//. The Ovule of Stangeria paradoxa. 303 
sporangium which still opens to the outside, so that free 
access is given to the spermatozoids to reach the megaspore 
though the latter is not liberated. As regards the structure 
of the opening, it can only be tentatively suggested that the 
pollen-chamber, with its thick-walled cells, may correspond 
to the ring of thick-walled cells around the apical cap of the 
microsporangium. 
Whatever the modifications were that led to the origin 
of the Cycadean type of ovule, no direct evidence of their 
nature is available at present, and it is thus necessary to 
connect the facts by as reasonable an hypothesis as possible. 
It may be pointed out, however, that the view above 
expressed, which rests primarily on the similar position held 
by the ovule of Stangeria to that of one of the sori of 
microsporangia, derives support from an examination of the 
relation borne by the individual sporangium to the sorus in 
some groups of Vascular Cryptogams. The reduction of the 
number of sporangia in the sorus to a single one occurs not 
infrequently in some species of Gleichenia and is the normal 
state of affairs in the majority of the Schizaeaceae , in which 
the monangial sorus has been definitely recognized. In these 
cases the variation is unconnected with heterospory, but its 
connexion with this is strikingly illustrated by Azolla among 
the Hydropterideae. It will be obvious that the reduction 
in number of the sporangia in a sorus, and along with this 
of sori on a sporophyll, would be still more necessary when 
the completion of the changes within the megaspore and the 
production of the embryo were all to go on at the expense 
of the parent plant. There would thus appear to be, on 
a priori grounds, good reason to assume that the need of 
such a reduction in number of sporangia in the sorus would 
have been even greater in the evolution of the ovule of 
Cycads than in the group of Hydropterideae , which are here 
cited as an unrelated but somewhat parallel case. 
A provisional examination of material of a number of other 
genera of Cycads has not disclosed anything inconsistent 
with the view of the nature of the Cycadean ovule here put 
