34 BOTANICAL GAZETTE [JANUARY 
revert to the homosporous habit, it would seem reasonable to conclude 
that the plant had attained to heterospory gradually, a few of the 
better nourished sporangia each developing a few spores slightly 
larger than the rest. Other things being equal, these spores in 
germinating would naturally develop the larger prothallia, because 
of a larger amount of food material within the spore. We might 
expect that these prothallia would bear archegonia. Thus the deter- 
mination of the sex of the prothallia would be at once shifted from 
dependence on the chance external conditions after the spore germi- 
nates, as found in the homosporous ferns, to dependence on nutritive 
conditions while the spore is being formed. This, it appears to me, 
must have been the process by which heterospory made its appearance 
in Marsilia. 
I should also state that megaspores were noticed which were packed 
unusually full of starch grains (figs. 44, 45). These forms were 
larger than any megaspores that I have seen and had developed no 
perinium. I was not able to observe stages in their development nor 
to account for the phenomenon. 
Theoretical discussion 
If the doctrine of recapitulation be applied here, we may say that 
Marsilia can be so manipulated as to show in the development of its 
sporocarps the various stages in the production of complete hetero- 
spory, through the abortion of many spores and the special nutrition 
of a few. 
We know that the heterospory described by WILLIAMSON and Scott 
in such forms as Calamostachys Casheana began in a very modest 
way, the diameter of the megaspores being not more than three times 
that of the microspores in the most extreme cases. We are also told 
that in the genus Calamostachys, C. Binneyana is homosporous and — 
C. Casheana is heterosporous; but even in C. Binneyana abortive 
spores are found, and the remaining spores are larger in proportion 
as fewer develop in each tetrad, the diameter being fully twice 4% 
great when only one develops as when three remain. 
Wi1u1amson and Scorr point out that in C. Casheana, also; 
abortion is evident, but is confined to the megasporangia; in fact, 
their figures show no abortion in the microsporangia. They com 
clude, therefore, that heterospory was reached by the abortion of the 
