Wright and Larsen — Quartz as a Geologic Thermometer. 433 



The abrupt change in expansion-coefficient observed by Le 

 Chatelier at the inversion temperature causes thick plates of 

 quartz to shatter more or less completely, the planes of fracture 

 running parallel with the unit rhombohedral faces ordinarily. 

 This shattering is so characteristic that it is practically 

 impossible to bring a thick quartz plate through the inversion 

 temperature, either up or down, without some cracking, usually 

 sufficient to spoil the plate for further use. Thin plates, on 

 the other hand, bend noticeably. Hiigge* observed the tem- 

 porary warping of a thin plate of quartz r03 mm thick) through 

 an angle of 3° at the inversion temperature. Above and below 

 the inversion temperature, however, the plate appeared per- 

 fectlv plane, thus indicating practical identity of the axial 

 ratios of the a and ft forms. 



The above determinations of the inversion temperature were 

 made with a direct reading Siemens and Halske millivoltmeter, 

 the scale divisions of which registered temperature intervals 

 of 10°, 0*1 division being equivalent to 1° and the absolute 

 accuracy perhaps ±5°. To standardize these readings and at 

 the same time to test the sensitiveness of this inversion in 

 quartz (which is exceedingly inert in its change to tridymite 

 and also to the viscous amorphous state), a quartz plate about 

 •2S lum in thickness and parallel with the principal axis was 

 selected and its birefringence bands in the Babinet compensator 

 observed, while the readings of the thermoelement on which 

 the plate rested were recorded on a potentiometer.! This system 

 wa^ sensitive to temperature differences of perhaps '02° and 

 its absolute accuracy was well within ± 0*5°. The point of 

 change was indicated in the microscope by a sharp movement 

 of the dark interference band of the Babinet compensator, and 

 at this instant the reading on the potentiometer was recorded. 

 The inversion, both on heating and on cooling, was observed 

 a number of times and the temperature of inversion found to 

 be remarkably constant and sharply marked. The readings are 

 included in Table Y, columns I-IY. The greatest difference 

 in the observed temperatures of the inversion during heating- 

 is 2 microvolts, or about 0*2 C , while that for cooling is some- 

 what greater. Although the thermoelement readings were 

 expressed in microvolts, the equivalent temperatures in degrees 

 are listed in Table V". It was found that notwithstanding 

 differences in rate of heating, the inversion point on heating- 

 was practically constant, whereas the reversion point on cooling 



*Loc. cit. 



f The methods here adopted for the testing and standardizing of the ther- 

 moelements used are due to Dr. A. L. Day, and to him the writers are fur- 

 ther indebted for all thermoelectric measurements recorded in the following 

 paragraphs. 



