590 
Journal of Agricultural Research 
Vol. XXVI, No. 12 
The guard cell walls are rarely swollen, although they are usually 
altered chemically. That these walls can be affected, however, is shown 
in Plate 6, C, where the cell lumen actually is pinched in two by the 
coming together of the opposite walls of the cell. It is noteworthy, too 
that although the walls are strongly affected, a part of the cell contents 
still is living. 
Two fungi entered adjoining stomata on the same leaf and grew in 
divergent directions. The two were of about the same age and size and 
the cells killed directly by the fungus were similar in appearance. A 
close comparison of the two revealed a notable difference in the matter 
of swollen walls. 
A portion of one of those two fungi is drawn in Plate 6, A, already 
referred to. A slightly swollen wall is found at g and nowhere within 
the influence of this fungus is there any more pronounced effect. In 
the other of the two fungi, on the contrary, there are several cells with 
markedly swollen walls (PI. 6, D). The lamination of such a wall can 
be made out clearly, and at a some of the wall material, perhaps the 
middle lamella, seems to be almost liquefied. 
These differences between infections have been noted repeatedly. It 
may be that there are two strains of the fungus here, differing only in 
their effect on the cell walls. The culture from which this material was 
inoculated was not started from a single spore, so this would be possible. 
Both give the o type of infection. The difference is one of degree, not 
of kind. Another less likely explanation is that the composition of host 
cell walls may differ slightly in different parts of the leaf. 
One odd fact is that host cells whose contents react violently to the 
fungus do not show swollen walls, and the swelling is often greatest some 
distance away from the fungus. In Plate 6, E, (four days after inocula¬ 
tion) the cell at a was entered by the fungus and killed and has completely 
collapsed. So far as can be judged its wall is still thin. In the near-by 
cells at b and c, on the contrary, the walls are swelling rapidly and have 
arched outward, severing the contact-wall between them. The granu¬ 
lar matter between the two swollen walls is probably the disintegrating 
remnant of the middle lamella. So, too, in Plate 5, B, the wall is most 
swollen at d , a little distance away from the fungus. 
The finest specimens of swollen walls are found in much older material. 
One of these was drawn at high magnification (X 1460) in Plate 7, A, 
from a leaf fixed 11 days after inoculation. These swollen walls take 
no stain (with the triple stain at least) and are transparent, but by 
reducing the light every detail stands out clearly. Mesophyll tissue has 
fairly thin walls, and it seems almost incredible that a corner where two 
or three cell walls meet, such as the one at c, could swell into the bulky 
masses seen at a. Yet, such is obviously the case, as each part of this 
enormous wall can be traced back directly into ordinary unchanged walls. 
The distance from a to 6 is the actual thickness of the wall, and every 
detail of its structure is proportionately enlarged. Such a wall must be 
almost liquefied. It is interesting to note that even here an occasional 
nucleus and a few plastids have survived. Nothing corresponding to this 
has been observed in infections of this rust on Baart or Kanred. 
