926 Gates.—Potien Formation in Oenothera gig as. 
as to occupy nearly the whole periphery of the mother-cell, but with a large 
cavity in its centre. It would seem that this change in the shape of the 
young pollen-cells must be from forces inherent in the cells themselves, 
with which the mother-cell can have little concern. The discoid shape 
of the young pollen-grains, and afterwards of the older ones, is clearly 
self-determined. 
Miss Lutz (’ 09 ) was the first to make the interesting observation that 
in Oenothera gigas the pollen-grains have characteristically four lobes or 
interstitial bodies, while all the other species of Oenothera , including 
O. Lamarckiana and its other mutants, are known to have usually three. 
These four interstitial bodies in the O. gigas pollen-grain are symmetrically 
placed at the corners of a square (PI. LXIX, Fig. 49), showing that the four 
are as characteristic of O. gigas as three, placed at the corners of a triangle, 
are of O. Lamarckiana and its other mutants. Sometimes, however, one finds 
five interstitial bodies symmetrically placed, as in Fig. 50. At the age of 
the pollen-grain represented by Figs. 49 and 50, the cytoplasm still fills the 
whole cell, the nucleus is centrally placed, spherical, and with a definite 
membrane. Its reticulum is rather coarse, and two or three nucleoli are 
usually present. 
Figs. 51 and 52 show a much later stage, in which the pollen-grain has 
undergone a great deal of growth, and the interstitial bodies give a more or 
less marked quadrangular appearance to the grain. The nucleus usually 
appears amoeboid in shape at this time, with very little chromatic content, 
and the cytoplasm is in coarse strands occupying only a portion of the cell. 
Beer (’ 06 ) has studied carefully various features in connexion with the 
growth and development of the triangular pollen-grains in O. longiflora and 
O. biennis. 
.In O. gigas, pollen-grains having more than four lobes are relatively 
common. Occasionally the lobes may then all be symmetrically placed, as 
in Fig. 50, but much more frequently the four original lobes are placed 
symmetrically, and the extra lobes are intercalated between these in various 
positions, frequently in a plane above or below the original ones. This is 
shown in Figs. 53 and 54, the first of which has four lobes rather regularly 
placed, and a smaller fifth lobe which is below the plane of the others. 
Fig. 54 has seven lobes, the three somewhat irregular and smaller additional 
lobes being all on the left side of the figure. It is therefore clear that in 
O. gigas the pollen-grain is constructed on a basis of four, just as in the 
other species it is on the basis of three, although grains with more than 
three lobes do occur. The reasons for this rather remarkable change are 
not clear, although one is tempted to speculate on the different size relation¬ 
ships of the cells and nuclei in the young pollen-grains, as a possible basis 
for this alteration. The doubling of the chromosome number, with the 
attendant doubling in the volume of the nuclei, but a different ratio of 
