1915] BURLINGAME—ARAUCARIA BRASILIENSIS 27 
Moreover, it requires the further derivation of other structures 
(proliferation of the prothallial cells) to compensate for the dis- 
advantages. 
It is comparatively easy to derive the pollination apparatus 
of the ordinary conifers from the araucarian type by réduction, 
for it can be shown that each step would be an advantage, and so 
likely to be retained whenever it chanced to occur. In the first 
place, any change that would bring the pollen grains nearer the 
micropylar end of the ovule would shorten the distance to be 
traveled and so be an advantage. Agathis shows such a change, 
and there are abundant reasons for thinking it less primitive in other 
respects than Araucaria. Saxegothaea shows a still further stage 
of reduction, and there are also good reasons of other sorts for 
believing that it too is related to the araucarians and derived from 
them. The podocarps and pines illustrate the final stage where 
the pollen tube forming grains reach the inclosed nucellar tip 
before germinating. Once the pollen was deposited directly on 
the nucellus, changes tending to cover the nucellus by the integu- 
ment and to draw the pollen down the micropyle by means of 
a pollination drop would be further advantages in the way of 
further protection to the germination tubes from drying, as well 
as some advantage in closing up the micropylar orifice in the 
maturity of the seed. 
As we have seen that it would be easy to derive the Pinaceae 
from the araucarians so far as the seed and pollination habits are 
concerned, and next to impossible to reverse the order, we may now 
inquire whether it is possible to derive the araucarians from the 
Cordaitales directly in respect to the same structures. There are 
abundant evidences that among paleozoic gymnosperms of both 
great groups the nucellus either protruded from the nucellus or 
projected far into it. So far as I have been able to find from the 
literature available to me they all show pollen chambers. I suspect 
that this preponderance of evidence in the published accounts is due 
in part at least to the general opinion that pollen chambers are 
primitive, and so this feature has been exploited. It is conceivable, 
at any rate, that some of the paleozoic gymnosperms did not have 
pollen chambers and were not pollinated on the nucellus, but on the 
