. C 53 ] 
find one fingle grain of metal in the remaining con* 
tents. 
This difappointment greatly puzzled me, till hav- 
ing thoroughly examined into the unexpedted event,' 
without being able to difcover any reafon fufficient 
to incline me to recede from my former opinion, as- 
to the component parts of the fand, I concluded that 
the flux might poffibly be a very improper one; for 
though it might have effedted the intended feparation, 
yet it might at the fame time be fufficiently powerful 
to divide the particles of the metal, when feparated, 
fo very minutely, as to be capable of fubliming and 
carrying them off imperceptibly : And finding the 
contents greatly diminifhed, fo that the quantity re- 
maining bore but a fmall proportion to that which 
was firfb put into the crucible, I concluded that thi3 
muft really have been the cafe, and that fo'me very 
different method muft be purfued in order to produce 
the defired effedts. I immediately determined to make 
a fecond trial, in which I proceeded in the following 
manner. I took the fame quantity of fand made 
ufe of in the former experiment ; and firft I fpread it, 
without any addition to it, upon an iron plate over a 
flrong fire, where I gave it a very powerful torrifica- 
tion (or roafling) to try if, by that means, I could not 
relax, and loolen the component parts to fuch a degree, 
as to make the feparation and redudtion of the me- 
tal more eafy, when I fhould bring it into the furnace. 
When I had fo done, I mixed it up with a flux of a 
very peculiar, but gentle nature, which I had before 
made ufe of for other purpofes with great fuccefs, and 
committed it (as in the former experiment) to the fur- 
nace, where I urged it by a very flrong fire for about 
i three 
