536 



MARSILEACE^. 



the greater degree of incrustation of the yellow sac, and its 

 three-lobed upper end ; and in the absence of the cellular 

 lobes. The development of the smaller, more numerous, 

 secondary capsules of the two may be said to be absolutely 

 the same, the fact of their presenting in Salvinia simple pe- 

 dicels, and a single mass being, at least so far as origin is 

 concerned, of minor importance. 



Many points of this communication are nearly untouched, 

 but I have not lately been able to procure either of the two 

 genera in fructification. I have not yet observed any thing in 

 connection with germination ; after three months immersion 

 in water neither kind of organ has undergone any change. 



There are some points of the above observations which 

 seem to me to call for remark, before passing to the more 

 general ones, which it is the aim of this to establish. 



A curious discrepancy to reduce, appears to me presented 

 by the body and its lobes surmounting the yellow sac in 

 Azolla. 



Analogy perhaps suggests its being of the nature of the 

 masses enclosing the trifacial cells in the other kind of se- 

 condary capsules. For not to mention the similarity in 

 appearance between the rather numerous condensed points, 

 visible in the grume surmounting the sac in Azolla, with 

 those, in thegrumous nucleus of both kinds of secondary cap- 

 sules in Salvinia,* I have seen appearances of their being sur- 

 rounded by membrane. And it has appeared to me certain that 

 at least more such membranes have been observed than 

 the subsequent number of lobes, although I have seen 

 more than one point of condensation enclosed in one mem- 

 brane. So that their origin from confluence of several dis- 

 tinct parts, and the enclosure of one at least of the original 

 points in the mass so resulting seems, though perhaps ob- 

 scurely, to parallel the formation of the masses properly so 



* Which would perhaps justify the hypothesis of their being the rudi- 

 ments of so many parent cells. 



