1907] GATES—POLLEN DEVELOPMENT AND MUTATION 85 
tapetum is normal at this time, but occasionally some of the tapetal 
cells or even the mother cells may be found degenerating in this 
“resting” condition; in jig. 3 both these conditions are present. 
Certain scattered tapetal cells show by their loss of turgor and their 
deeply staining contents that they are degenerating. Their nuclei 
may become flattened, but at first retain their nucleoli and nuclear 
wall. Later the latter breaks down and the nucleoli lose their outline 
and become indistinct, until finally the nucleus has disappeared. 
These changes are frequently accompanied by vacuolation. The 
cytoplasm stains progressively deeper with safranin and with haema- 
toxylin. This may be due to the distributed condition of the chroma- 
tin, which diffuses through the cell after the nuclear membrane 
disappears, or to the formation of some acid product in the cytoplasm 
as the result of the mucilaginous degeneration. Similar staining reac- 
tions are shown by degenerating mother cells. Fig. 4 shows a stage 
of this degeneration. 
Most of the recent papers on pollen development in sterile hybrids 
do not describe the condition of the tapetal cells. GREGORY (9) 
mentions that the somatic cells surrounding the pollen mother cells 
in a hybrid race of Lathyrus odoratus are normal. JUEL (13), in his 
description of pollen development in the sterile hybrid Syringa 
Rothomagensis, makes no reference to the tapetal cells; nor does 
Cannon describe the condition of the tapetum in his two papers 
(4,5). TiscHter (31) finds conditions in the tapetal cells of hybrids 
of Ribes very similar to those described in some detail by BEER (3) 
in Oenothera biennis and O. longiflora, and which I have observed 
also in O. lata and O. Lamarckiana. 
The tapetal cells in O. Jata are uninucleate during the “resting” 
stage of the pollen mother cells. During the early presynaptic stages 
in the mother cells the tapetal nuclei divide mitotically, and the 
binucleate condition persists for some time. This first nuclear 
division in the tapetal cells may take place almost simultaneously 
throughout a loculus, so that many mitotic figures in various stages 
are often found in the same section. Later divisions are more irregu- 
lar and are probably chiefly by amitosis, as stated by BEER (3) for 
O. longiflora and O. biennis, although occasionally twofspindles may 
be found in the same tapetal cell. 
