CIS 



Sir James Dewar 



[Jan. i: 



oil the outer glass surface was removed. The efficiency of the vessel 

 at the highest vacuum was thereby diminished to one-fifth — its rate 

 of evaporation being increased five-fold ; therefore approximately 

 80 per cent, of the total radiant heat seems to be reflected at the 

 outer silvered surface. At the same time the rate of increase of 

 evaporation with lov^ering of vacuum remained unaltered, as is shown 

 by curves (1) and (2) being parallel. When the similar value for 

 the nickel vessel is compared with this a much more rapid deteriora- 

 tion is found to occur, as appears from the steeper slope of curve (3). 

 The actual increments were found to be an additional litre per 

 minute per O'OOl mm. Hg increase in pressure in curves (1) and 

 (2), and approximately double this value in (3). Above 0*001 mm. 

 the increment is no longer linear, and when the pressure has risen in 

 the glass flask to 0*010 (teri times the previous limit) the increment 

 of evaporation is now one litre per minute per 0*0167 mm.; or 

 al»out one-seventeenth of the initial value. These values should be 

 considered in conjunction with the following table, comparing the 

 efficiency of flasks with different reflecting surfaces. They were 

 all exhausted to the limiting high charcoal vacuum, and filled to the 

 same extent with old liquid air : the steady rate of evaporation was 

 then observed. The total heat influx thus determined was reduced 

 to calories per unit area per minute, and gave the following resulr;S : — 



The •• Emissivity " values in the last column are loaded by the 

 neck losses, conduction and convection. These are relatively higher 

 for smaller flasks : so that Nos. 2 and 5 bear disproportionately high 

 apparent emissivities. No. 2 was deduced from evaporation values 

 given by Berry"' for an ordinary pattern 1 -litre silvered glass flask, 

 thoroughly exhausted by cooled charcoal. The neck dimensions of 

 such a flask are usually about 3 cms. diameter by 10 cms, long. The 

 emissivity value obtained from a vessel of similar capacity but much 

 narrower neck was even lower than for the 6i-litre vessel (No. 1 

 above), which however had the usual neck aperture of 3 to 4 cms. 



* Proc, Kov. Sec, A, Ixxviii. 456. 



