X- Rays from Boron and Carbon. 155 



when definite breaks were obtained, it was found that time- 

 changes were well marked. With a new surface the 

 photo-electric effect (referred to unit electron current from 

 F to T) was always small, and increased as the experiment 

 progressed to a value perhaps two to four times the initial 

 value. Later on it became much more steady. At first one 

 Was tempted to wait till the equilibrium condition had been 

 reached, as measurements seemed then to be more re- 

 producible, but unfortunately, at this stage, the breaks had 

 generally disappeared. It was therefore necessary to take 

 observations as soon as possible after the filament had once 

 been lighted, and then to take observations quickly at equally 

 spaced time-intervals from the highest accelerating potential 

 chosen to the lowest and back again, and repeated several 

 times if possible, so as to obtain data from which a mean 

 curve could be calculated. It is likely that the target got 

 covered with a thin layer of some foreign matter, possibly 

 evaporated tungsten. This would be a plausible explanation 

 of the obliteration of the breaks characteristic of the target, 

 for no doubt a very thin layer would be sufficient to absorb 

 th^se slow electrons. The increase in the intensity of the 

 radiation as time goes on fits in well with the known fact 

 that the general radiation from heavy atoms is more intense 

 than that from light atoms. 



Besides the fact that each curve from which the critical 

 values of the potential are derived is itself a mean of four, 

 six, or eight consecutive runs, each run itself showing the 

 break, there is other evidence in favour of the reality of 

 these small but definite breaks. That these breaks are not 

 due in some obscure manner to the particular apparatus and 

 measuring instruments is shown (1) by the fact that, both 

 types of apparatus gave breaks at the same place for boron, 

 and (2) that no evidence of the carbon breaks was obtained 

 when boron was used as target, and vice versa. Neither with 

 boron nor with carbon was there any trace of a bretik 

 between the K and L points, the curve between these points 

 being smooth and very nearly straight. 



Discussion. 



It is of interest to see how these values for the character- 

 istic radiations of boron and carbon fit in with the values for 

 other elements. The values of the Ka emission-lines are 

 known for the elements down to sodium, while the values of 

 the K critical absorption wave-lengths (Ka) are known 

 to magnesium. The question then arises, which of these 



