152 CHARLES J. CHAMBERLAIN 
and the Gnetales may be considered separately, although they have much 
in common. 
Familiar representatives of the Coniferales show more resemblances and 
fewer contradictions. The plant body is similar, and the internal structure 
of the stem often shows striking resemblances. The catkins of the Amenti- 
ferae may not differ much, morphologically, from some of the cones of the 
Coniferales; the pollen-tube structures of Angiosperms could be derived 
from those of Coniferales, and the embryogeny could be reconciled. The 
leaves are harder to reconcile, but leaves are very susceptible to environment. 
While these resemblances would not induce us to claim any Coniferales 
yet described as the ancestors of the Angiosperms, we believe they indicate 
the direction of the trail. We should remember that most of our Paleozoic 
and Mesozoic material is woody, and that there is a possibility — I believe 
there is a strong probability — that a great herbaceous vegetation has failed 
to be preserved, or, at least, has not yet been discovered. In such an herba- 
ceous vegetation, leading up to woody forms, I believe the missing links will 
be secured, and that the Angiosperms will be found to extend much farther 
back than any available records have indicated. 
The Gnetales show some striking Angiosperm characters. Most botan- 
ists, looking at the habit and leaves of Gnetum Gnemon, would call it a Dicot, 
and the histology of the stem continues the Dicot impression. In Ephedra, 
the habit, the strobili, and the spermatogenesis show Angiosperm features. 
It is so evident that the leaves have been reduced from more pretentious 
structures, that they need not constitute any objection. 
In this connection, the less said about the leaves and habit of Welwitschia, 
the better; but its flowers, especially the staminate flower with its sterile 
ovule, would pass for Angiosperm flowers. The only objection seems to 
be that definition relating to open and closed carpels. Fortunately we 
have reached a stage in botanical development at which definitions need 
not interfere with research ; for we do not put the Liliaceous Agapanthus in 
the Dicots simply because it has two cotyledons; or Nelumbo into the 
Monocots because it has only one cotyledon. So the open and closed 
carpel need not be absolute marks separating all Gymnosperms from all 
Angiosperms, and the presence of one condition or the other need not 
interfere with research into the origin of the Angiosperms. 
It is easy to be humorous and to say that an ancestor must be older than 
the offspring, and that, therefore, the Gnetales, with no geological record, 
could not qualify as progenitors of anything. But here, again, we must 
remember the possibility, or probability, of an extinct herbaceous flora, 
which, very late in its history, developed a few woody members. Earlier 
in its history, it may have given rise to herbaceous Gnetales and to primitive 
Angiosperms, which developed into the woody forms of the Cretaceous. 
We have tried to show that the Cycadophytes have come from the ferns 
and that they have not left any progeny, outside of the Cycadophyte line; 
