1914] BURLINGAME—ARAUCARIA BRASILIENSIS 495 
at another time. At maturity the gametophyte is 4-4. 5 cm. long, 
2 cm. wide, and 1.5 cm. thick. ‘The entire seed structure is 5-6 
cm. long, about 2.5 cm. wide, and 1.8 cm. thick. 
The megaspore mother cell appears about the middle of May in 
the upper part of the group of denser cells already referred to. A 
first it differs only in size. As it enlarges, the adjacent cells show 
signs of disintegration. Fig. 5 shows a mother cell in synapsis. 
To the right and to the left of it may be seen two cells that are 
being flattened and whose nuclei are apparently beginning to 
degenerate already. I secured only a few preparations of the early 
stages and so cannot say certainly whether more than one megaspore 
mother cell ever begins development or not. Unfortunately I did 
not observe the reduction of the mother cell for the same reason. 
The chromosome numbers would indicate that a reduction does 
actually occur. Fig. 6 shows the position of the megaspore. In 
fig. 7 there may be seen two small cells just above the functional 
megaspore, that are probably the remains of the other members of 
the tetrad. A similar group is present in fig. 8. Very early the 
-Megaspore becomes vacuolate, with the very scanty cytoplasm 
orming an extremely delicate lining to the embryo sac. The 
nucleus is flattened, yet its thickness is several times that of the 
layer of cytoplasm in which it lies. It increases its volume, but 
oes not immediately divide. Uninucleate embryo sacs are found 
in June, and binucleate ones (fig. 8) in July. By the latter part 
of the month as many as 64 nuclei may be present. Fig. 8 is that 
of a sac with two nuclei; in fig. 9 there are 8; and in fig. 10 a few 
less than 64. From this it would appear that the early divisions 
are simultaneous. As a matter of fact, I do not know that they 
_ are, for I have never observed a single mitosis in any of the more 
than 500 preparations of the free nuclear stages available for study. 
These preparations represent more than 50 collections of separate 
cones and two or three times as many separate ovules. It seems 
exceedingly strange that they should be so persistently missed. 
The same difficulty was encountered in a study made of the mitoses 
of the pollen mother cells (x). In that study I made the suggestion 
that those mitoses may occur at night. It has not yet been practi- 
cable to test out this hypothesis, and I mention it here merely 
