1022 



tlierefore, greater and iindei- cerlain circumstances may, it seems, 

 become 1., so that Roscoe and Bunsen's law would become valid. 

 According to measurements made by A. E. Webf.r ^) this would be 

 the case when the number of revolutions of the sector is more than 

 120 per minute. A similar result was obtained by F. Eckekt and 

 R. PuMMERER -) and also by H. E. Howe ^). A. Odencrants *) also 

 finds an increase of p, but he does not consider that Bunsen and 

 Roscoe's law may be applied. There is thus still some uncertainty 

 as to the foundations upon vvhich our method rests. This is, moreover, 

 increased by the fact that we determine the ends of the line that 

 are still just visible, and in consequence measure in the field of 

 sub-illumination. For the sake of simplicity we have, therefore, 

 in constructing the sector assumed the validity of Bunsen and Roscoe's 

 law, and subsequently investigated by measurements in how far 

 this assumption is borne out. 



The first sector was so constructed, that the length of the spectral 

 line was proportional to the intensity of the light. The intensity- 

 variations along the line were then, however, too small to be easily 

 measurable. We therefore proceeded to make a sector with which the 

 length of the lines becomes proportional to the logarithm of the intensity 

 (fig. I), with (his we obtained satisfactory results. For the sector used 



Fig. 1. 

 log-^^],h[I,-l,) 



where / means the length of the spectral line '"). 



3. A few measurements. 



In order to test the reliability of our method we carried out 



1) Weber. Ann. d. Phys. (45) 801. 1914. 



2) EcKERT and Pummerer. Zeitsch. f. physik. Ghem. (87) 605. 1914. 



3) Howe. Phys. Rev. 8, 674, (1916). 



4) Odencrants Zeilsch. f. Wissensch. Phot. (16) 69, 111. 1916. 

 ^) For further particulars cf Hamburger, Diss. Delft 1917. 



