420 Professor Sir James Dewar [June 5, 



purify the entering gases. After the charcoal condenser is cooled, the 

 vacuum in C becomes so high that the coil discharges across a 1-inch 

 spark gap outside rather than through C. The stopcock D is turned 

 to give a charge of dry air to fill the quill tube between D and the 

 capillary B. The conductivity of C now becomes high, the outside 

 spark gap stops acting, and the discharge again passes in C. The 

 vacuum, however, rapidly becomes high until C begins to show a 

 good glass phosphorescence. B is now cooled in liquid air, when a 

 rapid leak of gas through B into C becomes apparent by the red glow 

 of nitrogen round the positive pole through the viscosity of the air 

 being reduced by the cooling, thus mcreasing the amount of leakage. 



The converse experiment can be shown thus : when the vacuum 

 is not low, e.g. just after turning D to admit a charge of air, we heat 

 B with a spirit-lamp. The resistance in C now rapidly rises until 

 sparks pass across the outside gap due to the high exhaust in C. When 

 air is replaced by dried hydrogen the same phenomena can be shown. 



The expression for the viscosity /x of a gas at absolute tempera- 

 ture T is given by u = ?^T" where 71 and a are constants. Rayleigh's 

 value of a for air = 0* 75, so that a change from 15° 0. to - 183° C. 

 would diminish the viscosity 2^ times in any case. A series of 

 quantitative experiments must be undertaken on the viscosity of 

 hydrogen down to 20° absolute. 



Incandescence of a Thoria Mantle produced by Ceookbs 

 Rays in a Charcoal Vacuum. 



A small piece of an ordinary gas mantle about 1 cm. square is 

 supported by platinum wire near the focus of the concave aluminium 

 cathode of a vacuum tube 15 cm. in length and 6 cm. diameter 

 (Fig. 5). The platinum wire carrying the piece of Thoria mantle 

 is fused into a glass rod B wliich can be pushed into any position 

 by means of the indiarnl)ber joint fitting B. As gases diffuse 

 through the rubber tube at B a slight leak is maintained into the 

 vacuum tube. The final exhaust is controlled by means of a side 

 tube provided with a stop-cock and bulb containing 20 grm. of 

 charcoal (G) placed in liquid air. 



Starting with a few min. pressure in the tube, the cooled char- 

 coal is put into action by opening the stopcock. The pressure con- 

 sequently falls, as indicated by the discharge. In a short time a dull 

 red spot shows in the centi'e of the square of the mantle ; this rapidly 

 increases in intensity until a brilliant incandescence is obtained. Very 

 soon however, as the pressm'e continues to fall l^y the action of the 

 cooled charcoal, the incandescence again diminishes to a dull red spot 

 and soon goes right out. 



Now by closing the charcoal cock, the slight leak round the india- 

 rubber joint causes the pressure again to rise in the tube and the 



