GASEOUS SPECTRA UNDER HIGH DISPERSION. 417 



or definition to bring out many of the exquisite and close-set details which 

 spectroscopy is beginning to demand in the present day for any and every gas 

 that is now observed. And this, too, although those Upsala results did most 

 honourably transcend all others at the time they were published, and for 

 many years afterwards, indeed almost, if not quite, up to the present time. 



Hence I have felt it incumbent on myself, before venturing to trouble this 

 Society with new versions of any old phenomena, not only to increase the range 

 of subjects treated of, but to improve the instrumental apparatus employed on 

 all of them, until it was capable of some very remarkable performances in the 

 way at least of differential mensurations in the field of view or near it. 



Of such differences, however, only. For temperature changes in the fluid 

 prisms, although greatly reduced by a variety of contrivances, were yet continu- 

 ally at work, altering to so sensible an extent the value of the Dispersion scale 

 from its distant zero, as to prevent the absolute places given by this very 

 lately put-together apparatus being anything better than extremely rough when 

 over long ranges.* 



But condoning that one weakness, for the sake of other advantages, the 

 strength of the instrument for micrometrical detection and record of small 

 differences may be indicated thus — 



(A) The one-prism spectroscope employed in my collective and rudimentary 

 paper of 1880, possessed a little more than 3° of Dispersion from A to H, with 

 a magnifying power on the telescope of 10 diameters ; virtually broadening 

 those 3° to 30°. 



(B) The apparatus employed by MM. Angstrom and Thalen for their 

 Mesures Micrometriques, as nearly as I can gather in a general way, must have 

 had at the utmost a Dispersion A to H of 24°, with a magnifying power of 

 nearly the same number of diameters ; or equivalent virtually to a spectrum 

 600° long. 



(C) While my present arrangement has 60° of Dispersion A to H, with 

 magnifying powers on the telescope of inspection rising from 12 to 36, and a 

 further mechanical magnifying in the recording apparatus of 5 times ; 

 equivalent to 9000 of the same degrees altogether ; or to the action of 1800 

 simple prisms of 60° refracting angle in white flint-glass, viewed with the 

 naked eye.f 



* By observations on Green CO on November 3, 4, and 5, it was ascertained that, after every- 

 thing had been done at the place that could well be done to secure constancy of temperature in the 

 bisulphide of carbon prisms employed, a slow fall of 1° Fahr. increased the Dispersive power of the 

 collective train of prisms by 0'727 of a Eevolution of the very coarse screw motion ; or by close upon 

 10 inches on the surface of the recording barrel. The interval of time therefore between any two 

 mutually dependent observations was, after that, made as short as possible. 



|L The diameter of the objectives was the same as in my. earlier apparatus, viz., 2*25 inches, and 

 the refracting faces of the prisms presented nearly the same breadth to the entering ray. 



