—_—_———_——-_——-— . 
ar Scene a 
1907] CAROTHERS—GINKGO BILOBA 121 
the sac by deposit, therefore, corresponds to the formation of the walls 
of the microspore, and the fastening to it of the outer ends of the 
radial walls corresponds to the similar attachment of the walls cutting 
off the prothalial cells in the male gametophyte.* 
These walls, at right angles to the enclosing b and fastened 
to it, separate the nuclei, probably each one being enclosed in a cell 
of its own. In other forms the first radial walls are said to come 
in on the spindles of the last simultaneous free nuclear division. 
Since the later divisions of the free nuclei here are far from simul- 
taneous, this cannot be the case. Although the coming in of these 
first walls was not observed, the subsequent centripetal growth and 
division of the cells formed by their development are clearly seen 
(jig. 25). The inner cells of the centripetal rows are larger than the 
outer ones, the size of the latter being early diminished by repeated 
divisions, both tangential and radial (fig. 25). In most cases the 
nuclei of these cells, which are open to the interior of the sac (fig. 27), 
are much larger than those of the enclosed cells near the periphery. 
Most of the open cells have a single nucleus (fig. 27), but two nuclei are 
to be seen in some (jig. 28), and in a few, three nuclei. Binucleate cells 
(jigs. 29, 30) are common in the tissue near the open central space, 
and even a few multinucleate cells are present (jig. 31), but since cells 
with a single nucleus are not found in the mature endosperm, these 
very cells must later become uninucleate. Whether this change results 
from a fusion of the several nuclei, a degeneration and absorption 
of the superfluous number, or their separation by the formation of 
intervening walls is not certainly known. In all cases of wall forma- 
tion seen within cells, however, the wall is on a regular karyokinetic 
spindle. Usually the two nuclei in a cell are in contact, and it is not 
uncommon for one of them to be smaller and more dense than the 
other, suggesting its degeneration and absorption (jig. 30). From 
this it appears probable that cells which have at first two or more 
nuclei become uninucleate by a fusion of the several into one, or by 
the degeneration and absorption of the superfluous ones, and not 
by the formation of separating walls. 
of curious and irregular wall formation, abnormal and perhaps the 
result of Guilficlagical coma is seen ina — of June Ayal Rites pa a con- 
tainin, i, each 
free nuclei is as usu iti 
group quite separate from the sv and each with an decir membrane (fig. 26). 
