927 
Gates.—Pollen Formation in Oenothera gigas. 
► nuclear surface 
increase in the cytoplasm, gives a new ratio for - n7 -. 
& cell surface 1 
may conceivably have been the factor determining the appearance of 
a quartet instead of a triad of interstitial bodies. 
This completes the account of the normal development. Many cyto- 
logical features which were described in previous papers have not been 
repeated here, only the distinctly new features being emphasized. 
Sterility in O. gigas. 
I wish now to describe an interesting type of sterility in O. gigas % 
which differs in some respects from anything previously described in plants. 
Its peculiarity consists essentially in this : The cytological phenomena of 
reduction go forward in the normal manner, but the pollen mother-cells 
retain their archesporial appearance, their walls remaining in contact with 
each other and with the tapetum to form a compact tissue, so that they can 
never be set free or discharge pollen, even though the reduction processes 
are normally completed. This appears to be due primarily to a failure of 
the surrounding tissues of the anther to grow and form a cavity which 
allows sufficient space for the mother-cells to round off during synapsis 
and float freely in the cavity of the anther. Other features of this process 
will appear in connexion with the description. In all, four flowers, all from 
the same plant, were found to show this condition in various stages. 
Fig. 55 is from a longitudinal section, showing some of the pollen 
mother-cells and their surrounding tapetal cells as they appear under a low 
power. Though the tapetal cells are binucleate and the mother-cells are in 
interkinesis, there is no tendency for the mother-cell walls to break apart or 
separate from the tapetum. The walls of the mother-cells frequently 
become much thickened and cutinized at this time. Fig. 56 is from another 
anther of the same flower, and shows a portion of a row of mother-cells. 
In the two long central cells of the row there has been an attempt at wall- 
formation after the heterotypic mitosis, which resulted only in an incomplete 
segmentation of the cytoplasm. The four cells so formed are each under¬ 
going the second reduction division. There is frequently a more or less 
abortive attempt to form a wall after the heterotypic mitosis. 
Figs. 57 and 58 are drawn with a little higher power. They each show 
two pollen mother-cells from a row, and all are in the heterotypic telophase. 
In all there has been a partially abortive attempt to form a wall separating 
the interkinetic nuclei, and in two cases this has resulted in a complete 
constriction of the cytoplasm. The lower mother-cell in Fig. 58, particularly, 
shows evidence of irregularity in the chromosome distribution during the 
heterotypic mitosis. The reduction processes are not entirely normal, and 
some of the evidences of irregularities will be seen in later figures. PI. LXIX, 
Figs. 59-61, PL LXX, Fig. 62, represent stages of the homotypic prophase 
3 P « 
