386 DIVISION III. —MODE OF LIFE OF THE FUNGI. 
production of gonidia; and it may be assumed, though it has not been yet distinctly 
proved, that they may form germs capable of infecting, and that by means ofthese germs 
they attack the host-plant and thus return to a parasitic life, somewhat in the way 
described at page 373 in the case of Cordyceps. 
The facts stated in pages 265, 339 show that the behaviour of Exoascus and 
Exobasidium is quite similar to that of the Ustilagineae. 
The example of Cordyceps leads to the mention of Claviceps, and Epichloe 
typhina also approaches near to the latter genus in the points in question. The 
gonidia (and possibly also the ascospores) of these Fungi may develope small mycelia 
producing fresh gonidia in the manner described on page 227. That this facultative 
saprophytism in its various degrees is a frequently recurring phenomenon in parasitic 
Ascomycetes is to be expected, though more stringent proof of it is in most cases 
still to be desired. 
Of all the Ascomycetes the Lichen-fungi, according to our present knowledge, 
must be mentioned first as examples of strictly obligate parasites, after them the 
Erysipheae and Polystigma (section LXIII). It has still to be ascertained what are 
the exact conditions in this respect in the large number of parasites in the groups 
of the Hysterineae and Phacidieae. The Peronosporeae also contain excellent 
examples of the class which we are considering, for most species of Peronospora 
and all of Cystopus are strictly parasitic, the first stages only of germination being 
completed outside the host. This is the case also with Protomyces and many Chy- 
tridieae, some of which even commence germination on the surface of the host 
(see Chapter V). Lastly, only the strictest parasitism is known in the group of the 
Uredineae so rich in forms; they germinate if sufficiently supplied with water, and 
their further development takes place only on the proper host. 
Section CX. The Fungi which are parasitic on plants naturally exhibit within 
the limits of the chief phenomena of parasitic vegetaticn and its effects, which were 
pointed out on page 359, a variety of special adaptations in respect of their choice 
of a host and their spreading in, upon, or along with it. The reaction of the host itself 
which varies in each case corresponds again to the spread of the Fungus. We call 
attention to the following facts which are of general interest in relation to these 
points, again referring the reader to the former sections of this work and to the 
special literature of the subject. 
As regards the choice of the species to serve as a host, the rules stated on 
page 359 are of the first importance. Most parasites living on plants require a single 
proper host for the completion of their whole course of development, though they may 
enjoy a larger or smaller room for choice between different species more or less nearly 
allied to one another. Of all the hosts that are possible for a species of Fungus some 
may be more favourable to their development than others. Cystopus cubicus for 
example flourishes and forms abundance of gonidia on the leaves of species of Trago- 
pogon, Podospermum, and Scorzonera, but forms oospores almost exclusively on 
Scorzonera, especially on S. hispanica; oospores are extremely rare on Tragopogon 
in my experience. So it is with the Uredineae and species of Erysiphe. The best 
known Erysiphe is the Fungus of our grape-vine' which in Europe only forms gonidia 

1 See De Bary u. Woronin, Beitr. III. 
