Memoir upon Ultra-marine. 359 



mixtures made use of in my experiments ; suffice it to say, 

 that lime and dung, properly blended and fermented, proved 

 to be by far the most fertilizing of all that I tried. 



Having ascertained this fact, I was desirous to know 

 what proportion of lime was best to be mixed with the 

 dung. I found, after various trials, that the best was one- 

 fourth by weight ; the largest quantity of nitre being pro- 

 duced when the two were mixed in this proportion. 



To hasten the process of fermentation, I made my expe- 

 riments in a hot-house, kept at a proper temperature, 

 through which a current of air was allowed to pass. When 

 the fermentation was over I analysed the residue; and I con- 

 stantly found that part of the lime had disappeared, and 

 that the nitre formed was always nearly in the proportion of 

 two and one-twelfth to one of the lime missing. 



From the results of my experiments I was led to con- 

 clude, that potass takes about two-thirds of lime into its 

 composition. The remaining one- third I suspect is oxy- 

 gen, for reasons of which I will inform you in my next, 

 in which you will receive an accurate detail of my experi- 

 ments. 



LVIII. Memoir upon Ultra-marine. By Messrs. Desormes 

 and Clement. Read in the French Institute January 27, 

 1806*. 



JL he superb blue colour known by the name of ultra-ma- 

 rine has not as yet been an object of chemical inquiry: 

 chemists have hitherto only made experiments upon lapis 

 lazuli, which is, in some measure, the ore of ultra-marine, 

 and it is never found in the state of a crystal ; one specimen 

 only excepted, in the possession of M. Guy ton. 



A process is employed to extract ultra-marine from its ma- 

 trix which seems to have nothing analogous, and of the theory 

 upon which it is founded we arc entirely ignorant : it con- 

 sists in mixing well, pulverised lazuli, with a melted mastic, 

 composed of resinous pitch, wax, and linseed oil. When 



• From the Annolti <'«• Chlmit, torn. Ivii. p. 317. 



A a 4 this 



