382 BOTANICAL GAZETTE [| DECEMBER 
alcoholic solution of the pigment is then filtered through a care- 
fully cleaned bacterial filter (Chamberland, Kitasato, d’Arsonval, 
or others). The first portions of the filtrate are rejected, owing 
to the fact that a chemical action takes place at first, doubtless 
owing to air present in the pores of the tubes; the result is the 
production of a beautiful purple-red liquid. » After some 50%, 
more or less, have passed, the filtrate passes unchanged in color. 
It is then evaporated to a syrupy consistence at 50° to 60°C,; a 
higher temperature leads to reduction and decomposition of the 
pigment at this stage, owing to the presence of large amounts of 
sugars extracted from the potatoes. The thick deep-blue liquor 
is precipitated with strong alcohol (about 98 per cent.), the 
supernatant liquid poured off, and the precipitated pigment dis- 
solved in a very small amount of distilled water and again repre- 
cipitated with alcohol. This process is repeated as long as the 
alcohol seems to extract anything from the pigment. At this 
point the pigment is precipitated in such a finely divided condi- 
tion that it refuses to settle completely, and cannot be retained 
on filter paper ; a very small bacterial filter is therefore employed 
for the separations. The material is then carefully removed 
from the filter tube and dried. This dry amorphous powder has 
a grayish-blue color, and is completely soluble in water to 4 
beautiful pure-blue color. Although we have reason to believe 
‘that it is still impure, the amount of impurity is doubtless ° 
small that the pigment thus separated can be used as the basis 
of comparison, and for the reactions given in this paper. It can 
not be made to crystallize, and is insoluble in all ordinary sol- 
vents, such as ether, petroleum ether, benzene, chloroform, amy! 
alcohol, etc. oe 
If the blue aqueous solution is treated with a trace of acid, 
the blue is changed to a violet, a trifle more acid leads to = 
production of a beautiful purple (royal purple). An excess 0 
acid gives rise to a red with more or less of a purple tint. = 
sensitive is the compound to acids that carbon dioxid cause 
a change of color. It was at first thought that when se? 
acids were employed, a color change resulted which was differet 
