SOSMAN AND HOSTETTERI A VACUUM FURNACE 281 



connected this arm across to the vacuum apparatus by a wide 

 glass tube, and evacuated and sealed the open arm, thus making 

 a closed U-tube manometer of the apparatus. The three gages, 

 compared against one another at various pressures within their 

 ranges, agreed within the error of their readings. We have in 

 this combination of gages a range of pressure measurement from 

 0.000 001 mm. mercury up to about 2.5 atmospheres. 



The furnace jacket is an inverted iron pot, closed at the bottom 

 by a disk of Alberene stone through which pass (1) the two 

 conductors in parallel which carry the current to the lower end 

 of the platinrhodium tube, together with the tubes for the cool- 

 ing water, and (2) the steel extension of the platinrhodium tube. 

 The latter is surrounded by a water-jacketed steel tube and cap. 

 A little mercury is placed in the bottom of the steel tube to insure 

 good thermal connection with the furnace tube, so that the lower 

 end of the latter will always be kept cold. The joint between 

 the stone base and the flange of the jacket is made tight by 

 means of a plastic cement made bj^ Mr. J. Jost of the instrument 

 shop of this Laboratory. 



The jacket is evacuated independently of the inside vacuum 

 by means of a May-Nelson pump. There is no connection 

 between the inside high vacuum and the outside ''ordinary" 

 vacuum. No insulation or refractory material is used excepting 

 a magnesia tube surrounding the platinrhodium tube. The out- 

 side vacuum removes any danger of collapsing the hot furnace 

 tube at high temperatures, and also provides an efficient heat 

 insulation. The furnace jacket is cooled by a coil of lead pipe 

 on the outside carrying cold water. An observation window 

 permits optical temperatm*e measurements to be made from 

 without as a check on the thermoelectric measurements within 

 the tube. 



The magnesia tube does not greatly affect the loss of heat 

 from the furnace tube at the lower temperatures, where the 

 loss by radiation is not great, but does have a noticeable effect 

 on the efficiency at higher temperatures. For instance, 386 

 amperes raised the temperature to 790° without the tube, and 

 to 809° with the tube in place. 510 amperes, however, which 



