308 
THE FUNCTION OF TANNIN. 
the same cells and the same granules also responded, by a 
reddish coloration, to the action of iodine reagents. 
The use of these and other reagents showed, therefore, 
that we had here granules which— 
(a) Stain reddish, and subsequently bluish, with iodine 
reagents, and are therefore starch ; 
(b) Stain bluish witli iron salts, black with osmic acid, 
form brown balls with potassium bichromate, 
and are therefore tannin. 
In other words they show that these grains are at the 
same time starch and tannin. 
Through preparations which were treated with an iron- 
salt I passed for an hour or more, under the microscope, a 
current of water. This removed from cut cells the light iron 
tannin precipitate, but left the grains still coloured, but more 
lightly so, with the reagent. Each grain was distinctly 
bluish, but in no case did any grain which was so coloured 
dissolve. 
Next I treated sections, with which I had not used any 
reagents, with a stream of water. No solution of grains was 
observable. The use of iron salts showed that from the cut 
cells a considerable quantity of tannin had been removed, 
readily identifiable in the water, but that in no case, so far as 
I could tell, had any granules been dissolved, and many of 
the grains in the cut cells showed as before a blue coloration 
with ferric salts. 
In order to try whether by their differing solubility it is 
possible to separate tannin from glucose, I had had sections 
of this same plant m absolute alcohol for about three weeks. 
The tannin I had found to be not yet nearly dissolved out. 
Some of these sections, and others which had lain for a few 
hours in water, I tested with reagents. I found in no case a 
diminution in the number of the granule-containing cells, 
but I found that in the sections from water the granules 
no longer showed the tannin reaction, or only very feebly, 
while in the alcohol preparations some showed it slightly. 
These latter granules were in the same preparations treated 
with iodine, and showed themselves to be starch. 
These experiments I controlled by others when Q. peclun- 
culata was coming into leaf in late spring, and by yet others 
in summer. The granules (starch) had disappeared, but the 
tannin contents of the cells were still there, but in a fluid and 
not in a granular state. 
With arrowroot starch soaked in a solution of tannin, I 
have followed out all the reactions. The results practically 
are the same. Starch grains saturated with tannin take the 
