350 BOTANICAL GAZETTE [NOVEMBER 
agar without changing the composition of either. However, to 
accomplish this it was necessary to allow the tannin solution to run 
slowly from a pipette into the melted agar while the latter was 
constantly agitated. If the tannin was added too rapidly, or the 
agitation of the agar was insufficient, more or less coagulation re- 
sulted. With a 20 per cent aqueous solution of tannic acid, less 
than half this amount (0.8 per cent) could be added without coagu- 
lation. Moreover, even a very small amount of tannic acid in its 
solid form would cause coagulation in the agar. Another lot of 
MeErcx’s tannin was of such character that only about half as much 
tannic acid could be added to the agar without change. 
When first placed in the agar, the tannic acid caused a milky 
appearance, which disappeared on sterilization.. Where high per- 
centages of tannin were used, the agar upon sterilization showed 
a tendency to become viscid (about 0.8-2 per cent) or even 
liquid (about 2-2.5 percent). The transition between viscosity 
and liquefaction is gradual in such a series as we used, where 
each member differed from the next by 0.2 per cent of tannic acid. 
In no case was the distinct curd which various investigators have 
described to be observed. In agar with 3 per cent of tannic acid 
the entire mass of medium becomes a clear liquid with a thin film 
of solid matter on one side of the test tube if set for a slant, or in 
the bottom if set upright.. This solid material gives the same reac- 
tion both to Millon’s proteid test and to the ammonium molybdate 
test for tannic acid as does the solid agar of the lower members of 
the series. Similar results were obtained by testing the liquid 
portion of the medium. Evidently, the explanation of this lique- 
faction is to be found in some other direction than the chemical 
interaction of tannin and proteid. 
If the agar medium used is titrated to various degrees of acidity 
and a series of such tubes sterilized, it is found that the agar ranges 
from solid through viscid to liquid; that is, the same phenomenon 
can be induced by acidulating the medium as by the addition of the 
tannic acid. In each case a more careful test of the nature of the 
proteid substance in the liquid from the acidulated agar shows that 
proteid digestion has progressed so far that the power of solidi- 
fication has been lost. 
