20 BOTANICAL GAZETTE [JULY 
axis, which stands in ,the axil of a bract or leaf, or merely 
arises directly from the main axis among the bracts but without 
a fixed relation to any of them. Thirdly, one may apply the 
theory of CELAKOvSKI (16) of the pine cone to it, and suppose 
that the seed-bearing axis really represents a branch, in the 
axil of a bract, so intimately united with a sporophyll, which 
itself bears the seed, that no traces are left of its complex 
nature. 
Without attempting to review the extensive and well known 
literature relating to this third theory, the writer is disposed to 
admit that it offers a reasonable explanation of the cones of the 
Abietineae. It seems much less probable when applied to the 
araucarian cone or to the ovulate structures of the podocarps and 
taxads. The attempt to explain the cordaitean cone according 
to it would appear to be beset with very many grave difficulties. 
In the first place, most of the evidence used to support it for the 
modern forms is here unavailable. In the second place, there is no 
indication in the cordaitean cone itself of such a union of branch 
and sporophyll. In the third place, it is hardly to be supposed that 
if such a process had taken place in the ancestors of the pines, 
there would be still in the present geological age clear indications 
of it, and that the paleozoic ancestors would have apparently gone 
so much farther than their modern representatives as to have 
made their cone appear even simpler than any of them, including 
even the apparently simple araucarians. Evolution plays strange 
tricks, it is true, but it really puts a considerable strain on one’s 
credulity to believe, as I think we must if we accept the theory that 
araucarians are derived from Cordaiteans through the Abietineae, 
that a complex branch system was reduced to a cone in the paleozoic 
-cordaiteans, showing practically no trace of its complexity, then 
reverted in the abietinean descendants to a stage where the evi- 
dence of complexity is again clear, and finally passed on into the 
araucarians, where the evidence of complexity is again at least 
doubtful. 
It seems to the writer far simpler to make no such difficult 
assumptions, but to consider that the cones of Cordaites and 
Araucaria are no more complex than they appear to be. In any 
