156 CARNEGIE INSTITUTION OF WASHINGTON. 



(26) Tlic cheini.strv of i)(jrtl:ui(l cciiH'nt. (1. A. Rankin. Concrcto (Cement Mill Section) 



S, ir,-20 (1910). 



Reprinted from "The chemistry of portland cement" (Address delivered at 

 the Twelfth Annual Convention of the American Concrete Institute, Chicago, 

 February 14-17, 1916). Keviewed under No. 25 above. 



(27) The chemistry of portland cement. G. A. Rankin. Cement Era, 14, 79-8^ (1916). 

 Reprinted from "The Chemistry of portland cement" (Address delivered at 



the Twelfth Annual Convention of the American Concrete Institute, Chicago, 

 February 14-17, 1916). Reviewed under No. 25 above. 



(28) PortlaiKl cement. G. A. Rankin. J. Franklin Inst., 747-784 (1916). 



An address given at the Frankhn Institute, Philadelphia, embodying most 

 of the results of the investigation of the relations between lime, alumina, and 

 silica in portland-cement clinker, which was first published under the title 

 "The constituents of portland-cement clinker" (J. Ind. Eng. Chem., 7, 466, 

 1915), and reviewed on page 165 of Year Book No. 14 (1915). Some interest- 

 ing applications of these conclusions to the development of portland cement 

 are pointed out. 



(29) Chemical analyses of igneous ro(;k!-:, published from 1884 to 1913, with a critical 



discussion of the character and use of analyses. H. S. Washington. U. S. 

 Geol. Survey, Professional Paper 99, 1,200 pages (1916j. 



This is a revision and expansion of Professional Paper 14, published in 1903. 

 The text is devoted to discussions of the importance of chemical analyses in 

 the study of igneous rocks; the general characters of rock analyses, referred 

 especially to their accuracy and completeness; the possible sources of error 

 involved in making them; the method adopted of "rating" them, that is, 

 determining their relative w^orth for use; remarks on several features of the 

 tables; and a list of the publications consulted, with some minor matters. 



The tables are divided into four parts: I, superior analyses of fresh rocks; II, 

 incomplete, but otherwise superior, analyses of fresh rocks; III, analyses of 

 weathered rocks and of tuffs; IV, inferior analyses. 



The total number of analyses amounts to about 9,500, while the earlier 

 edition embraced 2,881. The search for analyses throughout the literature 

 has been very thorough, chiefly based on the Library of the U. S. Geological 

 Survey, and it is believed that the collection is practically complete for those 

 which have been published between 1884 and 1913, both inclusive, with the 

 addition of many hitherto unpublished analyses contributed by petrographers 

 and official surveys of many countries. 



The analyses in Part 1 (which are all of superior quality and therefore the 

 most useful) number about 4,950. The "norm" of each has been calculated 

 by the author, and these analyses are arranged according to the quantitative 

 system of classification. The analyses in the other parts are arranged 

 according to the usual system. 



It may be added that all the original cards on which the analyses were 

 collected are preserved and on file in the Geophysical Laboratory. It is 

 purposed to arrange them geographically, and to use them, in conjunction 

 with the published work, in studies of the distribution of rocks over the earth, 

 the characters of petrographic provinces, and other subjects. 



(30) The common refiactorv oxides. Robert B. Sosman. J. Ind. Eng. Chem., 8, 985-990, 



(1916). 



Some of the properties of the individual oxides are compared and a graphical 



summary is given of all the possible two- and three-component systems of 



SiOz, AlzO.-i, ^IgO, CaO, FeO, and FezOa. The l^inary compounds are in most 



cases made up of the individual oxides in simple proportions, usually 1 : 1 or 



