274 Lord Rayleigfr's Investigations in Optics. 



barely resolved. A soda-flame was generally used a though 

 similar observations have been made upon the D line of the 

 solar spectrum. When the adjustment was complete, the aper- 

 ture along the face of the prism was measured, and gave at 

 once the equivalent thickness, i. e. the difference of thicknesses 

 traversed by the extreme rays, since the prism was in the posi- 

 tion of minimum deviation. Care, of course, was taken that 

 no ordinary opiical imperfections of the apparatus interfered 

 with the experiment. 



One observer, familiar with astronomical work, fixed the 

 point of resolution when the thickness amounted to from 1*00 

 to 1*20 centimetre. I was myself less easily satisfied, requi- 

 ring from 1'35 to 1*40 centimetre. But even with a less 

 thickness than 1 centimetre, it was evident that the object 

 under examination was not a single line. With the same 

 prism I found the thickness necessary to resolve b 3 b 4 in the 

 solar spectrum to be about 2*5 centimetres. According to (11), 

 the thickness required for b 3 b 4 should be 2*2 times that 

 required for D L D 2 - Probably something depends upon the 

 relative intensities of the component lines. 



From (1) and (11) we see that if a diffraction and a dis- 

 persion instrument have equal resolving-powers, 



**$: ...... (13) 



so that the power of a dispersion instrument relatively to that 

 of a diffraction instrument varies inversely as the third power 

 of the wave-length. 



For the kind of glass considered in (10), and for the region 

 of the D lines, 



^■^^o < u > 



To find what thickness is necessary to rival the fourth spec- 

 trum of a grating of 3000 lines, we have merely to put m = 4, 

 ?i = 3000j so that the necessary thickness is about 12^ centi- 

 metres — a result which abundantly explains the observations 

 which led me to calculate the power of prisms. 



Terling Place, 

 August 12, 1870. 



[To be continued.] 



