Chemistry and Physics. 157 



various formulas have been proposed, and Daniour found the 

 composition variable in the same specimen. Professor J. D. 

 Dana* also suggests that a portion of the silica may be basic. 

 On the whole it appears to me that the constitution of this 

 mineral and of the allied milarite is insufficiently established, 

 and that it is not as yet necessary to assume for them an acid 

 not known to exist elsewhere. 



The hypothesis that the silicates can be reduced to two 

 series and their combinations would greatly simplify the dis- 

 cussion of eruptive magmas. This supposition would cover 

 nearly the same ground as Bunsen's famous suggestion of a 

 normal pyroxenic and a normal trachytic magma, afterwards 

 more or less differently developed by Durocher and others. 

 Magmas would then be composed of substitution products 

 (acid, normal or basic) of polysilicic and orthosilicic acids. 

 U. S. Geol. Survey, San Francisco, March, 1889. 



SCIENTIFIC INTELLIGENCE. 

 I. Chemistry and Physics. 



1. On the Vapor Density of Hydrogen fluoride. — Thorpe 

 and Hamblt have confirmed by further experiments the opinion 

 expressed by them a year or more ago that the process of break- 

 ing up by heat of the molecular grouping of hydrogen fluoride in 

 the state of vapor, is analogous to that of acetic acid vapor, and 

 that there is no evidence that a molecule corresponding to the 

 formula H 2 F 2 exists through any appreciable range of temperature. 

 The commercial aqueous acid was purified by adding to it a 

 strong solution of potassium permanganate and then potash, and 

 redistilling from a platinum retort. The acid potassium fluoride 

 was prepared either by adding the theoretical quantity of pure 

 potassium carbonate to the hydrofluoric acid in a platinum dish, 

 or by adding to the normal potassium fluoride the required quan- 

 tity of the acid. The residue after evaporation was fused, and 

 yielded on cooling a white opaque mass only slightly hygroscopic. 

 About 40 grams of this salt were placed in a platinnm-iridium 

 retort, fitted, by grinding, to a condensing tube and receiver made 

 of the same alloy. The condensing tube which was 50 cm. long 

 was immersed in a mixture of calcium chloride and ice, which 

 gave a temperature of —25°. The receiver was similarly cooled. 

 To determine the vapor density, a platinum vessel of known 

 weight and of a capacity of 288*6 c. c, was placed in a bath of 

 glycerol at the required temperature and connected with the 

 receiver containing the liquefied hydrogen fluoride. On surround- 

 ing this receiver with water at 25°, the liquid rapidly volatilized 



* System of Mineralogy, 5th Edition. 

 Am. Jour. Sci.— Third Series, Vol, XXXYIII, No. 224.— August, 1S89. 

 10 



