244 Mr Kaye , The Selective Absorption of Rontgen Rays. 
by rays from other metals — to an extent depending on the 
proximity of the atomic weight to that of platinum, e.g. lead and 
silver have their former values cut down by about 20 per cent., 
while nickel and aluminium are about 35 per cent, less than they 
were with the aluminium screen. 
Now compare the column of values for the copper screen with 
those for the aluminium screen. For ease of comparison copper is 
given the value 35 — the same as it had with the aluminium screen. 
It will be seen at once that there is evidence of marked selective 
transmission. A copper screen is now easily first in transparency ; 
the next is nickel whose value exceeds those of silver, platinum 
and lead, although their atomic weights are much higher. 
Similar results will be observed when the columns for nickel 
and iron screens are also compared with the column of values for 
the aluminium screen. Take the case of the iron screen, and 
compare it with the aluminium screen. Then giving the radiation 
from iron the same value 30 in each case, it will be seen that the 
new values for lead, platinum, and silver are about a quarter, and 
those for copper, nickel, and aluminium about a third of their 
values with an aluminium screen. This again indicates that the 
degree of transparency of a metal to Rontgen rays, is a function of 
the proximity of the atomic weight of the radiating metal to that 
of the screen. 
It is worth noticing the big jump that occurs in the character 
of the radiation in passing from nickel to iron, and the resemblance 
between the radiation from nickel and from copper. For instance a 
nickel screen is more transparent to radiation from copper than to 
radiation from iron ; a copper screen transmits radiation from nickel 
far better than radiation from iron, while a screen of iron transmits 
radiation from copper and from nickel almost equally badly. 
Barkla*, working with secondary Rontgen radiation, has noticed a 
similar resemblance between copper and nickel, and a correspond- 
ing jump between nickel and iron. 
Yet it must be remembered that the atomic weight of nickel 
(58’7) is nearer the atomic weight of iron (56) than it is to that 
of copper (64). Nickel and iron show some chemical resemblances 
and are placed in the same group (vm) of the Periodic Classifica- 
tion, copper being situated in group I. On the other hand, a 
point worthy of remark is that copper and nickel have densities 
which differ but little — 8‘95 and 8'8 respectively, while the 
density of iron is distinctly less — 7'8. It would seem therefore 
that the density of the absorbing screen, rather than its atomic 
weight is the important factor in this screening effect. 
Barkla, Nature, Feb. 14, 1907. 
