Crossed Spectra hy Combinations of Interferometers. 127 

 Plate : length 20 cm., breadth 3'5 cm., thickness 1*0912 cm. 



SXmax. =120-9 m.A.U. for 5790 

 = 120-0 „ „ 5769 



= 106-6 „ „ 5461. 



We have published in another place * our results con- 

 cerning the structure o£ mercury lines obtained by crossing 

 the echelon-grating spectrum with that given by the plate, so 

 that it will be unnecessary to reproduce them here. When 

 only one instrument is used, there is much difficulty in 

 interpreting the results owing to overlapping of several 

 orders of spectra, and to the appearance of ghosts. These 

 two points are the chief objections raised by Wood | against 

 the use of echelon or plate, but the difficulties can be over- 

 come by simply crossing the instruments. 



As regards the elimination of ghosts, the discussion given 

 by Gehrcke and v. Baeyer for a combination of plates 

 applies equally to that of echelon and plate. The ambiguity 

 as to the orders of spectra is easy to settle, as the interference 

 points belonging to the same order of spectrum lie on a 

 parabola, whose position and dimension are easily determined. 

 The simple process of arranging the points on a section 

 paper is sufficient to decide this puzzling question, which can 

 hardly be answered when an interferometer is used singly. 



It will not be out of place to compare Wood's measure- 

 ments of the satellites of mercury lines observed with a 

 concave grating with our results made on crossed spectra. 



o 



Wood 





Nagao 



fa and Takamine 



f + 222 m. 



A.U. 





+ 224 



4- 164 







+ 164 



+ 131 







+ 130 



o 











- 124 







-119 



- 191 







-191 



- 944 







-938 



1^-1007 







-999 



f + 42 







+ 44 















- 55 







- 50 



1 -112 







-121 



* Nagaoka and Takamine, Proc. Phys. Soc. xxv. p. 1 (191:?). 

 + Wood, Phil. Mag. xxv. p. 443 (1913). 



