( 353 ) 



Our object of investigation is tlie relation between asymmetry and 

 strength tit' Held. 



The measurements were made in the following way. The negative 

 was placed on the comparator, in such a manner, that the middle 

 line of one of the triplets was contained between the two parallel 

 wires in the reading microscope. The parallel wires had been placed 

 previously at right angles to the direction of motion of the negative. 

 It appeared that if with one of the triplets the desired coincidence 

 had been obtained, this was also the case with the other. An extra 

 system of cross wires, crossing under an angle of about 50°, was 

 used in the measurements and made it possible to determine the 

 resolution in the selected point of the lines. 



The resolution of one line having been measured for a definite 

 value of the magnetic force, the corresponding resolution in the 

 corresponding point of the second line was determined immediately 

 afterwards. 



The line 5770 appeared to be divided almost exactly symmetri- 

 cally, so thai the resolution could be taken as a measure of the 

 magnetic force. 



On the obtained negatives 34 series of measurements were made. 

 They relate to different points of 10 negatives made at different times. 



The vacuum tubes used were intentionally made somewhat dissimilar. 



In order to control the results the negatives were taken with 

 different maximum intensity of field. 



Finally the negatives obtained can be distributed into two groups, 

 differing by the position of the grating. After taking 7 negatives I 

 resolved to rotate the grating in its own plane through 180' in order 

 to see whether this had some influence on the asymmetry. 



This appeared to be not the case, but the apparent distribution of 

 intensities changed in a remarkable manner. Whereas in one position 

 of the grating figures 1 — 5 were obtained, the middle line being- 

 strong and the outer components rather weak, the distribution of 

 light after rotation became reversed. In this position of the grating 

 Fig. 1 of my last paper was taken. (See these Proceedings October 

 1907). The middle line is very weak and the outer components 

 predominate. 



5. The results of the measurements are dealed with in the following 

 way. The amounts of separation of line 5791 towards the red and 

 towards the violet are supposed to be functions of the separation 

 of line 5770, which may be treated as proportional to the magnetic 

 force. The separations of line 5770 may be taken as abscissae, the 

 two other separations as ordinates. 



