G. R. Wieland — Proembryo of the Bennettitece. 447 



this genus the oospore enlarges at the expense of the adjacent 

 tissue. Later free nuclei become very abundant in the central 

 region, and then disorganize, all the cytoplasm massing at the 

 base of the spore, and parietally, with a single parietal layer of 

 equi distantly imbedded nuclei, except at the base, where there 

 is some massing of nuclei. Still later the sac-like cavity of 

 this stage is partly filled up by tissue preceding suspensor 

 development. The proembryo of Cycas is, in a word, sac-like, 

 and the endosperm large, the size of the latter in a way corre- 

 sponding to the excess in size of the whole seed over that of 

 the Bennettitese. 



In GingJco, after repeated nuclear division of the oospore, 

 there is no parietal grouping, but instead the oospore enlarges 

 and comes to be compactly filled with undifferentiated cellular 

 tissue, in which proembryo, suspensors, and embryo are all 

 merged. This must clearly now be regarded as absolutely the 

 most primitive condition known amongst the existing gymno- 

 sperms. 



In the organization of the GingJco embryo, the mass of tissue 

 just noted as filling the entire oospore takes part, the endosperm 

 being directly invaded without the formation of suspensors. 

 Two cotyledons remarkably like those of the Bennettitese in 

 both size and general appearance are produced ; but their 

 earliest stages have unfortunately not been figured so far as 

 known to the writer. 



Comparison with the other gymnosperms shows that the 

 proembryo of the Bennettitese is unique in oocupying the 

 entire nucellus, although this character loses not a little of 

 its isolation from the fact that the nucellse of the existing 

 Cycads are almost of the same size, increase in the size of 

 the seed having been plainly bound up with endosperm 

 development. Again it is supposable that a progressive reduc- 

 tion of endosperm had taken place in the Bennettitse and was 

 perhaps a cause of the disappearance of the group. 



The most distinct agreement of the Bennettitean proembryo 

 is clearly with Gingko, long known to have much in common 

 with some ancient Cycadean ancestry or relationship. In 

 both these proembryos, as has been seen, large-celled homo- 

 geneous tissue fills the oospore, and the formation of dicoty- 

 ledonous embryos takes place without the intervention of sus- 

 pensors. The present discovery unmistakably determines for 

 the first time that the embryogeny of Gingko is the most 

 primitive amongst existing gymnosperms. 



Between the existing Cycads and the Bennettitese the com- 

 parison is a more general one, there doubtless having been 

 agreement in the early history of both, and the more general 

 facts favoring the inclusion of the Bennettitese within a single 

 great group, the Cy cad ales. 



