16 BOTANICAL GAZETTE [JULY 
either end of the cell near the nucleus. These areas are fre- 
quently to be noticed in the megaspore of Taxodium, but I have 
not been able to establish any definite relation between them 
and the spindle-formation (figs. 38-43). A slightly younger 
stage is shown in fig. 38, where the nucleus is of the usual struc- 
ture and has not approached the tip of the cell, in which posi- 
tion it is always to be found before the first division. Stages of 
the first division are shown in figs. go-42. The chromosomes 
were not counted, but are evidently not far from twelve in num- 
ber. This division cuts off a large lower cell and a much 
smaller upper cell. The lower cell immediately prepares to 
divide again. The second spindle is shown in fig. 43, and in 
jig. 44 the division is complete. The starch has begun to dis- 
appear during these divisions, but some is present until the 
conclusion of the second. Strasburger ('79) describes it as dis- 
appearing in Larix europaea before the second division; the same 
is true in L. sidirica (Juel, ’O1) and Pinus Pumilio (Coulter and 
Chamberlain, ‘ot). 
The upper of the two cells formed at the first division does 
not divide again, and its nucleus never reaches the resting stage, 
or indeed approaches it. fig. 42 shows the difference in the 
nuclei of the upper and lower cells of the first division. The lower 
is developing as usual, but the upper has formed no reticulum, 
and in fact never reaches a more highly organized stage. Its 
chromosomes remain fused and lumped, and soon present merely 
a disorganized, homogeneous appearance. This history of the 
upper nucleus is repeated in detail by that of the upper cell of 
the second division. There are thus formed in Taxodium only 
three cells from the division of the megaspore mother-cell, but 
as the lower divides twice, it is in every respect the equivalent 
of a pollen grain, as much so as if the upper cell of the first 
division had divided, as is the case according to Juel (’00) in 
Abies sibirica. Strasburger (’79) gives three, or seldom more, 
as the number of potential megaspores derived from the mother- 
cell in Taxus. He also gives the same number in Larix europaea, 
but as Juel has found four in L. sibirica it is possible that Stras- 
burger may have overlooked one in Z. . europaea. Coulter and 
