Absorption of X- Rays. 485 



oxidized by the contact of air and hydrolyzed by dilution, 

 partly passes into a colloidal state. As I pointed out earlier, 

 other solutions, that behave in the same way when diluted, 

 display the same property with respect to absorption. The 

 cause o£ absorption being greater in colloidal solutions seems 

 to be the fact that, as shown by Galecki *, the number of 

 particles irradiating a colloidal solution by X-rays is strongly 

 diminished, in which process part of the radiation energy 

 is required for the disintegration of the particles. 



When values for bivalent Fe have been found in greater 

 concentrations, they seem to be more exact than those found 

 at a greater dilution. The mean value, 41'8, of the two 

 first determinations is considered to give the probable value 

 of X. a. k. for bivalent Fe, this value, however, seems to be in 

 reality somewhat too high. The difference between this and 

 the value of trivalent Fe, 45'5, found on examination, is 

 scarcely within experimental error, and, therefore, bivalent 

 Fe is likely to have a lower absorption coefficient than 

 trivalent Fe. 



X. a. k. of carbon from cyclic combinations (benzene, aniline, 

 pyridine, nitro-benzene) seems to get a lower value than that 

 of carbon from the alifatic series, which fact may be traced 

 to some difference in the inner structure of the two kinds of 

 carbon. Since, however, as it appears from the figures quoted, 

 a slight uncertainty with respect to X. a. k. of the elements 

 forming combinations with carbon may in a high degree 

 influence the results, there cannot, it seems, at present 

 be drawn any trustworthy conclusions from the differences 

 that have been observed in the measure of the absorption 

 coefficients for different organic combinations. By continued 

 researches the author hopes to be enabled to elucidate this 

 question. As the sensibility of the methods used here can be 

 increased, it will be possible to increase the exactness in 

 determining the relative absorption coefficients. 



In the same range of ideas, it seems worth noticing that 

 the error in the values found for X. a. k. must be regarded as 

 greater for the elements calculated from combinations whose 

 X. m. k., estimated in percentage, is mostly determined by 

 other component parts of the combinations, e. g. Xa from 

 XaCl and Al from A1 2 (S0 4 ) 3 . 



The comparative slightness of the error is made plain by 

 the good agreement between the values deduced from the 

 above combinations as well as from the same elements in 

 other combinations or in form of metals, e. g. for Na from 

 XaOH and for Al from aluminium metal. 



* kolloidal Z. x. p. 149 (1912). 



