1898] THE ORIGIN OF GYMNOSPERMS 167 
SUMMARY. 
1. A great Cordaites plexus, more extensive than the one 
usually included under that name, represented the characteristic 
Paleozoic seed plants. 
2. It was probably derived from homosporous-eusporangiate 
Filicales, represented today most abundantly by the Marattia 
forms and their allies, and was the most common Palzozoic 
type of Filicales. 
3. From it the gymnosperm lines, at least the cycads and 
conifers, were derived, the usually recognized Cordaites repre- 
senting a transition stage towards conifers. 
4. The frequent independent appearance of hetérospory is to 
be expected, as it probably results from inequalities of nutri- 
tion in connection with the development of antheridia and 
archegonia. 
5. The retention of the megaspore, resulting in the seed 
habit, follows the extreme sterilization of the megasporangium, 
Which is attained with the organization of but one megaspore. 
With the development of a single megaspore imbedded in 
sterile tissue, shedding becomes mechanically difficult, unnec- 
essary, and even disadvantageous from the standpoint of nutri- 
tion. 
6. The retention of the megaspore was followed by the 
development possibly of seed coats, through the well-known 
effect of fertilization upon adjacent tissues ; by immediate germi- 
nation of the oospore, on account of the favorable conditions 
and the abundant supply of available nutrition; and by the 
checking of the developing embryo by the mature seed struc- 
tures, resulting in the characteristic intra-seminal and extra- 
seminal stages of germination. 
7. The first retained megaspores were doubtless directly 
exposed to the microspores, and in Cordaites and cycads a pol- 
len chamber of varying depth and extent is associated with the 
