( 352 ) 



The most striking example of asymmetrical resolution that I know 

 of, occurs in (he case of one of the yellow mercury Tines (5791). 

 The structure of a line like this one cannot be made out by means 

 of Michelson's interferometer. Indeed the assumption of symmetry, 

 which is, as has been proved by Lord Rayleigh 1 ), necessary to 

 deduce the structure from the visibility curve in this case certainly 

 is unjustified. 



3. Following the method described in the first part of this paper 

 1 have made some experiments concerning the mentioned spectral 

 line in strong fields. For the Rowland grating used in my obser- 

 vations I am indebted to the dutch Society of Sciences at Haarlem. 

 Presently I hope to give an account of results obtained in weak 

 lields by means of an interference method. 



The grating has 10.000 lines to the inch and a radius of curvature 

 of <>.5 M., the divided part being of 14 cm. width. In the use of 

 my method the grating necessarily should be mounted in such a 

 manner that to every point of' the slit corresponds only one point 

 of the spectral image. Rowland's concave grating can be mounted 

 in a non-astigmatic manner as has been remarked by Runge and 

 Paschen s ) and this arrangement was made use of in former inves- 

 tigations by myself, n ) Hallo and Geest. ') 



All observations recorded in the present paper were made with 

 the spectrum of' the first order. 



4. Whereas the mercury line 5791 is resolved asymmetrically, the 

 neighbouring line 5770 is resolved by (he magnetic field into a 

 perfectly symmetrical triplet, or at least very approximately so. 1 

 have used this circumstance for applying the optical method of 

 measurement of field (see § 1), the mentioned yellow lines being 

 easily photographed simultaneously. 



Fig. 1 represents a ninefold enlargement of one of the negatives. 

 According to measurements of Fabry and Perot the difference of 

 wavelength of the yellow mercury lines is 5790.66 — 5769.60 = 21.06 

 A.U., hence 1 m.m. in Fig. 1 corresponds to 0.551 A.U. Inspection 

 of Fig. 1 clearly shows that line 5791 is asymmetrically resolved. 

 Perhaps this is still more evident in the enlargements Figures 2 and 

 3 of parts of Fig. 1. 



') Rayleigh. Phil. Mag. November 1892. 

 5 ) Runge and Paschen. Wied. Ann. Bd. 61. p. 641. 1897. 



:; ) Zeeman. Archiv. Néerl. (2) T. 5. 237. 1900, T. 7. 405. 1902. These Proc. 

 May 1902, May 1903, Dec. 1904. 



*) Hallo. Archiv. Need. (.2) T. 10. p. 148., Geest. (2) T. 10 p. 291. 1905. 



