296 BOTANICAL GAZETTE [NOVEMBER 
n “ : ; ; 
——— one grew and three failed! Witha strictly non-volatile 
102400 
substance we should expect less variation, because with less dif- 
ference between the vapor pressures of the hanging-drop and 
that of the water below, the changes in the concentration of the 
deleterious agent in the hanging-drop would necessarily take 
place more slowly. Of the volatile properties of HgCl, we 
shall speak later. 
In looking for a method that would meet every requirement 
of the case, many preliminary tests. were made with all five 
molds and with different deleterious agents. Two such tests 
with potassium cyanid are here presented in detail. It is of 
interest to note that while KCN is not in itself a volatile com- 
- pound, in aqueous solution more or less hydrolysis takes place, 
resulting in the formation of a corresponding amount of HCN 
(Shields ’93), which is quite volatile. Hence, aqueous solu 
tions of KCN are in their behavior quite typical of volatile com- 
pounds, 
Column 1 gives the culture label by means of which the ie 
tures were identified; column 2 the concentrations of KC} 
in the hanging-drops of the various cultures; column ; me 
solution—if any—used in the bottom of the cell. “Dry 
implies that no water or other solution was placed in the ee? 
‘of the cell. Under the head “germination” the percentage 
spores germinated is given for three observations, at 12, ee 
36 hours, respectively. Under “development”’ the lenge 
germ tubes of spores showing an average development Is 8 
in micromillimeters. in the 
These data not only show that cultures having ae 
bottoms of the cells are unreliable and vary acco ae 
amount of water present, but that “dry” cells oe cult 2 
of the , 
s but one fourth eo 
” former, how” 
of the KCN present in cultures labeled ‘‘A9, the he hanging” 
ever, had several drops of the solution from which ' : d 
drop was made placed in the bottom of the cell wens 
customary drop of water. The striking difference 1? 
results : 
ate 
