442 BOTANICAL GAZETTE [NOVEMBER 
seeds in which the embryo is beginning to break the stony layer, the 
membrane reaches a thickness of 9-10, but the membrane is homo- 
geneous, there being no differentiation into layers as in earlier stages. 
The membrane clearly consists of two layers, which may be called 
the endospore and exospore (jigs. 12,18). The outer layer in young 
ovules is three or four times as thick as the inner, but in older ovules 
the difference is not so great. The inner layer seems perfectly homo- 
geneous under the highest magnification. The outer layer, under 
moderate magnification, looks somewhat like the dense outer layer 
of the megaspore membrane of Marsilea; but with a Bausch and 
Lomb ;}, immersion and Zeiss oc. 12 the structure appears to be com- 
paratively loose (fig. 18). The entire outer layer consists of club- 
shaped bodies. In other gymnosperms similar bodies have been 
described as prismatic, but in Dioon they really consist of a globular 
or ovoidal outer portion connected with the inner layer of the mem- 
brane by a stalk. In surface view these bodies are seen to be regu- 
larly arranged and to be quite uniform in size, the diameter of the 
head in November ovules being about 0.85 #, and in the following 
February about 1. As will be seen from the figure, the stalks form 
a comparatively open structure between the heads and the endospore. 
This open region appears as a nearly black line when seen in sections 
more than 1 or 2» in thickness, so that there seem to be three layers, 
the extra layer being only an optical effect caused by the stalk region. 
The membrane covers the entire female gametophyte, but after 
the pollen chamber breaks through the base of the nucellus the portion 
of the membrane covering the archegonial chamber is ruptured, so 
that the two chambers form one continuous cavity. This cavity is 
moist, but is not filled with a liquid. Only a few chemical tests were 
made and the results agreed with those of Tomson (21), who found 
the outer layer of the membrane to be suberized, while the inner layer 
iss uberized only where it is in contact with the outer layer. The inner 
portion of the inner layer, next the gametophyte, consists of cellulose. 
That the megaspore membrane is a vestige surviving from an ancestry 
which shed the megaspores seems too probable to be questioned. 
If it could be assumed that the ancestors of all the living cycads had 
megaspore membranes of equal thickness, and that in the surviving 
forms the membrane has been reduced uniformly in all genera, then 
