Appendix to Lecture of Jan. 22, 1915 821 



charcoal. Inside A is a smaller tube about 5 mm. in diameter 

 reaching- to the charcoal ; it is sealed to the top of the outer tube 

 and continues up to the sparking-tube B. The Regnault three-way 

 stopcock C is sealed near the top of the annular space of A by a 

 short connection of about 2 mm. bore. The ^'ertical branch of 

 C terminates in a ground cone as close to the stopcock as possible. 

 The brass cone of an indiarubber membrane support D fits this 

 vacuum tight. To limit the free space below the membrane the 

 drum support is reduced to very little more than the depth of a 

 few thicknesses of stiff dried porous paper, which are placed inside 

 to support the membrane. The remaining branch of C serves for 

 connecting the preliminary exhaust. The membrane support passes 

 through a cork into an inverted bell jar E, from the loose cover of 

 which a dish F of phosphorus pentoxide is hung. The air gaining 

 access to the membrane through the loose cover is thus dried. To 

 equally serve this purpose the membrane may be enclosed in its 

 gas-box, Fig. 1 (a), with the inlet tube bent down into a U immersed 

 in the liquid air round A. 



It is only necessary, after heating and exhausting out the charcoal 

 and sparking the Pt electrodes of B, to turn C to connect D to the 

 charcoal tube, and then arrange a flask of liquid air to keep the 

 charcoal continuously cooled. The annular construction of A, when 

 in the liquid air, keeps B free from any volatile impurity coming 

 from the membrane or the stopcock lubricant. 



The air diffusion rate of the membrane used was previously 

 determined in each case ; thereby a knowledge of the amount 

 of air diffusing through and condensed in the charcoal in 

 any time is determined, the area of the membrane being also 

 known. 



For a fairly rapid indication, a membrane of 50 sq. cm. area, 

 having an air rate at 15' C. of 1-44 c.c. per day per sq. cm., was 

 employed with half a gram of charcoal. Within 12 hours the whole 

 helium and neon spectrum was visible when a discharge was passed 

 between the platinum wire poles of B. o6 c.c. of air would now be 

 condensed in the charcoal. The F line of hydrogen was faintly 

 seen, with the beginning of some nitrogen lines. This experiment 

 continued for four days, when on removing the charcoal from the 

 liquid air 290 c.c. of gas was obtained containing oG per cent of 

 oxygen. 



To study the manner of growth, of the gas discharge more 

 steadily, another experiment was made with a thicker membrane 

 of half the area, i.e. 25 sq. cm. ; the air diffusion rate was only 

 0'82 c.c. per day per sq. cm. A contained 2 grams of charcoal. 

 The resulting observations were as follows : — 



18 hours : 15J c.c, condensed : He yellow line alone, quite bright, with 

 much phosphorescence, 'therefore low pressure. Nothing further 

 showing until 



