NATURAL PHILOSOPHY. 129 



All the substances which give colored flames, give also colored images 

 by the light. If we take any of the substances which do not give color 

 to the flame, we do not obtain colored images by the light ; we pro- 

 duce upon the plate a negative image, composed merely of black and 

 white, as in the ordinary photographs. Those substances which give 

 white flames, as the chlorides of antimony, lead and zinc, yield no color 

 by luminous action. All the colors of the picture have been produced 

 by preparing a bath composed of the deuto-chloride of copper ; this 

 salt thrown into burning alcohol produces a variegated flame, accord- 

 ing to the intensity of the fire ; and it is nearly the same with all the 

 salts of copper mixed with chlorine. If we put a salt of copper in 

 chlorine liquid, we obtain a very sensitive surface by a single immer- 

 sion ; but the result of this mixture is seldom good. I prefer taking the 

 deuto-chloride of copper, to which I add three or four pounds of water ; 

 this bath gives good results. A mixture of equal parts of chloride of 

 copper and chloride of iron, with three or four parts of water, is, how- 

 ever, the best. The chloride of iron has, like that of copper, the prop- 

 erty of being impressed on the plate of silver, and of producing many 

 colors ; but they are infinitely more feeble, and the yellow always pre- 

 dominates ; this agrees with the yellow color produced in flame by 

 chloride of iron. If we form a bath, composed of all the substances 

 which separately give a dominant color, we obtain very lively colors ; 

 but the great difficulty is the mixing in proper proportions, for it 

 happens, nearly always, that some colors are found excluded by others. 

 By care, however, we ought to arrive at the reproduction of all the 

 colors. There exist many difficulties, more indeed than in any of the 

 ordinary processes of photography. We cannot always depend upon 

 obtaining the same results with the same materials, owing principally 

 to the difficulty of preserving the solution at a uniform strength. 

 Liquid chlorine is necessary ; the application of dry chlorine will not 

 produce the same result. The action of heat upon these prepared 

 plates is, in some respects, analogous to the effects of light. By warm- 

 ing a plate over a spirit-lamp, we produce successfully the following 

 tints : brown red, a cerise red, scarlet, and red having a whitish tint. 

 Numerous experiments have been made by M. Niepce to produce the 

 colors upon the salts of silver and copper spread on paper, but hitherto 

 without success ; a metallic plate of silver the plated copper answers 

 must be employed. Iodine and bromine, and their salts, have been 

 tried, but they will not produce a surface capable of developing colors. 

 Chlorine, in the state of chlorates or chlorides, is the only substance 

 which possesses the property of being colored by light, when chemi- 

 cally combined with metallic silver. 



The mode of operating recommended is, to form a bath with one 

 fourth by weight of the chloride, and three fourths of water. When 

 the muriatic acid is used with a salt of copper, we must add one tenth 

 of water. When the bath is composed of several substances, it is 

 essential to filter the solution carefully, so as to obtain very transparent 

 solutions, and it must be preserved in a well-stoppered bottle. The 

 quantity necessary to prepare two or more plates should always be 

 taken, because the bath is weakened considerably at each operation ; 



