230 
SELECTED ARTICLES. 
concentrated and the more the empyreiimatic oil becomes 
acrid and disagreeable, and that at last a fourth product is 
added to the preceding; this consists of long filiments which 
attach themselves to the sides less heated; a little warm water 
easily removes them, and the filtered solution, upon eva- 
poration, yields crystals of caffein. It is, probably, on this ac- 
count that some have thought that benzoic acid was disengaged 
during torrefaction. M. Labillardiere, an old member of the 
Academy, has informed one of us that he has collected these 
crystals by means of a cone of paper placed over burning 
cofiee. We have repeated this experiment by substituting a 
glass vessel for the paper cone, and we have noticed it covered 
with long delicate arborizations presenting the characters of 
caffein. 
Not having gained any thing by this procedure, we direct- 
ed our researches to roasted coffee, for here at least, thought 
we, it is very certain that the active principle must be found 
developed, — since it is prepared thus that it produces so great 
an efiect upon the animal economy. But all our efforts were 
useless; the burnt coffee preserved its secret as completely 
as the green. It is true, however, that ether separated from 
it an abundant brown fixed oil highly charged with its aroma- 
tic principle, but nothing could isolate the latter; thus it was 
boiled in vain with water in a distilling apparatus; no product 
could be obtained representative of the taste or smell, and 
stiiriess of the properties of coffee. It would appear then, that 
this principle is fixed, or at least, that it remains adherent to 
the oil; which, moreover, undergoes but little alteration by 
this slight torrefaction. 
It remains for us to state, in order to complete the table of 
failures, that the presumed presence of gallic acid has induced 
us to try the method by fermentation upon green cofiee, and 
we have thus been enabled to determine a species of putrefac- 
tion in this grain. Coffee becomes very white after being 
macerated some days, the water acquires a high degree of 
colour and a viscid consistence, and, if submitted to evapora- 
tion, an extract is made, which re-dissolved by alcohol, affords 
