Effusion of Argon, Helium, and some other Gases. 439 

 A repetition of this test on the following day gave : — 



m. 



Oxygen .... 4 2*9 

 Hydrogen ... 1 3*3 



These results lead to a much higher viscosity-correction 

 than in the case of the effusion-plug first used. We get 

 /e 1 = 57*6, /c 2 =12'4:. Measurements with cyanogen gave the 

 following results : — 



m. s. m. s. 



Oxygen 4 5 



Cyanogen 5 4*9 5 4'2 



Oxygen 4 4 



Mean time for Cyanogen . 5 4*6 



Graham gives '506 for the viscosity of cyanogen referred 

 to oxygen. The calculated value is therefore 



(_ 244* 5 \ 



^26 x 57*6 x |jg^ j + (12-4 x -506) = 5^ VS. 



There is a deviation here of '9 per cent, in the direction of 

 slower effusion. The time calculated from the densities 

 alone is 5 m 11 s * 6. 



The helium previously used was sparked over potash with 

 oxygen and the oxygen removed hy heated copper. After 

 being dried by P 2 5 it gave the following results : — 



m. s. 



Oxygen . . . 4 3'8 

 Helium .... 1 37'1 

 Oxygen ... 4 3*4 



The viscosity of helium referred to air was found by Lord 

 Rayleigh to be *96, so that its value referred to oxygen is '865. 

 The density of this specimen of helium was determined and 

 found to be 2*314, so that it contained a considerable amount 

 of argon. The calculated time is l m 38 *6, from which it 

 would appear that the helium effuses 1*5 per cent, faster than 

 as calculated. This helium after being further dried over P 2 5 

 was examined again, with the following result : — 



m. s. 



Oxygen 4 3*2 



Helium 1 37'6 



Oxygen 4 4'0 



Helium (calc.) 1 38*4 



Helium (calc. from density only) . 1 32*6 

 Here the deviation only amounts to *8 per cent., which is 

 about half the previous amount. Possibly this want of agree- 

 ment in the results is partly due to the increased likelihood of 



