1915] BURLINGAME—ARAUCARIA BRASILIENSIS 19 
probability of a real relationship between the araucarians and 
podocarps. 
Though it has been generally recognized by botanists that the 
protruding nucellus is correlated with the method of pollination 
and extensive growth of the pollen tube, it does not appear to me 
that this very peculiar situation has received anything like the 
attention that it deserves. I have elsewhere (3a, 3b) expressed 
the opinion that pollination of the scale, coupled with an extruded 
nucellus, is more likely to indicate the retention of an ancient habit 
than the acquisition of a new one. 
It must be admitted that we know comparatively little about 
the structures and affinities of paleozoic seeds and pollination 
devices. In the absence of present knowledge we must resort to 
more or less probable conjectures in our attempt to relate the 
already known facts. We do know enough, however, to make it 
very probable that the earliest known gymnospermous seeds are 
very far from being representative of the beginnings of the seed 
habit. They had already acquired numerous complexities. It is 
scarcely credible that the actual first seeds should have been pro- 
vided with a deep and narrow micropyle, with devices to draw the 
pollen grains down into it and on into a chamber specially prepared 
for their reception by the breaking down of the cells of the nucellus. 
It is further to be noted that seeds of this type have in their pollen 
chambers pollen grains that show no signs of having possessed 
pollen tubes. It seems evident that this complexity of devices 
must have had a more or less extended history, and that to under- 
stand it we must try to conjecture the conditions and structures 
that would have been likely to be developed as intermediate stages 
between heterosporous pteridophytes and these paleozoic gym- 
nosperms. 
It is not alone that we do not know the history of the seed 
structures of these early gymnosperms that makes the problem 
difficult. The difficulties of relating the structures known in more 
modern plants to these ancient ones is no less difficult. 
An analysis of the known facts will show that there are four dis- 
tinct methods of accomplishing pollination and fertilization now 
known among gymnosperms: (1) the Cordaitales and Cycadofilicales 
