Wave-lengths in the Invisible Prismatic Spectrum. 201 



eye in the case of the visible, with the bolometer in that of 

 the invisible ray. 



To illustrate the use of the above-described apparatus under 

 somewhat unfavourable circumstances, let us consider as an 

 example the observations of June 13, 1882, which were taken 

 far down in the spectrum, where the heat is feeble and the 

 galvanometer-deflection small, requiring a widely open slit. 

 The apparatus having been previously adjusted, and the sun- 

 light properly directed by the siderostat, the visible Fraun- 

 hofer line J) l of the third spectrum of the grating was caused to 

 fall upon the slit S 2 of the spectro-bolometer. Then, according 

 to the theory of the grating, there passed through this slit 

 rays having the wave-lengths — 



0589 (3rd spectrum — visible), 

 1'178 (2nd spectrum — invisible), 

 1*767 (1st spectrum — invisible). 



The prism having been removed and the telescope brought 

 into line, an image of S 2 , of the same size as the slit itself, 

 was formed in the focus of the object-lens ; and on testing 

 with the bolometer, whose face was covered with a screen 

 pierced centrally with a 2-millim. slit, the heat of this image 

 produced a deflection of the galvanometer-needle of about 

 thirty divisions. The prism was then replaced on the auto- 

 matic holder and set to minimum deviation, and the image of 

 the slit containing superposed rays, whose combined effect had 

 produced the deflection just mentioned, was separated into 

 three similar images* (fig. 2), each composed of nearly 



Fiff. 2. 



homogeneous rays. Of these three bands, only the first or 

 most refrangible, containing the D x line, was visible, and its 

 deviation w T as found to be 47° 41 7 , agreeing with the value 

 given by the table. It was the object of the experiment to 

 find the place of the lower invisible band, by groping for it ; 



* These three images, being composed of rays of three different wave- 

 lengths, could not all be in the same focus of the same lens at the same 

 time, since the collimator and objective of this spectrometer were simple 

 lenses. The lenses were adjusted by means of a table of focal distauces 

 previously prepared, so as to throw a sharp (invisible) image of the band 

 to be detected. 



