426 Relative Amounts of Krypton and Xenon in Air. [Mar. 9, 



Total krypton obtained 0*0028 gramme. 



xenon obtained 0-0005 „ 



Percentage krypton in gaseous air 0*000014 by weight. 



xenon „ „ 0-0000026 



Krypton equal to 1 part by weight in about 7 millions of air; by 



volume, 1 part in 20 millions. 

 Xenon equal to 1 part by weight in about 40 millions of air ; by 

 volume, 1 part in 170 millions. 



As before remarked, it is not maintained that all the krypton and 

 all the xenon have been separated; it is likely, however, that the 

 separation of the xenon was more perfect than that of the krypton. 

 The results are merely brought forward as the result of a careful 

 experiment to quantitatively isolate these gases. 



I have to express my cordial thanks to Mr. E. C. C. Baly and to 

 Mr. Inglis for aid in carrying out part of these operations. 



As a quantity of pure krypton, sufficient for determination of 

 density, had been collected, occasion was taken to redetermine the 

 density of that gas. It was submitted to careful fractionation; a 

 considerable portion was rejected as possibly containing argon, and the 

 dregs were set aside as possibly having contained xenon. The sub- 

 stance weighed had a low vapour pressure, — about 15 mm. at the 

 temperature of the liquid air used in fractionating. The separation of 

 the lighter and heavier portions was repeated four times, the density 

 having been determined on each occasion, with only small differences. 

 Finally, a very careful determination of density was carried out, with 

 the following results : — 



Volume of density-bulb 7*268 c.c. 



Temperature 15°*57C. 



Pressure on gas, corrected 754-0 mm. 



Weigh t 0-02488 gramme. 



Hence density, compared with 0=16... 40-81. 



Previous determinations with two samples of gas, one fractionated 

 from argon, the other fractionated from xenon, gave 40*82 and 

 40*73 respectively as the density. The result given above is in 

 perfect concordance with these figures. The chief cause of error is 

 in the weight ; I think it would be fair to regard two units in the fifth 

 place as the limit of error, which gives a possible divergence of about 

 1 part in 1200. 



The atomic weight of krypton would accordingly be 81*62; the 

 mean of former determinations is 81*28. This is in accordance with 

 its position in the periodic table, which lies between bromine, 80, and 

 rubidium, 85. 



