270 BOTANICAL GAZETTE [Apri 
the position in which the pore first appeared (fig. 7). The spindles 
in this division are more distinct and not quite so parallel as in sub- 
sequent conjugate divisions. Each shows a few short polar rays and 
to each pole are drawn several (certainly more than two, though the 
number was not determined) distinct chromosomes. The two 
nucleoli are at this time lying in the cytoplasm outside the spindle 
figures. They are less dense than at earlier stages and appear to be 
disintegrating. Two of the daughter nuclei, one from each spindle, 
wander back into the bases of their respective cells. The other two 
remain lying side by side and move into the distal end of the fused 
portion of the gametes, which now enlarges and elongates (fig. 8). 
A cell wall now cuts off the distal portion of this” region, which 
contains the upper pair of nuclei, and thus the first aecidiospore 
mother-cell is formed (fig. g). This cell, as has been described fre- 
quently, at once divides into two unequal cells, the aecidiospore and 
the small intercalary cell. 
After one spore has been formed, the nuclei in the bases of the 
fused gametes again move upward into the conjugated portion, 
which has meanwhile grown in length, and the process of division is 
repeated. In this way a single row of aecidiospores is formed from 
each pair of the gametes (fig. 10). During these processes the cyto- 
plasm retains the same granular appearance. The nuclei, excepting 
in figures showing division, always have the same general structure. 
From the fact that the bases of the conjugating cells diverge 
widely, it is suggested that the conditions shown in jig. 6 might be 
produced by the bending of a single hypha, as is the case in the 
development of an ascus. The earlier development of the gametes, 
however, shows conclusively that nothing of the sort occurs here, 
and that the cells which fuse may belong to distinct hyphal branches. 
We find thus a typical case of the fusion of gametes at the base of , 
each row of aecidiospores, with the difference that the nuclei do not » 
fuse and the cell produced by the fusion germinates at once. The 
- subsequent history of Phragmidium has been well worked out and I 
shall not take it up here. ~ ; 
My material of Cacoma nitens did not afford so complete 4 sc 
of stages, but I have been able to trace with perfect certainty the sil 
tory as far as the formation of the fertile and sterile cells. As in 
