DODGE: STUDIES IN THE GENUS GYMNOSPORANGIUM 139 
not occurred (Text-fig. 5, D). Remnants of buffer-cell walls were 
now difficult to find. A mature spore is shown in Text-figure 5, EF. 
The mycelium of G. fraternum penetrates through the leaf in every 
direction. The mesophyll cells are usually somewhat enlarged and 
are packed rather closely together with small intercellular spaces. 
Haustoria are quite abundant in such cells. The cells surrounding 
the vein are especially affected. They appear to be filled with minute 
granules and as many as eight or ten haustoria can be found ina 
single cell, the more common number being two to four. A complete 
ring of these large cells is formed. This is due to an increase in 
number as well as their larger size. This, taken in connection with 
the increase in the mesophyll tissue, gives the leaf a slightly thicker 
and more compact appearance. The mycelium does not invade the 
Fic. 5. a, pseudoparenchyma with buffer cells; 6, 2-nucleated stage of the 
young teleutospore; c, 4-nucleated stage; d, 6-nucleated stage with cross walls; e, 
small spore after nuclear fusion, 
vascular tissue even to the extent of penetrating the phloem of the 
stem. Both G. transformans and G. fraternum are capable of producing 
sori two or three years in succession, the latter may produce a sorus 
even after the leaf has apparently died. While haustoria are more 
numerous and attack the individual cells more vigorously in the case 
of the latter species, G. transformans seems in some way to be more 
destructive, as infected leaves more frequently die after maturity of 
the first sorus. Of the two species, G. fraternum is clearly the more 
nearly related to G. biseptatum, both from the nature of their teleuto- 
spores and the similarity of the aecidia in the two species. If the 
mycelium of G. fraternum is ever able to push in beyond the large cells 
surrounding a vein and get into the central cylinder of the stem, 
therefore nearer the cambium, we should look for a stimulus such as 
