64 
the sprout conidia of the different kinds of smut fungi named always 
possess a definite and characteristic form, and intruding germs can be 
distinguished without difficulty. It was no trouble to keep unlimited 
quantities of these conidia, since a large number of these flasks for 
securing the material for infections were always prepared at the same 
time, in order subsequently to unite the sprout conidia out of these dif¬ 
ferent flasks. 
From the previously more fully described details as to the develop¬ 
ment of these fungi through the sprouting of their conidia, we now know 
that the multiplication ceases with the exhaustion of the nutrient solu¬ 
tion, and that when this happens there immediately occurs a pushing 
out of the conidia inU) germ threads,,which in turn, in the space of at 
most two days, lengthen out, but then cease to grow, and gradually 
perish. By means of these threads must the fungous germs pene¬ 
trate into the host plants, when they begin to germinate upon their 
surface. Since the development into germ threads immediately stops 
the increase of conidia by sprouting, and since the germ threads, as they 
continue to develop, grow into the host plants, it follows that the sprout 
conidia must be transferred to the host plants, if the infection is to be 
attended with the best results, at a time when they are still sprouting 
and have not yet grown out into threads. The conidia which have 
already grown out into germ threads are very liable to injury by their 
transferrence to the host plants, and as soon as they have grown out 
are scarcely able to penetrate into the latter. The most favorable 
period for infection is very transitory in the rapidly growing conidia, 
and if it is missed, a normal success of the experiment is not to be ex¬ 
pected. In view of this, the necessary precautions were taken to have 
the plants which were to be infected always ready in the various stages 
of development required for the individual experiments at the same 
time that the sprout conidia just described had reached their most favor¬ 
able point of development for infection. 
The transfer of the sprout conidia to the host or experiment plants was 
done with the help of an atomizer which Wolff had already used in his in¬ 
fections with smut fungi, and which I had myself formerly put to manifold 
uses in my mycological investigations. The sprout conidia from the dif¬ 
ferent culture flasks were quickly united in one flask, in theneck of which 
the atomizer, cautiously tested for satisfactory performance, had been ad¬ 
justed beforehand. Without the use of an atomizer it is impossible to 
bring the fungous germs upon the experimental host plants in the neces¬ 
sary degree of dispersion. Only in the tiniest drops do the sprayed fluids 
remain sticking to the parts of the plant upon which they have fallen; in 
case of larger droplets there occurs at once a union into drops which flow 
off, and consequently hinder the development on the plant and the pene¬ 
tration into it of the germs transferred with the droplet. 
But it is now known that fungous germs are easily injured when 
taken in lull vegetation and suddenly transferred from nutrient solu- 
