280 SOSMAN AND HOSTETTERI A VACUUM FURNACE 



Branches from the glass tube connect with the Gaede molecular 

 vacuum pump on the one hand and with the pressure gages on 

 the other. These connections are made with wide bends ^to take 

 up stresses due to movements of the furnace and gages with 

 changing temperature. All the glass tube connections are large, 

 to facilitate rapid pumping to low pressures. The molecular 

 pump delivers into a Gaede oil box pump, with which it is con- 

 nected by a wide glass tube. The glass connection from the 

 furnace to the molecular pump is provided with a removable 

 conical ground cap (as shown in fig. 1), so that the pump can 

 be washed without interference with the rest of the apparatus. 

 A stopcock is also attached here, so that the pump can be used 

 for other work. A CaCl2 drying tube connected between the 

 molecular and box pumps allows the pumps to be filled with 

 dry air before stopping. The interior of the furnace and gages 

 is kept dry by a tube of P2O5 in the bottom of the steel extension 

 of the furnace tube. 



The charges are contained in a small platinum crucible sus- 

 pended by two platinum wires sealed into a glass cap, which is 

 attached to a ground glass conical joint at the top of the tube 

 ' above the furnace. The two wires of the platinum-platinrhodium 

 thermoelement are also sealed through this glass cap, and the 

 four wires are insulated from one another by transparent fused 

 silica or Marquardt porcelain capillaries. The two suspension- 

 wires may also be used as the leads for the determination of 

 melting points, et cetera, by the quenching method. 



There are three pressure gages: (1) a McLeod vacuum gage 

 of 500 cc. capacity from Leybold (shown at Mi), (2) a similar 

 gage of 50 cc. capacity made in this laboratory (Mo), (3) a mer- 

 cury manometer read by means of parallel knife edges and a 

 vernier, and reading to 0.01 mm. The latter is the manometer 

 used in previous work on the nitrogen thermometer.^ It has 

 an excellent scale, calibrated throughout its length to 0.01 mm. 

 by the Normal-Aichungs-Kommission in Charlottenburg. We 

 removed the fixed level point used with the gas thermometer, 



* Am. J. Sci. (4) 26: 415. 1908. Carnegie Inst. Washington, Publ. No. 157, 

 * p. 19. 



