382 



ROMBERG. 



soldered into saw-cuts in the copper plug which joins the two parts of 

 the case. Compensating leads, carr^ang about | ohm of platinum, are 

 provided. Figure 4 shows the thermometer somewhat distorted, the 

 upper portion being drawn too short. These thermometers show 

 a temperature rise of .02 degrees with the measuring current of 3 

 milliamperes. 



The thermostats, which contain about 6 cubic feet of water each, 

 are so lagged and heated that they would attain without automatic 

 regulation a temperature too low by perhaps a degree. Electric heat- 

 ing by bare-wire helixes with alternating current at 30 volts, controlled 

 by a toluol regulator, supplies the further heat needed. The tempera- 

 ture is thereby kept fluctuating about a mean, and the variations, of 

 perhaps .005°, are so smoothed out by the heat capacity and imperfect 

 conduction of the streams in the coils, as to keep these streams well 

 within .001° of the desired temperature. 



n^.4 



Copper ieavCi 

 Silver .5ol<^e red '"fo J"a*^ """ 

 "7 Cofj^^w P/u<j 



The regulators consist of about 60 feet of \" copper tubing, closed 

 at one end, the other terminating in the top of a piece of half-inch iron 

 pipe 6" long, capped at both ends. Through this top extends also a 

 steel tube \" diam., to near the bottom of the pipe, where it is amal- 

 gamated with mercury. All joints are soldered. Mercury fills the 

 lower portion of the pipe; toluol the rest and all of the copper tube. 

 "When the toluol expands, mercury is forced up the steel tube into a 

 glass capillary attached to its upper end, and there makes an electric 

 contact which operates a relay and circuit-breaker, stopping the heat- 

 ing current or introducing resistance. The purpose of amalgamating 

 the end of the steel tube is to keep the toluol from creeping. The 

 success of these regulators, whose task is made severe by the flow of 

 unthermostated water, is attributed to the close thermal contact of the 

 toluol with the bath, to the large quantity of toluol used, giving ample 

 sensitiveness, and to the generous size of the bath. No contact 

 difficulties were experienced, and no drifting of temperatures noticed. 



Although the coils of the thermal coupling were made as symmetrical 

 as practicable, it was feared that when there was only a small quantity 



