GE;OPHYSICAr, LABORATORY. Id 



of equilibrium of the mineral types thus far chosen for investigation, leaving 

 their application to natural rocks to the geologist whenever sufficient data 

 shall have been gathered to serve the purpose effectively.* In the case of 

 quartz, direct application was made of these data to a considerable collection 

 of typical specimens of natural quartz from known localities with most inter- 

 esting results, illustrative of the possibilities of more extensive studies of 

 this kind, using all the temperature data available from the common 

 minerals. A brief review of this paper follows (Quartz as a geologic ther- 

 mometer).! The application to natural rocks will become even more con- 

 clusive and productive when a greater body of such data has been gathered 

 and when the role of water in rock formation has become more thoroughly 

 understood, but there would appear to be no longer any doubt of the ulti- 

 mate practicability of establishing such formation temperatures with great 

 certainty. 



TEMPERATURE STANDARDS. 



The research on the absolute measurement of high temperatures with the 

 gas-thermometer, of which the preliminar}^ results were published a year 

 ago, J has been successfully continued throughout the year and is expected to 

 be completed about January i, 1910. The most important features which 

 have developed since the publication referred to are perhaps these: 



(i) An alloy of rhodium and platinum has been substituted for platin- 

 iridium as the bulb material, which reduces the contamination of thermo- 

 elements necessarily exposed in the furnace with the bulb to about one-fifth 

 of its former magnitude. 



(2) New determinations have been made of the expansion coefficient of 

 the new bulb material (80 parts platinum, 20 parts rhodium). 



(3) A new source of error has been discovered in the radiation of heat 

 from the ends of the bulb toward the cold ends of the furnace which operates 

 to raise the published temperatures of a year ago about 1.5° C. The error 

 is readily eliminated. 



(4) Observations with the gas-thermometer have now been successfully 

 continued beyond the limit of a year ago (1100°) to the melting-point of 

 nickel (1452°) with the same order of accuracy (copper melting-point, 

 1082 ± 1° ; nickel melting-point, 1452 ±2°). 



(5) Absolute determinations of the temperature of melting copper with 

 the new bulb (platin-rhodium) agree with those made with the old bulb 

 (platin-iridium) within 0.5°, although the form of the bulb, the furnace coil, 



* G. F. Becker, Harald Johannsen, and others have already made considerable use of 

 data of this kind in the discussion of important geologic problems. 



t Also the binary systems of alumina with silica, lime, and magnesia. Am. Journ. 

 Sci. (4), vol. 28, 322, 1909. 



I Some new measurements with the gas-thermometer. Am. Journ. Sci. (4), vol. 26, 

 40s, 1908. 



