Mr. Talbot's Facts relati?ig to Optical Science. No. III. 3 



duced by this cause acting more energetically at a high tem- 

 perature. 



On Prismatic Spectra. 

 It is much to be desired that an extensive course of experi- 

 ments should be made on the spectra of chemical flames, ac- 

 companied with accurate measurements of the relative posi- 

 tion of the bright and dark lines, or maxima and minima 

 ,bf light which are generally seen in them. The definite rays 

 emitted by certain substances, as, for example, the yellow rays 

 of the salts of soda, possess a fixed and invariable character, 

 which is analogous in some measure to the fixed proportion 

 in which all bodies combine, according to the atomic theory. 

 It may be expected, therefore, that optical researches, care- 

 fully conducted, may throw some additional light upon che- 

 Inistry. Some experiments which I formerly made upon this 

 subject will be found in Brewster's Journal for 1826*. In 

 addition to the substances there enumerated as giving a pecu- 

 liar optical character to flame, I have found that the salts of 

 copper are exceedingly remarkable. They give spectra so 

 covered throughout with dark lines as to resemble in that re- 

 spect the solar spectrum. The flames of boracic acid and 

 nitrate of barytes also possess somewhat of a similar character. 

 The most convenient way of obtaining brilliant spectra of 

 these substances is to deflagrate them with chlorate of potash, 

 but this is attended with the inconvenience that the spectrum 

 produced by the chlorate is seen in conjunction with the other, 

 and an allowance must be made accordingly. Another good 

 method is to sprinkle the substance in powder on the wick of 

 a spirit-lamp, and direct a current of oxygen upon it. With 

 regard to the accurate measurement of the lines, it requires 

 the use of very superior apparatus. I have sometimes made 

 approximate measurements by fixing a divided scale trans- 

 versely to the linear aperture through which the light of the 

 burning body was observed. This aperture was then ex- 

 panded by the prism into a spectrum parallel to the scale, by 

 means of which it could then be measured. An objection 

 may, perhaps, occur to the reader, that the scale would thus 

 be as much refracted as the light itself, and therefore could 

 not serve to measure it. But this difficulty was avoided by 

 a simple contrivance, viz. by illuminating the scale with ho- 

 mogeneous light. 



Spectra of various Galvanic Flames. 

 Silver leaf tleflagrated by galvanism gave a spectrum wilh 



* Vol. V. p. 77- Sec also the present .louinal, vol. iv. p. 114. 



B2 



