COLOURS OF JIEATED METALS. 135 



oxidation, because they have supposed they observed a great 

 similarity between the appearances in question, and those of 

 several metals placed in circumstances under which they are 

 actually oxided. This subject however deserves at least a 

 closer examination, and the following is the simple method 

 I have pursued. 



1 held a steel watchspring across the flame of a candle a A steel watch- 

 few seconds in a fixed position. After it was cold and clean- ^vThe flame of 

 ed", I found both on the right and left of the central point, a'candle 

 where the flame had been, a series of colours more and more acquired the 

 faint [degradees], with periodical recurrences, such as would P"sniatic co- 

 have been exhibited by a small band cut precisely from the ly in concen- 

 middle of a circle formed of a series of concentric coloured ^"*^ ""^* 

 rings. The nature of the phenomenon then is very dis- 

 tinctly shown here, particularly as the exterior ring was 

 nearly 3 cent. [1 1*7 lines] in diameter, and the others de- 

 creased interiorly with intervals of a few millimetres. No- 

 thing was wanting to have completed the circles, but to have 

 operated on a bioad plate of steel suspended horizontally 

 over the point of the flame. 



Not having such a plate at hand, I took a sheet of tin, A sheet of tin 

 which, with appearances analogous to the preceding, af- 

 forded very vivid colours in consequence of its natural 

 whiteness.' With a proper degree of inclination the colours 

 are most lively ; particularly tlie yellow, I'ed, and blue, which 

 form together a spot, in which the blue occupies the centre, 

 surrounded by the red, and beyond this with the yellow, with 

 the intermediate tints and gradations. 



There is nothing in the property tin has of being oxided Not fromoxi- 

 and forming salts, that indicates colours corresponding with "^^''^'^' 

 these : on the other hand the periodical recurrences on the 

 steel spring evidently belongs to a series of rings : we must 

 conclude therefore, that this phenomenon is simply of the 

 class of coloured rings. 



Another trial made with a gold ring equally produced re- q i j ^u 

 peated traces of rings, and here the suspicion of oxidation 

 will hold still less. 



A copper wire gave me similar indications, though more Copper and 

 faintly; but I have observed them very striking on copper 

 chimneys of stoves. Lead that has just been melted ex hi- lead similar. 



bits 



shows this 

 more plainly. 



