MEMOIR XIV.] THE PKOCESS OF DAGUERREOTYPE. 207 



does not yield to theory : on one point alone there is ob- 

 scurity. Why does the vapor of mercury condense in a 

 white form on those portions of the film of iodide which 

 have been exposed to the influence of light ? condense 

 to an amount almost proportional to the quantity of 

 incident light? 



Even on this point there are facts which appear to 

 have a bearing. 



(a.) It has long been known that if a piece of soap- 

 stone or agalmatolite be made use of as a pencil to write 

 with on glass, though the letters that may have been 

 formed are invisible, and though the surface of the glass 

 may subsequently have been well cleaned, yet they will 

 come into view as soon as the glass is breathed on. 



(&.) I have often noticed that if a piece of very clear 

 and cool glass, or, what is better, a cold polished metallic 

 reflector, has a little object such as a piece of metal laid 

 upon it, and the surface be breathed over once, the object 

 being then carefully removed, as often as you breathe 

 again on the surface a spectral image of it may be seen, 

 and this singular phenomenon may be exhibited for 

 many days after the first trial was made. 



(<?.) Again, in the common experiment of engraving on 

 glass by hydrofluoric acid, if the vapor has been very 

 weak, no traces will be perceived on the glass after the 

 wax has been removed ; but on breathing over it, the 

 moisture condenses in such a way as to bring the object 

 into view. 



(d.) In a former Memoir (XIH.) I described a phe- 

 nomenon relating to the crystallization of camphor on 

 surfaces of dry glass, on which invisible traces have 

 been made by the pressure of a glass rod ; this also ap- 

 pears to belong to the same class of effects. 



Berzelius (jBvz&?,VoL II., p. 186) has attempted to ex- 

 plain (a) and (<?) on this principle, that the changed and 



