138 BROOKLYN BOTANIC GARDEN MEMOIRS 
the rust. The difference in the shape of the buffer cells of G. trans- 
formans and G. fraternum corresponds roughly with that between the 
teleutospores; they are comparatively long in the latter species. 
On January 30, 1917 a potted plant naturally infected with G. 
fraternum was taken from the cold frame and examined. Several 
leaves showed by the presence of slight yellowish spots that they 
would develop sori. Sections of one leaf cut on this date showed 
Fic. 4. Section of a leaf of the white cedar infected with G. fralernum show- 
ing the layer of buffer cells at the time when the epidermis has been quite completely 
disorganized. At the right a large palisade cell in the process of disorganization, 
but the nucleus is still visible and a haustorium is present. 
that the pseudo-parenchyma or teleutospore primordium was well 
marked. The buffer cells were mostly without granular contents and 
nuclei (Text-fig. 5, 4). In a few cases fragments of the degenerated 
nuclei could be seen. By February 2, a sorus taken from the leaf of 
the same plant showed great numbers of teleutospore buds in the 
2- and 4-nucleated stages (Text-fig. 5, C). A few buds had 6 nuclei, 
and the stalk cells of these had been cut off. Buffer-cell walls 
were just visible as narrow irregular lines showing most distinctly 
at the base of the young teleutospore. Two days later, February 4, 
cross-walls had been formed in many spores but nuclear fusion had 
