MEMOIR XVI.] CONDITION OF A DAGUERREOTYPE SURFACE. 927 



parts, that is to say, the iodide is neither upon nor beneath 

 the mercury, but both are, as it were, in the same plane. 



Soon after I had ascertained the action of gum-arabic, 

 some of it was applied to the surface of a plate on which 

 an impression had just been formed in the mercury-bath. 

 This was without removing the coat of iodine. On dry- 

 ing it, the gum chipped up, as was expected, bringing 

 away with it all the lights of the picture, and leaving 

 a uniform coat of yellow iodide of silver beneath. It 

 seerns, therefore, that the film of iodide coheres more 

 strongly to the metal plate than the amalgam ; and, far- 

 ther, from this result we should judge that the amalgam 

 is on the surface of the iodide. 



But this is not true; for on three different occasions 

 I found that when Russian isinglass was employed in- 

 stead of gum for the purposes presently to be related, 

 the isinglass, from its stronger cohesive power, chipped 

 off in the act of drying, tearing up the yellow film from 

 end to end of the plate, and leaving the amalgam con- 

 stituting the lights undisturbed. It is here to be under- 

 stood that this action takes place without the smallest 

 disturbance of the lights and demi-tints, the plate remain- 

 ing in all the beauty and brilliancy and perfection that 

 it would have had if it had been carefully washed in 

 hyposulphite of soda. 



This is a result, however, which cannot be produced 

 with uniformity. Most commonly the lights are torn 

 up with the iodide. Had it occurred but once, I should 

 still have cited it with decision, for from the very char- 

 acter of it, it is impossible to be mistaken or to commit 

 an error of judgment. It proves that the film of iodide 

 may be mechanically torn off from the metallic surface 

 as perfectly as it can be dissolved off by chemical agents 

 a singular fact. 



This result, therefore, proving that we can tear off 



