60 
But even four years ago I had carried on the culture of smut fungi 
in nutrient solutions for a long time, in order at least to make the first 
part of the investigations as complete as possible. I had cultivated the 
conidia of oat smut and corn smut from generation to generation for 
more than a year. Every four days the nutrient solutions were ex¬ 
hausted and the mass of yeast conidia was deposited in the culture as 
a distinct sediment. A few germs from the exhausted culture were 
always introduced by a needle-point into a new nutrient solution, and 
in another four days this was also exhausted. The serial cultures 
amounted to more than a hundred, which must have corresponded 
to about fifteen hundred continuous generations of the yeast conidia 
produced exclusively by sprouting. Yet the conidia produced in the 
last culture were of the same form as in the first. According to this, 
the sprout conidia in their unbroken succession are to be regarded as 
the exclusive product of the growth of these smut fungi in nutrient 
solutions outside of the host plants. This result is noteworthy in the 
same degree as the long-known fact that the smut spores are exclu¬ 
sively the form of the same smut fungi inside of the host plants. 
Only this one change was to be observed in the continuous genera¬ 
tions of thesproutconidia—they gradually pushed out into threads more 
slowly when the nutrient solution was exhausted. After ten months 
culture, after more than 1,000 sprout generations were formed, the ger¬ 
mination in threads ceased entirely, the conidia swelled up somewhat 
and divided perhaps into two cells, but then remained passive. If we 
reflect that the conidia can penetrate into the host plants, to produce 
smut, only by means of their germ tubes, then with the disappearance 
of the tube germinations their infective power must also necessarily 
cease. 
Consequently in the loss of this morphological character we have 
found a natural explanation for one of the much discussed special cases, 
viz, why the infective power of fungous germs should cease with lapse of 
time and with exclusive maintenance outside of the host. I will show 
later that as a matter of fact infections with these germs were without 
result, but first I will state briefly that in its composition and concen¬ 
tration the nutrient solution remained exactly the same during the en¬ 
tire period of the serial cultures; that therefore influences of nutrition 
and of the method of culture could not have brought about in the 
conidia the gradual cessation of thread germination. 
The pure and satisfactory material from the sprout conidia of these 
smut fungi and of some others was also incidentally examined as 
to its possible power to induce alcoholic fermentation in nutrient so¬ 
lutions rich in sugar. But the forms investigated proved incapable of 
fermenting sugar, and could not grow at all in some of the larger masses 
of fluid. In these the sprouting remains nearly stationary and the 
germs finally die, probably from lack of sufficient access to air. From 
this behavior of the sprouts of the smut fungi in large masses of fluid 
