478 Dr. Tycho E:son Auren on the 



wave-lengths, as shown by the experiments of Barkla. When 

 making controlling experiments, the absorption vessels being 

 placed at a distance of 16 cm. from the ionization-chambers, 

 the same values of p have, within experimental error, been 

 observed. 



Although the equation (1) whence (4) is deduced only holds 

 when the radiation is homogeneous, this formula, according to 

 Barkla's law, confirmed by various experiments (pp. 472-473) 

 must hold as well when non-homogeneous radiation is used, 

 supposing that the wave-lengths of the rays are not the same 

 or nearly the same as those of the characteristic radiation of 

 the element to be examined, and that rays of too great a 

 wave-length are not used. 



In the radiation used by the author from a bulb with an 

 anticathode of tungsten, and with a spark-gap of at least 

 10 cm., the rays have been filtered through an aluminium 

 screen 1*25 mm. thick. By the heterogeneous radiation, 

 the rays of a greater wave-length have thus been filtered 

 away, and the rest must certainly have got a decided 

 maximum of intensity. When heterogeneous rays from a 

 bulb with a spark-gap of some centimetres are transmitted 

 by an aluminium screen sufficiently thick, the passing rays 

 may practically be considered as homogeneous *. When 

 changing the spark-length at the bulb, there has not in any 

 case been observed either in lighter or heavier substances any 

 influence on the determinations of the relative atomic ab- 

 sorption coefficients. Within the range of the wave-lengths 

 used by the author, the determinations of these coefficients 

 are independent of the wave-length, and thus Barkla' s law 

 must be applicable. 



By determining in the ordinary way the thickness of the 

 aluminium sheet necessary to diminish the ionization by 

 about half, the absorbability of radiation could, by means of 

 formula (1), approximately be determined. Thus the ap- 

 proximate wave-length could be evaluated f, and as a mean 

 value \ = 035 . 10~ 8 was obtained. Among the substances 

 examined certainly W, and probably also Ce and Ba, have 

 entered into the range of selective absorption. 



The rays I have made use of have not been homogeneous, 

 it is true, but in determining X. a. k., with regard to absorption 

 for different sheet thicknesses, theymave behaved in the same 

 way as if they had been homogeneous. Indeed, it appears 



* Compare Kave, ' X-Rays,' London, 1914, p. 103. 

 t Barkla and Dunlop, Phil. Mag. xxxi. p. 229 (1916). 



