462 Proceedings of the Royal Society 
turned his attention to the spectrum of the stars, of which he 
described several. He likewise described the spectrum of electric 
light, but only that of sparks passing through atmospheric air. 
He has likewise left on record a very curious observation on the 
spectrum presented by the exterior flame of a wax candle. When 
the bright flame is intercepted by a screen, and only the faint ex- 
terior flame viewed, he found it to consist almost entirely of homo- 
geneous yellow light ; but his skill as an observer was so great that 
he perceived this yellow light to consist of two distinct rays very 
close together, and only separable by an excellent prism, and a 
very narrow aperture. As he remembered that there was a similar 
double ray in the yellow part of the solar spectrum which he had 
named D, the happy thought occurred to him of transmitting solar 
light through the same aperture. He did so, and found that the 
two rays of the line D coincided most accurately with the double 
yellow ray given by the exterior flame of a wax candle. He does 
not appear to have prosecuted this interesting research further. 
He merely records the fact. He was not aware that the yellow 
light of the candle was in any way caused by the presence of 
sodium , the existence of which in a wax candle would probably not 
occur to any one, unless perhaps to an experienced chemist on the 
look out for some extraneous substance. 
About the same time Sir D. Brewster had been seeking for a 
source of homogeneous light, for the purpose of improving the 
microscope by destroying all chromatic aberration of the lenses. 
See his paper of 1822 in the Transactions of the Boyal Society of 
Edinburgh, vol. ix. p. 433. Although acquainted with the effect of 
salt on the flame of burning alcohol, he had evidently only cursorily 
examined it, since he says “ salt or nitre f which is incorrect, and 
speaks of its causing the flame to yield “ insalubrious vapours.” 
He therefore rejects the use of it, and merely recommends that the 
alcohol should be “ largely diluted with water.” The yellow light 
so obtained he refers to “ imperfect combustion” (p. 435), and not 
in any way to sodium , observing that the combustion of paper, 
linen, cotton, or the flame of a blow-pipe, also contain the same 
homogeneous yellow light in tolerable abundance. His observa- 
tions, therefore, have a certain resemblance to those of Fraunhofer. 
About the year 1824 or 1825, Dr Wollaston gave one of his 
