128 



rof . N 



agaoka and Mi 



\ T. Takamine on Crossed Sjiectra 





(Centre displaced +14). 







r+212 





+ 226* 



— 





+ 168 





+ 182 



+ 182 





+ 130 





+ 144* 



— 





+ 91 





+ 105 



+ 105 





f 29 





+ 43 



+ 43 











— 



+ 26** 



Oi 



— 





— 



+ 17** 



cSH 



o 













-41 



-34 





- 20 



- 23 



- 48** 





-111 





- 97 



- 97 





-124 





-110 



-110 





-174 





-160 



-163 





-219 





-205* 



— . 





-2S2 





-268* 



_ 



Wood's values agree tolerably veil with ours for satellites 

 of two lines 5790 and 5769 ; the presence of the lines —938 

 and —999 in the former has been confirmed in our case by 

 arranging the interference points on parabolas before 

 mentioned : further, the existence of the second satellite on 

 the negative side of the latter can hardly be doubted. As 

 to the satellites of the violet line 4359, we have to remark 

 that the position of the principal line is rather doubtful. 

 There is a satellite very close to the principal and another 

 not far from it on the positive side. In order to bring 

 Wood's measurement in harmony with ours, it is necessary 

 to displace the centre towards the negative side by about 

 14 m.A.U. ; by so doing the numbers in the second column 

 were obtained, agreeing quite well with ours. We missed 

 the lines marked with an asterisk, and Wood those with double 

 asterisks. This must be ascribed to the high resolving 

 power of echelon and plate, while the new lines found by 

 Wood must be attributed to the large range covered by the 

 concave grating. Jt is, however, questionable how ghosts 

 and false lines in the spectra formed by a concave grating 

 can be discriminated from true lines without some means of 

 crossing them. We have no doubt that there is a group of 

 satellites near the principal line as shown in PI. I. fig. 1, 

 in which the principal line appears on both sides, and the 

 branches of the parabola on which the interference points lio 

 run oblique to the echelon spectra, which are vertical. The 

 arrangement of the satellites is shown in the diagram 



(fig- !)• 



We believe that by crossing the concave-grating spectrum 



