546 Professor Sir James Dewar [Jan. 



9'-) 



to further uDify the heating, which was supphed by a small fan- 

 shaped bunsen flame, just not luminous. On either side of the 

 oxidised copper the tube is bent into a U shape for cooling in liquid 

 air, to condense any volatile hydrocarbons from the stopcock and the 

 oxidised products from the action of the hot CuO on any combustible 

 material in the gas sample. 



The exact measurement of samples small in volume is made by 

 lowering the - pressure to any convenient amount in the Imrette C, 

 in which the gas collects during the purification by heated CuO. 

 The burette opens below under the surface of merciiry in a vessel D 

 with a connection to the exhaust pump ; an exhausted tube E as 

 barometer also opens into the mercury in the same vessel. From 

 the difference of levels of the mercury in C and E the real vohime 

 of the gas sample at normal temperature and pressure can be deter- 

 mined. After measuring the gas, the cock F is turned to connect to 

 the main stopcock G, by which the gas is cautiously admitted to the 

 cooling bulb H, and through this to the McLeod gauge ; G- also con- 

 nects by another way to the molecular pump, by which the cooled 

 products can be highly exhausted. The bulb H is formed of two con- 

 centric tubes Ih cm. and h cm. diameter respectively. The inner tube 

 is continued up to connect to the gauge, while the connection to G 

 is on the annular space ; by this means no gas can reach the gauge 

 until it has passed through the cooled part of the bulb H immersed 

 in the liquid hydrogen. 



When the charcoal and liquid air method is employed, 10 grams 

 of charcoal are placed in an enlarged part of the tube H, which is 

 immersed in the liquid hydrogen. Liquid air or liquid nitrogen or 

 oxygen replaces the liquid hydrogen in the inner of the vacuum vessels 

 J, the outer still containing liquid air, under exhaust if necessary. 



The volume of the cooling bulb and McLeod gauge up to the 

 main stopcock G was carefully determined. The volume of the gas 

 sample admitted through G being also measured, the ratio of this to 

 the total volume of the gauge and cooling tube is known, from 

 which the partial pressure which would result if the gauge had been 

 filled with the gas at atmospheric pressure could be deduced. 



This is given by observing the partial pressures with successive 

 increasing volumes of gas occluded in the charcoal. Plate 2 expresses 

 the results graphically. The partial pressures observed are ordinates, 

 while the abscissae are the gas volumes admitted. The two lowest 

 graphs, Xos. 1 and 2, show the values given when 10 grams of 

 charcoal at the respective temperatures of 77° Abs. (boiling liquid 

 nitrogen) and 80' Abs. (old liquid air, temperature controlled by 

 oxygen vapour tension thermometer) are employed as cooling agents. 

 The volume of successive increments was 5i) or 100 c.c. ; the 

 volume of the apparatus was just over 350 c.c. ; therefore not until 

 this volume had been admitted does the recorded partial pressure 

 correspond to 700 mm., or one atmosphere, in the gauge. This 



