344 GEORGE MASSEE [NOVEMBER 
pustules, therefore no comment is necessary; only further experiments 
in the same direction will be made in the future. 
Tradition acts as a powerful bias, even in scientific matters; 
immediately following De Bary’s brilliant discovery of heteroecism, 
the condition of rust on the Barberry alternating with that on some 
graminaceous plant was considered indispensable for the continuation 
of the species; eventually it was discovered that the stage on Barberry 
could be dispensed with, and yet the rust appeared as rampant as ever ; 
in fact in Australia, where rust is more abundant and injurious than in 
Europe, the aecidium condition is unknown; in India also, where ruts 
is very destructive, no aecidium condition is known to exist within 
hundreds of miles of the wheat-growing districts. At the present day 
it is generally accepted that the uredo-spores only retain their power 
of germination for a very limited period, and that the uredo-spores 
must be in contact with the host-plant to effect inoculation. The 
experiment just recorded modifies this idea to some extent. Teleuto- 
spores, again, are considered at present as being only able to infect the 
host that bears the aecidium stage; however, their production in such 
immense numbers in those countries where no aecidium stage is pro- 
duced, or required, suggests that they may possibly play some part in 
the reproduction of the fungus hitherto undiscovered. 
Numerous preparations of rust-shrivelled grains of wheat have been 
examined microscopically, and an abundance of mycelium detected in 
the outer layers of the grain, correctly speaking, in the pericarp; but 
not in a single instance have I been able to detect mycelium in the 
embryo; and in those cases where the grains were allowed to germinate 
and form a tiny plantlet up to half an inch in length, the mycelium 
never appeared to pass into this part. On the other hand, when 
sections of diseased wheat were placed in culture media, hyphae 
frequently radiated from the section on all sides for some distance. 
May not similar hyphae radiate in the soil from diseased grain 
when sown under natural conditions, vegetate for some time in a 
saprophytic manner, and finally, if conditions are favourable, infect the 
young plantlet at, or just below, the ground level? Sufficient of 
obviously rust-shrivelled grain is frequently used as seed; and if, in 
addition, plants are infested with mycelium, which for some at present 
unknown reason does not produce spores, as I have shown to be the 
case with Scilla bifolia, and also recorded by Bolley (11) as frequently 
occurring in the case of wheat attacked by Tvlletia levis, Kiihn; and 
assuming that this mycelium also passes into the grain, then we should 
be able to account for a considerable quantity of the rust prevalent, 
without introducing a new factor—mycoplasma—into the theory. 
Mycelium of the rust fungus has been observed in the grain of 
wheat by Eriksson, as shown by the following quotation (12), and if 
inoculation of the young plant is effected by means of mycelium 
originating from the grain, and growing for a longer or shorter period 
