1905] FRYE & BLODGETT—LIFE HISTORY OF APOCYNUM 51 
hedral form, varying only in degree of rotation and in mutual adjust- 
ment. Both forms may thus occur in the same plant. However, 
the tetrahedral form is not common; most of the groups are like 
figs. rr and 12, similar to those found in Typha latifolia (3). A 
case like Zostera (4), with its long pollen mother cells dividing length- 
wise, makes it doubtful whether pressure is much of a factor in deter- 
mining the direction of the spindles. The spindles in the microspore 
daughter cells of A. androsaemijolium seem to lack definiteness in 
direction. 
The formation of the generative and tube nuclei occurs when 
the ovules are in the sporogenous cell stage. The division is not 
simultaneous in the same sporangium, nor even in the same tetrad; 
the division is complete by the time the embryo sac has reached 
its 8-celled stage. The generative cell is lenticular or fusiform, as 
in Asclepias. Two spherical male cells are formed about the time 
the embryo sac is ready for fertilization, and while the pollen is still 
in the anther. STRASBURGER (5) observed a small and a large nucleus 
in the pollen of Vinca major, and again found both nuclei in cultures 
of the pollen tubes. If these were tube and generative nuclei, Vinca 
differs from A. androsaemijolium in the time of the division of the 
generative cell. STRASBURGER also observed (5) a tube nucleus 
and two smaller ones, probably male cells, in Amsonia salicifolia, 
which seems to indicate that this one agrees with A. androsaemi- 
jolium in the time of male cell formation. 
At the base of each petal and alternating with the stamens are 
five glands resembling those in Asclepias. They originate shortly 
after the floral parts appear and are said to be nectariferous (6). 
The two carpels unite at their tips before ovules are formed, 
and just after sporogenous tissue appears in the stamens. The 
tips form a rounded lump or head with glandular epidermis over 
large portions of it, as in the Asclepiadaceae. The ovules are arranged 
in the same way as in the family just mentioned, and have the same 
form. The archesporial cell is of hypodermal origin and does not 
divide to form a primary parietal cell (figs. 15 and 16). A single 
integument deeply buries the nucellus and primary sporogenous cell. 
The latter divides into four megaspores, any one of which may become 
the embyro sac. In fig. 19 the innermost spore becomes the sac; 
