-C:::: 



134 Prof. Nagaoka and Mr. T. Takamine on Crossed Spectra 



points. The accompanying figure gives the sketch of how 



the rings and the lines are superposed on coincidence. The 



coordinates may be utilized for de- F" 3 



ducing the position of the centre 



by the geometrical property of £ 5 



circle. A more practical way of ^ 



finding it is to photograph a small 



portion of the system of rings after 



removing the echelon ; this can be 



done in the arrangement of the 



apparatus constructed by Hilger. 



The method of procedure is as 

 follows : — The apparatus is en- 

 closed in a wooden box lined with 

 thick cork plates, to avoid changes 

 of temperature. The photographs 

 of crossed spectra for the reference 

 line and for the line in question 

 are taken, the temperature and 

 pressure of the air in the box being- 

 noted. Sometimes it will be neces- 

 sary to note the hygrometric state 

 also. Then x- and ^/-coordinates, of 

 interference points, are measured 

 hj a delicate micrometer on the 

 photographs. The position of the — j i + gj 



centre of the rings is determined 3 £ £ 3" 



from these values ; thus O's are -g ■%' 



easily calculated, and consequently ^ £* 



e i € , rr 1 • l Echelon and Fabry-Perot Plate, 



m m' ar6 ktlOWn - 1W0 tnal Coincidence of the Satellites 



m . -, -f 123 and —242. 



experiments are suthcient to make 



— differ by less than one per cent, from — , . As the tem- 

 m J L in 



perature is generally changing slowly, it is necessary to 

 measure e before and after the experiment. 



We give an example of our determination of S\ made on 

 the yellow lines of mercury and another on two satellites of the 

 green line. As a reference line, we assumed the wave-length 

 of the yellow line 5790'6593 according to Fabry and Perot, 

 and that of the neighbouring line to be about 5769*698. 

 These lines were photographed in crossed spectra of echelon 

 and plate as in PI. I. fig. 4. The temperature of the air 

 was determined by means of Assmann's aspiration thermo- 

 meter, which is indispensable for accurate works with air- 

 plate, as a thermometer simply hung in the neighbourhood 



