^'ith making computations 
of the analyses of others. Among these are the analyses taken 
by Rosenbusch which were made, at least in part, on rock? 
