Nuclear Divisions in the Rusts . 
35i 
cells seem to be cut off in such instances (see Fig. 11) more or less in- 
definitely from certain well-nourished hyphae which line the bottom of the 
cup. To all appearances this process is quite similar to the cutting off 
of sterile cells in the caeoma forms. 
With Christman, I should regard such sterile cells as having but 
little phylogenetic significance, and as serving merely as ‘buffer cells’, 
to assist in rupturing the epidermis. Such sterile tips, at least in the case 
of the caeoma type of Rust, are clearly gametophytic cells which finally 
die from lack of food, or which fail of fertilization. But the presence 
of a sterile cell at the tip of only one of the fusing cells in the caeoma and 
not on the other has, in my opinion, an important physiological significance 
which has not heretofore been pointed out. This phenomenon indicates, 
to my mind, that those hyphae which bear such cells at their apices grow 
up from below some time before those hyphae which do not possess such 
a structure. The first gametophytic cells to group themselves more or less 
regularly upright and parallel under the epidermis of the host, apparently 
thus grow vegetatively for a time and cut off sterile tips ; but so far as 
my observation goes, these early hyphae do not fuse among themselves. 
Other distinct hyphae appear to grow up later from below and to press 
in between the earlier growths, thus providing the fertilizing gametes. 
That the conjugating pairs of gametes are thus differentiated in time of 
development, is clearly indicated by the two highly significant facts noted 
above ; viz., that one of the gametes bears a sterile cell while the other 
does not, and, secondly, that one gamete, in my experience, generally lies 
somewhat below the other. To these points might be added also the 
important fact that after the earlier hyphae are grouped under the epidermis 
a period of considerable duration undoubtedly elapses before the binu- 
cleated condition arises. 
These facts, in my opinion, point to the interpretation just mentioned 
as a more simple and reasonable one than that proposed by Blackman. 
The latter’s idea that the sterile cells are abortive trichogynes, while of 
great interest and theoretic importance, has therefore but little basis 
in fact. 
It is quite evident that the nuclear divisions described above are 
to be classed as indirect divisions rather than amitotic, as Blackman regards 
them in the forms he has investigated. The latter author studied examples 
in which the chromatin in each vegetative nucleus was fused together 
so as to form a dumbbell-shaped mass and was otherwise poorly differ- 
entiated during division. Such a condition is apparently common among 
Rusts, since every one of the previous investigators who had studied the 
subject described similar appearances to those he figures. Holden and 
Harper probably rightly conclude that, in such vegetative mitoses, poor 
fixation must be responsible for the lack of differentiation. Blackman, on 
