294 GEOLOGY OF THE COMSTOCK LODE. 



lated as desired and kept constant during the whole time of experimentation. 

 Oil could be supplied without interfering with the flames. Trimming of 

 wicks was seldom necessary, and, there being four flames, gave rise to no 

 serious disturbance. To diminish the heat lost by I'adiation as much as 

 possible, the whole apparatus, with the exception of the bottom, was cov- 

 ered to a thickness of about three-quarters of an inch with cotton batting, 

 wrapped in layers and surrounded externally by heavy paper. Finally, 

 the water lost by evaporation was replaced drop by drop by means of a 

 pneumatic arrangement placed upon the boiler, but not shown in the 

 figure. The number of drops fed in a given time was so regulated by 

 the aid of a small faucet as to keep the level I I of the water in the boiler, 

 as indicated by the gauge m n, approximately at a constant height. The 

 ebullition was not allowed to become sufficiently intense to produce an 

 increase of pressure in the interior. 



To recapitulate: By the aid of a fairly constant source of heat, the 

 ebullition from a water level of constant height coixld be maintained at a 

 nearly constant intensity. It was believed, therefore, that a .stationary ther- 

 mal condition would soon set in and continue indefinitely. Errors due to 

 fluctuation of the barometric column, this being as likely to produce posi- 

 tive as negative eff"ects, could be excluded by proper methods of reduction. 



Thermo-eiement. — To mcasurc the Small increments of temperature, a thermo- 

 pile composed of three bismuth-silver elements was first used. Though 

 this acted well, there was danger, in consequence of the amount of sulphur 

 in the rocks (Fe S^), of complete destruction of the silver terminals during the 

 course of the experiment. This metal was therefore discarded, and platinum, 

 which is not thus affected, chosen in its stead. The bismuth was cast in the 

 shape of three adjacent sides of a rectangle, the length and width chosen being- 

 such as to allow the two ends to occupy the positions x and z shown in Fig. 

 19. Of course care was taken to insulate the whole thoroughly from the 

 walls of the boiler, this being accomplished by surrounding the element on 

 all sides by strips of thick sheet rubber. The parts of the element were 

 kept from touching each other by pieces of glass tubing suitably placed. 

 The terminals — which, to prevent confusion, are not indicated in the figure — 

 nere themselves insulated by a covering of rubber hose of small caliber. 



