Jack — 3tagnetic Resolution of the Spectrum Lines of Niobium. 43 



improvement in the apparatus, these times, which were necessary when the 

 photographs for tungsten and molybdenum were taken, might with advantage 

 have been considerably shortened. The observations were made in the 

 direction perpendicular to the magnetic field. Photographs were taken 

 separately of the vibrations perpendicular (s) to the magnetic liaes of force, 

 and of those parallel (2^) to these lines. By the aid of those two sets of plates 

 the kind of vibration corresponding to any line could be known, and, further, 

 the cutting out of one set of vibrations made the intervals between the 

 neighbouring components larger, and hence these could be more easily and 

 more accurately measured. Photographs were also made with all the 

 components included, that is, containing the vibrations both perpendicular (s) 

 and parallel {p) to the field. By means of these a better estimate of the 

 relative intensities of the components could be formed, and the question of 

 the dissymmetrical separation of the components of some of the lines could 

 be studied. The dependence of the relative intensities of the components on 

 the state of polarization of the vibrations when falling upon the grating was 

 first pointed out by Zeeman,* and in my Dissertation (/. c.) the effect of the 

 quartz lens used in these experiments on the state of polarization, and hence 

 on the intensities of the components of all the lines in the spectrum, was 

 shown. On account of this the observation of the relative intensities under 

 such conditions is of less importance in some respects than it was formerly 

 supposed to be; but in other respects it may be of much importance. If, 

 however, the spectrum lines are not very far apart, the relative intensities of 

 their components may be used in comparing, since in a short distance the 

 influence of the rotation is small, but for lines further apart a comparison 

 can only be made between the mean intensities of the components for each 

 line. 



As in the cases of tungsten and molybdenum, the strength of the magnetic 

 field was determined from the separations of several lines whose intervals are 

 well established. The field was found to be 24,500 C.Gr.S. units. T'o facilitate 

 a comparison of che results with those for the above-mentioned substances, 

 the measurements were reduced to a field of 23,850 C.G-.S. units. The plates 

 were measured with a dividing instrument by Brunner of Paris, which is in 

 the Observatory, Meudou, Paris. The majority of the Imes were measured 

 in the second order, but very many of these were again measui-ed in the first 

 and tliird orders. I'he intensities were all reduced to those of the second 

 order. With the dividing instrument one could measure to '001 mm. Owing 

 to such causes as the variation in sharpness of the different lines, the 



* P. Zeemiin, K. AU.id. v. "Wet., Amsterdam, Oct., 1907. 



[7*] 



