186 Dr. Tycho E:son Auren on 



arranged in the proximity of the nucleus, more especially if 

 they are arranged in some sort of planetary orbits. If the 

 scattering phenomenon is produced solely by the outer 

 electrons, the absorption due to this process according to 

 the said theory must be proportional to the number of such 

 electrons. Of the two terms of the formula for /c a , the term 

 seems to be determined by the number of outer electrons. 



In the elements, except in the lightest, absorption is chiefly 

 due to the inner electrons and the manner in which they 

 are grouped in relation to the nucleus. On increasing the 

 atomic number, presumably rearrangements of electrons must 

 take place in certain elements, whether according to Bohr 

 one assumes that new electron rings are formed or simply 

 assumes a discontinuous change of concentration by sudden 

 leaps. When such a rearrangement takes place it must make 

 itself perceivable by a sudden increase in absorption. In 

 previous experiments I found that a perceivably sudden 

 increase in absorption takes place in the elements 0, S, Mn,, 

 and Cu, which in accordance with the same view might be 

 assumed to be due to rearrangements of the inner electrons 

 of these elements. In the experiments I have given an 

 account of above, a similar sudden increase in absorption has 

 been found again in the same elements except Mn. Owing 

 to an error in the analysis of the first-examined solution of 

 CrCl 3 , somewhat too low a value was obtained for Or. The 

 point where rearrangements seem to be occurs at this element* 

 In addition, similar points seem to be at the elements Mg 

 and Ca. 



At the elements following from Cu there seems not to be 

 any rearrangement of the inner electrons to Pd ; but as the 

 experimental errors may possibly be a little greater here than 

 in the preceding elements, this point may not be quite certain. 



To obtain comparable values of the absorption coefficients 

 for the lighter elements, it is necessary to correct the values 

 found in relation to the absorption due to scattering. For 

 Cu, with the mean wave-lengths used by me, the term () a in 

 the formula (4) compared with k a X b is certainly very insig- 

 nificant and may, in maxima, be estimated to 1 per cent, of 

 the whole. If this term is neglected for Cu, we obtain the 

 following formula for calculating fc a /Q n : — 



or 



Vcu=>/oa-r^-TT- • • • (<>> 



Uu Lu 



