THE LIVING CYCADS AND THE PHYLOGENY OF SEED PLANTS 1 49 
If we consider only the nine genera of living cycads, as we know them 
today, the answer is easy: they are not responsible; they are the last of 
their race, restricted in geographical distribution, restricted in numbers, 
and struggling for their very existence. 
To some this may seem like too positive a statement; but if it should 
be challenged, we should ask, "To what could the cycads have given rise?" 
The only possibilities are the Cordaitales, Ginkgoales, Coniferales, Gnetales, 
and the Angiosperms. 
The Cordaitales, as the ancestral stock of the Coniferophyte line, might 
be expected to show resemblances, if any were to be found; but in habit 
they are very different from the Cycadophytes. They are the forest types, 
while the Cycadophytes bore somewhat the same relation to them that the 
ferns of today bear to the forests in which they occur. The leaves are 
prevailingly simple, contrasting sharply with the prevailing pinnate or 
twice pinnate leaves of the Cycadophytes. Not enough is known of spore- 
producing members in the Cordaitales, to make safe comparisons, but the 
Cordaitales certainly had well-developed cones;' so that, in this feature, 
they had progressed far beyond the Cycadofilicales. The fact that the 
cones were compound, while those developed later in the Cycadophyte line 
were simple, would indicate that the Cordaitales were from a different 
stock. We believe the available evidence indicates that the Cordaitales 
have come directly from the Pteridophytes ; but whether they have come 
from the fern section or from the lycopod section is a problem in the solution 
of which morphological characters of still undetermined value are balanced 
against each other. 
In the Ginkgoales, the pollen-tube structures, with the two motile 
sperms, present a startling resemblance to the corresponding structures in 
the cycads, even to the blepharoplasts developing into spiral ciliated bands, 
the peculiar behavior of the persistent prothallial cell, and the haustorial 
habit of the pollen tube. The extensive free nuclear period in the develop- 
ment of an embryo with two cotyledons is common to the cycads and 
Ginkgo; but here the resemblance ceases. The plant body and the strobili 
make relationship seem impossible. As far as the Mesozoic cycads are 
known, they afford no better Ginkgo resemblances. 
In my opinion the Bennettitales are no more nearly related, although I 
once tried to compare the long-stalked ovules of Ginkgo with the ovulate 
strobilus of the Cycadeoidea type. 
Even if we go back to the Paleozoic Cycadofilicales, it seems no easier 
to establish a relationship. Besides, the Ginkgoales can be accounted for 
quite naturally as an offshoot from ancient Cordaitales stock. 
A relationship with any of the Coniferales would be even more difficult to 
establish. Corresponding structures are too contradictory. The large pin- 
nate leaves of the Cycadophyte line do not compare well with the small, entire 
leaves of the Coniferophytes; nor does the unbranched trunk of the former 
compare well with the profusely branched trunk of the pines and Ginkgo. 
