930 Gates—Pollen Formation in Oenothera gig as. 
normal growth. It is evident that this type of sterility is not due primarily 
to any failure of the cytological processes concerned with reduction, for 
these are followed through in many cases without irregularities. The 
tapetal cells, after the karyokinetic division of their nuclei which makes 
them all binucleate, do not undergo the further nuclear divisions and other 
changes which are characteristic of these cells in ordinary cases, but their 
later alterations seem to be suspended. Yet they seem well equipped for 
performing their glandular function, and the mother-cells are apparently 
not at all lacking in nourishment. This condition, in which the pollen 
mother-cells retain their archesporial condition and remain in contact 
although their nuclei undergo the normal reduction processes with few 
irregularities, has been found in four flowers of O. gigas , all taken from one 
plant. In one of these nearly all the anthers were in different stages 
of reduction, a condition already mentioned as being of much value, in 
connexion with the interpretation of synapsis stages. In this flower, 
in anther no. i the pollen-tetrads had been more or less successfully formed, 
as in PI. LXX, Fig. 83, but the walls of the mother-cells were very thick and 
cutinized. They still formed a compact tissue, and evidently could never 
round off and separate to form pollen-grains or free pollen-tetrads. The 
mother-cells and also the tapetal cells of this anther have very little contents 
and are evidently lacking in nourishment. In anther no. 2 of this flower the 
mother-cells are in interkinesis (Plate LXX, Fig. 65). In some cases a wall 
has been formed between the daughter nuclei, in others there is only a seg¬ 
mentation of the cytoplasm or an evanescent cell-plate. Anther no. 3 is in 
much the same stage as anther no. 2, some of the cells being in the homotypic 
prophase. In anther no. 4 the two reduction divisions had been completed, 
as in Plate LXX, Fig. 80, and small extra nuclei in the cytoplasm indicated 
frequent irregularities of the spindles. In anther no. 5 the mother-cells had 
undergone the reduction division, followed by segmentation of the cytoplasm, 
and the daughter nuclei had passed into the resting condition. In no. 6 
the mother-cells were mostly in the early homotypic telophase, and a seg¬ 
mentation of the cytoplasm had followed the reduction division. In no. 7 
the mother-cells were chiefly in interkinesis, showing the chromosome bi¬ 
valents of this time. Anther no. 8 was in a stage just following synapsis, 
as illustrated by Plate LXVII, Figs. 15-19. In all the anthers except the 
first the tapetal nuclei are binucleate and well filled with cytoplasm, in the 
condition they normally exhibit about the time of synapsis. The other 
three flowers which showed this type of sterility were more uniform in 
the stage of development of the mother-cells throughout each flower. 
Somatic Mitoses. 
No figures of the somatic mitoses in O. gigas will be presented in the 
present paper, but the fact that they are much larger than in the forms with 
