Anode Rays to the Investigation of Isotopes. 865 



the metals, showing that they had less energy . The difference 

 presumably corresponds to the anode fall of potential. 



These lines were always faint, but were not much affected 

 by whether the anode was working well or not, so that they 

 are relatively more important on the fainter photographs. 

 The following j arabolas of this kind have been measured : — 



(1) Carbon (12). This seems to appear more strongly 



when the anode is nearly worn out. 



(2) Water (18). This appeared strongly when SrBr 2 



was used in the anode. This salt had had water 

 of crystallization which had been removed by 

 heating, and it is possible that a little still 

 remained. It also appeared more faintly with 

 other anodes. 

 (3,) Fluorine (19). This occurred fairly often when a 

 fluoride was used. 



(4) Carbon monoxide (28). Perhaps the commonest 



non-metallic line. 



(5) Chlorine (35 and 37). Doubtful if ever observed. 



(6) Carbon dioxide (44). Aiwa) s very weak. 



(7) Iodine (127). Fairly common when iodine was in 



the tube, but always weak. 

 Bromine was never observed. Hydrogen molecule was 

 observed twice ; it was probably often present, but the 

 deflexions were too great for it to appear on the plates. 



Advantages and disadvantages for isotope separation 

 of various types of ray. 

 It may be of interest to compare the different methods 

 which have been used up to the present to produce posi- 

 tively charged metallic atoms for the purpose of investigating 

 isotopes by electromagnetic analysis: — 



(1) Passing a discharge through a volatile compound of 

 the metal. This has been used with success by 

 Aston * for nickel, using nickel carbonyl ; but it is 

 not an easy method, as most of the volatile metallic 

 compounds are easily decomposed and cease to 

 exist after the discharge has been passing for a 

 short time. Even in the case of nickel, several 

 unsuccessful attempts were made to obtain it in 

 the early days of positive ray analysis. The ap- 

 plicability of the method depends entirely on a 

 sufficiently stable compound being available, and 

 it is doubtful if more than a low metals can be 

 done in this way. 



* 'Nature,' June 28rd, 1921. 



