I20 BOTANICAL GAZETTE [FEBRUARY 
large, the numerous divisions and their irregularity making the num- 
ber larger than 256, which is so common a number among the gym- 
nosperms. It is not impossible that the irregularity in division of the 
free nuclei, varying as it does from the simultaneous division said to 
obtain in other forms, may have been due to the unnatural conditions 
—growth after separation from the tree—although an effort was 
made to render conditions as natural as possible. 
Wall formation—About this time there appears on the outer 
surface of the plasmic sac a delicate membrane (fig. 24). This is not 
a Hautschicht, nor is it in any way attached to the original spore wall 
or embryo sac. It is a true cell wall of the type known as walls of 
deposit, and is formed by the protoplasm which lines it. Then walls 
appear in the cytoplasm, perpendicular to this enclosing membrane 
and with their outer edges fastened to it (fig. 25). The current 
accounts of wall formation following free nuclear division in the 
embryo sac, beginning with Miss Soxotowa’s (10) description and 
continuing through subsequent papers by other investigators, state 
that the walls appear at right angles to the embryo sac or spore wall, 
and with their outer edges fastened to it. This is distinctly not the 
case in Ginkgo; and the ease with which the spore membrane may be 
peeled away from the prothallia of many other gymnosperms, even 
when the latter are but partly developed, suggests the probability that 
it is not the case in them. 
It has long been known that the walls of microspores are quite 
separate and distinct from the wall of the mother cell in which they 
lie, and that the spores when mature escape from the mother cell 
membrane through ruptures or by its solution. In sectioned micro- 
sporangia showing developing tetrads, these separate surrounding 
walls of the mother cells may be readily seen. It is known also that 
the outer portion of the microspore wall, that next the wall of the 
mother cell, is formed by deposit, and not upon spindle fibers, as are 
the walls between the spores themselves. The megaspore wall, as 
can be seen in the figures showing the tetrad formation, with the excep- 
tion of that comparatively small portion which divides it from the 
non-functional spores, is the old membrane of the mother cell and 
corresponds to the mother cell wall enclosing the microspore tetrad. 
The formation” of this separate membrane upon the outer surface of 
