178 BOTANICAL GAZETTE [MARCH 
in the details of their development from the haustoria of Ery- 
siphe; but their abundance and the many weeks for which the 
prickly-ash bush must have been infested when the material was 
gathered, make it evident that they are modified haustoria. It 
is to be seen that there is some resemblance between figs. 27 and 
29 and the early stages of penetration represented in figs. 2 and 
3 from Erysiphe. The haustoria of the fungus on Xanthoxylum 
certainly answer very well to the knob-shaped structures De Bary 
described, which were probably young haustoria. It is impos- 
sible to decide absolutely that they were, however, because of 
their similarity to figs. 27 and 29. Certainly if they represent 
dead haustoria, as De Bary thought, the vigorous mycelium must 
be nourished in some other way. 
GENERAL CONSIDERATIONS. 
The phenomena exhibited by the intercellular hyphe of 
Phyllactinia are interesting in connection with what has been 
ascertained by several investigators in relation to the nutrition 
of fungi. It has been mentioned that the majority of the inter- 
cellular hyphz in several of the hosts studied (excepting Xan- 
thoxylum) take a more or less direct course to the regions 
near the bundles. The development by the fungus of the 
absorbing organs in regions abundantly supplied with available 
food, such as the parenchyma sheath of the bundles, indicates a 
selective chemotropism in the fungus. A selective chemotropism 
was reported by Miyoshi for several fungi in his paper already 
cited. Phyllactinia thus offers under normal conditions of 
growth a demonstration of the selective reaction which Miyoshi 
demonstrated by artificial means. This reaction enables Phyl 
lactinia not only to ‘surround itself with conditions which will 
insure it an uninterrupted supply of food and water such as 1S 
not insured to it while living as a purely epiphytic parasite, but 
also it is able to select within the leaf of the host those regions 
better supplied with food and water than the loose parenchyma. 
If the fungus is stimulated to place haustoria in cells joining 
the bundles to the palisade layer, or if haustoria enter the 
i itt 
