UI 



since Wood has found the scattered rays to be unpolarized, 1 am 

 i-ather doubtful as to tlie propriety of applying the above theory to 

 his experiments. For tliis reason, I shall no longer dwell on this 

 question. ') I shall only add that the value //„ which we found, leads 

 to a value of /i„i„ considerably below 1. 



§ 10. The fornuilae (24) and (25) may also be used for calcula- 

 ting the total absorption, integrated over the whole width of the 

 line, for a certain thickness of a given gas. On the other hand this 

 absorption can be measured by a simple photometric experiment. 

 Dr. G. J. Elias was so kind as to do this for iodine vapour. 



A beam of yellow light was passed through an evacuated lube 

 containing some small iodine crystals and heated to 89° C. The 

 beam was obtained by isolating from the spectrum of an arc lamp 

 a portion corresponding to the distance between the D lines. In a 

 layer of 2 cm. the absorption amounted to 15 7o- 



In discussing this result, I shall remark in the first place that the 

 distribution of light in an absorption band will depend on different 

 circumstances, e.g. on the tliickness of the gas traversed. It may be 

 that at the middle of the line and within a certain distance from 

 it practically all light is absorbed, the absorption diminishing gradu- 

 ally on both sides. However this may be, one can always define 

 a certain width A ;« such that the amount of light absorbed by the 

 gas is equal to the quantity of light that is found in the incident 

 rays within the interval AP.e. The magnitude of AP.^,, which we may 

 call the "effective" width of the line, can be immediately deduced 

 from a photometric measurement. 



The absorption spectrum of iodine vapour has a very complicated 

 structure, containing somewhat over 100 lines between the D lines. 

 Dr. EuAs's observation shows that the effective widths of all these 

 lines taken together amount to 15"/„ of the distance between the D 

 lines, i.e. to 0,9 A.U. We shall therefore not be far from the mark 

 if for one line we put on an average 



A}.e = O.QQS A.U. 

 If J (Ü is the intensity of the incident light within the interval 

 (Ü-, we have for the absorption over the whole width of a line by 

 a layer of thickness (f 



7 r{\~e--i!>^) dX. 



1) According to more recent measurements by A. v, Malinowsky (Resonanz- 

 strahlung des Qiiecksilberdampfcs, Ann. d. f^iiysik 44 (1914), p. 935) /(q = 1,55. 



