440 Dr. F. Gr. Donnan on the Relative Rates of 



error in noting the points of passage owing to the very rapid 

 rate of effusion. 



Owing to the high viscosity-correction it was thought 

 advisable to re-measure argon and carbon monoxide as a 

 check on the results given by this effusion-plug. Results: — 



IB. S. 



Oxygen .... 4 3*8 



Argon .... 4 21-1 (density 19-71) 



Oxygen .... 4 3 # 4 



Argon (calc.) . . 4 29*9 3*3 per cent, faster. 



Oxygen . . . . 4 3*2 



CO 3 46-4 



Oxygen .... 4 3*4 



CO (calc.) ... 3 46-6 



These results are in fairly close accordance with those pre- 

 viously obtained*, so that the method of correcting for viscosity 

 appears to be justified in the main. 



Owing, however, to the unpleasant prominence of the 

 viscosity-correction in the case of helium, several attempts 

 were made to obtain a better effusion-plug. An aperture 

 was finally obtained showing a much smaller viscosity- effect. 

 The pierced disk of thin foil was sealed transversely in a tube 

 of flint-glass and the latter sealed on to the rest of the appa- 

 ratus. This forms the simplest and most secure method of 

 attaching the platinum-foil disk to the effusion apparatus. 

 With hydrogen and oxygen the following results were 

 obtained : — m. s. 



Hydrogen . . 2 7*9 

 Oxygen . . 8 21*3 



The time for hydrogen calculated from the densities is 

 2 m 5 s * 8, so that a small viscosity-effect still exists. These 

 results give % = 122*5, « 2 =11'3. Helium and argon were 

 now remeasured. 



He Hum. —Density 2*314. 



in. s. 



Oxygen ... 8 22*2 

 Helium ... 3 16*8, 3 16-6 

 Oxygen ... 8 24'2 



Mean time for He observed .... 

 Time calculated w T ith viscosity-correction 

 „ „ from densities alone 



* As evidenced by the density, this specimen of argon was not very 

 pure. It probably contained some hydrogen. This would account for 

 the rather low " deviation " of 3*8 per cent. 



m. 



s. 



3 



16-7 



3 



16-7 



3 



11*2 



