1907] GATES—POILLEN DEVELOPMENT AND MUTATION 87 
lata pollen development is normal up to the tetrad stage; that the 
tapetal cells later elongate and multiply, filling the whole loculus; and 
that the pollen cells in the mean time have broken down and disap- 
peared, only a few small scattered grains remaining. The first state- 
ment is erroneous, because degeneration may take place in the pollen 
mother cells before the reduction divisions (figs. 3, 4). Pout’s 
account of the history of the tapetal cells is also contrary to my observa- 
tions. It is true in general that the tapetal cells continue active 
during synapsis and the reduction divisions in the pollen mother cells, 
although breaking down frequently begins at this time, or even earlier, 
in the “resting” stage, as already mentioned. But in later stages, 
when the pollen cells have largely disappeared after the reduction 
divisions, leaving an empty cavity, the normal condition is not as 
PoHL states—ingrowth of the tapetal cells to fill the cavity—but 
partial or complete disappearance of the tapetum; and in the latter 
case subsequent ingrowth of the middle layer cells behind to fill 
more or less completely the cavity (figs. 5, 6). Stages of this process 
are found in which the disintegrating tapetal cells form a thin layer lin- 
ing the locular cavity (which is usually empty under these conditions), 
the remains of this layer sometimes appearing later as a thin, homo- 
geneous, deeply staining ring (fig. 5), which may afterwards completely 
disappear. The middle layer cells, which now line the cavity, in the 
meantime become rounded on their inner surfaces, owing to the release 
of the pressure of the tapetal cells. They then push in to fill the 
cavity and may even multiply by mitotic division. The cavity of the 
loculus may thus become completely closed (jig. 6). 
Pout’s error in mistaking these middle layer cells for tapetum 
was probably due to a hasty and insufficient examination, for in care- 
fully prepared sections the former have a totally different appearance 
from the tapetal cells, for which they could not possibly be mistaken. 
While the latter are easily recognizable by their densely granular con- 
tents and conspicuous nuclei, as well as by their rectangular shape, 
the middle layer cells have no granular contents, are uninucleate, with 
small nuclei having little chromatin, and are rounded and irregular 
in shape. Pout’s error is probably to be attributed to his failure 
to observe the degeneration and disappearance of the tapetum before 
the ingrowth of the middle layer cells takes place, as well as to distin- 
