with a Liquid Mercury Anticathode. £27 



Table I. 



\ . 10 n cm. V /N. 



I 1418-3 6425 



1249-7 729-2 



1238-5+0-3 735-69 



a 2 1249-7 729-2 



t) 11619 784-3 



/3 6 1077-4 845-8 



& 1068-6 852-7 



/3, 1045-8+0-6 871-35 



j3 2 1037-5+0-4 878-33 



|3 3 1030-1 884-6 



/3 7 /5 ... 1007-8 904-2 



y 5 914-4 9966 



y x 893-5+0-7 1019-9 



y2/3 ... 869-5 1084-0 



y 4 834-8 1091-6 



The third column of Table I. contains the frequency 

 of the Hg lines divided by the Rydberg frequency N. 



Another test was made by finding the frequency differences 

 of the doublets according to Sommerfeld's theory. The 

 following table shows that these differences are nearly 

 constant. The accuracy obtained in this work is not high 

 enough to decide whether the small lack of constancy has a 

 physical meaning or is simply due to experimental errors. 



Azz/N. 

 A(5 x a 2 1421 



A 7l /3 7 141-6 



Ar\l 141-8 



^72/3^7/5 143-8 

 Ay 5 & 143-8 



It is well-known that the wave-lengths of the corre- 

 sponding lines in the X-ray spectra of the elements are 

 continuous functions of the atomic number. Considering 

 this fact it was easy to calculate from the experimental data 

 of Coster I. c. that the lines <y 2 , 73 should be very near 

 together for the element with the atomic number 80, that is 

 for mercury. 



This was confirmed by the spectrograms, where only one 

 line appeared instead of a pair. 



In conclusion I wish to express my best thanks to Sir 

 Ernest Rutherford for his interest in the work and for the 

 readiness with which he placed the necessary apparatus at 

 my disposal. 



Cavendish Laboratory, Cambridge. 

 May 1921. 



