1912-13.] Determining Vapour Densities at High Temperatures. 7 
thermometer which had been calibrated at the Reichsanstalt. Immersed 
in a bath of sodium and potassium nitrates, the thermocouple and 
the thermometer gave readings which agreed to within one degree at 
400° C. 
The manometer was kept at a constant temperature of 580°d=5° C. in 
another furnace, which was heated by a spiral of platinum wire. 
The ends of the furnace which contained the bulb A were closed by 
means of tight-fitting asbestos plugs. This was found necessary to prevent 
the disturbing effect of draughts of air. The end of the manometer furnace, 
through which the ray of light from the illuminated slit had to pass, was 
closed by means of a polished plate of transparent quartz. 
Without this plate, irregularly heated vortices of air were set up at 
the mouth of the furnace, which, on account of their varying refractive 
index, caused the images of the slit to flicker, and made the reading of the 
zero point very difficult. 
With the plate in position the images were practically steady, and the 
zero point could be read with ease and certainty. 
A space of 6 cm. separated the ends of the two furnaces, and had to 
be traversed by part of the capillary B (fig. 3). This portion of B was 
wrapped in asbestos fibre and heated by means of a bunsen flame. A 
correction was made for the vapour in this part of B. Its temperature 
was estimated to be 800° C. An error of several hundred degrees in the 
estimation of this temperature would have no appreciable effect on the 
accuracy of the result. 
01030 gram of Kahlbaum’s purest mercury was weighed out in a small 
quartz tube and introduced into the bulb A (fig. 3). The whole apparatus 
was then evacuated by means of an automatic Topler mercury pump. 
When a vacuum of less than '01 mm. of mercury had been attained, the 
bulb A was sealed off at D with an oxygen-gas blowpipe. 
The Herseus electric furnace was then moved into the position which 
by previous trial had been found to give the most uniform temperature in 
A. The furnace was mounted on iron rails to facilitate this operation. 
The determination was then proceeded with as already described. At 
the conclusion of the determination the apparatus was taken to pieces and 
the volume of the various parts was determined by weighing out with 
mercury. 
The volume of the bulb A, including the tube C and the portion of B 
contained in the large Herseus furnace, was found to be 63‘73 c.c. 
The volume of the connecting portion of the capillary tube B which 
was heated by means of the bunsen flame was CK)6 c.c. The volume of the 
