420 MR WILLIAM SWAN ON THE PRISMATIC SPECTRA OF THE 



imperceptible, shines out in great brilliancy; while the lines /3, 7, 8, and £ remain 

 totally unchanged in position, colour, and intensity. 



While the line a is thus exceedingly variable in its brightness, the lines /3, 7, 8, 

 and £, on the other hand, are perfectly steady ; and being never absent in 

 carbohydrogen spectra, there is every reason to believe that they are really 

 characteristic of the body undergoing combustion. Beyond a on the less refracted 

 side there is a faint trace of red light, which, as it becomes so feeble as almost to 

 disappear when the light is derived from the lowest point of the flame of the 

 Bunsen lamp, is probably due to the exterior envelope of the flame, and not to 

 the interior cone. The line a is separated from (3 by an extremely dark space, 

 almost destitute of light. The line (3 is of a faint yellowish green colour, but well 

 defined, and is accompanied by four almost equidistant lines /?, (3 2 , &c, which 

 diminish in brightness as their distance from (3 increases. After another very 

 dark interval, the extremely beautiful line 7 follows, which is exceedingly bril- 

 liant, and of such absolutely definite refrangibility as, like a, to form a perfectly 

 sharp image of the slit through which the light passes. Its colour is a fine 

 slightly bluish or pea green, and it is accompanied by a fainter line 7, . The 

 next line 8 is the less refracted edge of a broad band of light containing four 

 fine lines. This group, which is of a pale ashy colour, is separated by dark in- 

 tervals from 7 and 5. The line £ belongs to a brilliant but not very well defined 

 band of a fine purple tint, which is accompanied by a fainter line f . 



I have completed observations of the minimum deviations for the lines a, /?, 7, 

 8, and £ ; and also for the principal lines of the solar spectrum, which are given in 

 Series 1, Tables II. and III., pp. 427, 428. From an examination of these tables it 

 appears, that while several lines in the carbohydrogen spectrum coincide nearly in 

 position with remarkable lines in the solar spectrum ; yet in no case, if we except 

 the line a, has the observed coincidence been exact. The observations, therefore, 

 rather tend to prove that the bright lines of the carbohydrogen spectrum coincide, 

 not with the dark lines, but with the bright spaces of the spectrum of sun light. 



Postscript added since the preceding Paper was read* 



From the well known coincidence discovered by Feaunhofee, to exist between 

 the line R, in the spectrum of a lamp, and D of the solar spectrum, taken in con- 

 nection with similar phenomena, which have since been observed, it might be 

 inferred, as a general law of the spectra of flames, that their bright lines always 

 coincide with dark lines of the solar spectrum. 



The result of the investigations which have now been detailed, is obviously 

 unfavourable to such a conclusion. In publishing observations bearing on a ques- 

 tion of so much interest and importance, I was anxious, if possible, to leave no 



* Printed by permission of the Council. 



I 



