NATURAL PHILOSOPHY. 169 



* 



of one part of chloride of iron with four of a salt of copper in 300 

 parts of water, all the colors will be obtained, with white grounds, 

 but they are not very bright. If a mixture is made of 100 parts of 

 chloride of magnesium with 50 parts of sulphate of copper, all the 

 colors will be copied, and they will be brighter than the preceding, 

 but the ground will always be dark or pink. 



I shall now speak of the experiments that I have made on colored 

 flames, and on the relations of the color of these flames with the 

 colors that are developed on the silver plate prepared with the bodies 

 that give colors to these same flames. 



It has been observed that chlorine alone gave to a silver plate the 

 property of being differently colored when exposed to the influence 

 of the light. It was therefore of importance, according to my idea, 

 to dye with it alone flames of all the colors, which I did by means of 

 the following experiments. If there is placed in pure alcohol a small 

 quantity of pure hydro-chloric acid, there will be obtained on burn- 

 ing it, at first a yellow flame ; then, if there be gradually added new 

 hydro-chloric acid, by shaking the liquor in a capsule, there will be 

 successively obtained flames of all the colors of the spectrum, begin- 

 ning from the yellow ray to the violet, which will be produced by the 

 greatest quantity of hydro-chloric acid that can be placed in the 

 alcohol without extinguishing it ; only I premise that the flames are 

 not very brilliant. They will be rather more so if there be substi- 

 tuted for the hydro-chloric acid a chloric ether, or Holland gin, and 

 preferably, the chlorides of carbon, particularly the sesqui-chloride, 

 which has given me colored flames from the yellow ray to the violet ; 

 but I have not been able to obtain red and orange-colored flame. 

 That arises probably because the heat was not sufficiently strong, or 

 because I did not employ sufficient chloride. The proto-chloride of 

 carbon has given me flames of a rather^ great intensity of color, by 

 exposing it on a sheet of platinum, or by saturating a stout match of 

 asbestos, heated by the flame of an eolipile lamp. I have by this 

 means obtained violet flames of a deepish tinge, but nothing further. 

 This arises, no doubt, from the great volatility of the proto-chloride 

 of carbon, which does not allow of maintaining it at a sufficiently 

 high temperature to produce red and orange rays, as can be done 

 with the chloride of copper. It resists, however, much better the 

 action of fire than the sesqui-chloride of carbon, which, being treated 

 in the same manner, immediately volatilizes, producing only a yellow 

 flame. 



Having formed a chloride of silver with nitrate of silver and 

 chloride of magnesium, which, having been previously well washed, 

 was afterwards exposed to the flame of the eolipile lamp, there was 

 produced a fine yellow flame. But to obtain flames of all the colors 

 of the spectrum* with a great depth of color, a chloride of copper 

 must be formed, by mixing hydro-chloric acid (or a chloride whose 

 base does not give coloring to flame) with a salt of copper, or prefer- 

 ably, by adding one of these two things to the per-chloride of cop- 

 per. Colored flames of all the rays of the spectrum can be by this 



