340 
SELECTED ARTICLES. 
sure to the atmosphere, presented the following characters. 
Heated to 100°, they lost 23 to 100 of their weight, and as- 
sumed a reddish tint; by continuing to heat them in a glass 
tube to 250°, they became brown, without undergoing decom- 
position. At a more elevated temperature, exposed to the 
air, the substance was blackened, inflamed, and left as a re- 
sidue a white ash retaining the form of the crystals. If heated 
to redness, the residuum entered into fusion, and left, upon 
cooling, a hard mass, which, with re-agents, exhibited the 
characters of carbonate of soda. 
110 parts of boiling alcohol at 30°, dissolved 10.3 of gen- 
tisate of soda, and when cold, 7.0. This vehicle allows crystals 
of gentisin to be deposited, and retains the soda. The so- 
lution of gentisate of soda is of a beautiful yellow colour. If 
a current of carbonic acid gas is made to pass through it, the 
liquid is deprived of its colour, and deposits gentisin of a 
white colour, which, however, resumes the yellow tint when 
desiccated. 
Upon burning gentisate of soda in a platina crucible, 
and adding sulphuric acid, I obtained sulphate of soda, the 
weight of which indicated the quantity of soda. It results 
from this experiment, which has always given the same weight 
when repeated, that the gentisate of soda is formed of 6.81 
soda, and 93.19 gentisin. 
The action of concentrated acids upon gentisin aiSbrds no- 
thing remarkable. With sulphuric acid the colour is deepened, 
is suspended, and disappears in part; heated, it becomes red, 
then black, and if it is evaporated, nothing but carbon remains. 
There is no action when the acid is diluted. Concentrated 
nitric acid exhibits no action at the ordinary temperature; if 
heated, the substance becomes coloured greenish yellow; and 
if carried to evaporation, it becomes black, and leaves carbon 
as a residuum. 
I should state here, that recently, and about the time that I 
announced that the crystalline substance of the gentian is not 
the bitter principle, results analogous to those obtained by me 
were published by M. TrommsdorfT. He, however, asserts 
