GEOPHYSICAL LABORATORY. 91 



experimental demonstration of the fact that the order of crystallization is 

 determined by the composition of the magma and not by the melting tem- 

 perature of the individual crystals. While such a relation has been predicted 

 by physical chemists from general considerations, this is probably the first 

 time it has been experimentally shown for the more complex silicate magmas. 

 It is likely to prove of immense geological importance in its future appli- 

 cation. 



PORTLAND CEMENT. 



Inasmuch as the three oxides above referred to also form the basis of 

 Portland cement, it has proved possible for the first time in the history of 

 cement study to establish the true composition of cement clinker. It has. 

 been shown from what initial compositions the clinker will contain free 

 lime, and that clinker which does not contain free lime can crystallize to one 

 or the other of two different but definite mixtures, depending on the initial 

 concentration. This is also likely to prove of considerable importance, not 

 only in determining what components of Portland cement are essential and 

 what ones are merely accidental, but may also form a basis for establishing 

 standards of pure cement. It also furnishes the basis for a rational expla- 

 nation of the setting and hardening phenomena of both Portland and slag 

 cements ; it will assist in clearing up the many anomalies noted in the practi- 

 cal working of concretes ; and, finally, the completed diagram will explain the 

 fundamental nature of blast furnace slag and furnish a sound basis for a 

 study of its mode of action. 



In view of the number of measurements to be made where so many rela- 

 tions are involved, the present announcement is preliminary and does not 

 cover the whole ground. The work will be continued during the coming 

 year. 



SULPHIDES OF IRON. 



Partly in order to ascertain whether physico-chemical methods of analysis 

 would be of assistance in the study of ore deposition, and partly by reason of 

 the economic importance of the sulphides of iron, a laboratory study of 

 pyrite, marcasite, and pyrrhotite was begun nearly three years ago. But 

 here also a difficulty cropped out very early in the investigation, of precisely 

 the same character as the others referred to earlier in this report. Ana- 

 lytical chemistry was unable to provide methods for the determination of the 

 amount of sulphur contained in a given product which were correct within 

 2 per cent. The investigation was accordingly halted until a method could 

 be developed. This painstaking effort has now been rewarded, and is pub- 

 lished in two papers which are reviewed on pages 96 and 97. Determina- 

 tions of this sulphur can now be made which are accurate within 0.1 to 0.2 

 per cent. 



Continuing upon the original plan, we now find pyrite and marcasite to be 

 different crystal forms of a true compound, FeS 2 , the marcasite becoming 

 unstable and going over into pyrite at about 450 . The obvious inference is 



