

tgor | THE LIFE HISTORY OF ERYTHRONIUM oho 
farther up among the masses of dark-staining material and are 
thus difficult to identify. 
In the liliaceous types the division of the generative nucleus 
takes place in the tube; in many Helobiae and some other 
monocotyledons, in the grain before it is shed. In Typha the 
generative nucleus does not divide in the grain, while in Sil- 
phium °(22), one of the highest types, the division is in the grain 
and the sperm cells are elongated and even coiled like spermato- 
zoids. Such elongated sperm cells are also common in Alisma. 
It appears, therefore, that progressive reduction of the male 
gametophyte has not been uniform in the various lines of 
angiosperms. Search was made for the division of the tube 
nucleus, as is common in certain species of Lilium (8) and 
Hemerocallis (13), but nothing was found. Such a division 
probably never occurs in Erythronium. 
The style has a large continuous: canal, from the stigma to 
the cavities of the ovulary, for the conduction of the pollen tube. 
This canal is lined by a layer of glandular cells for the nourish- 
ment of the tubes (figs. 73, rg). The pollen tube does not grow 
through any tissue until after it passes into the micropyle. It is 
not difficult to see how such an angiosperm could develop from 
a Symnospermous condition. 
THE DEVELOPMENT OF THE MEGASPORANGIUM AND THE 
REDUCTION DIVISION 
As stated before, the archesporial cell begins to enlarge about 
the first of October, and by December first the chromatin net- 
Work is very distinct and is being transformed into the continu- 
ous spirem (jigs. 22-24). Inthe following stages the nucleus 
becomes very large, and the same is true of the chromosomes. 
This makes Erythronium a favorable subject for the study of 
these structures. After December the nucleus probably goes 
into a partial state of rest until early in the spring, at which 
time development and division continue. During this period it 
will be convenient to call the cell a megaspore. The spirem is 
at first very long and slender and the chromatin granules are 
