446 G. JR. Wieland — Proembryo of the Bennettitece. 



clearly to be seen abutting on the wall of the nucellus. There 

 are also especially to be noted in the transverse sections several 

 irregular ribbon-like traces about the thickness of the cell 

 walls, extending quite across the large-celled mass, filling or 

 nearly filling the nucellus. These traces or rather surfaces 

 occur too often to be considered wholly accidental, but are not 

 supposed to be either suspensors, or tubular ospores, or cells 

 such as precede embryo formation in Ephedra. Their fuller 

 explanation doubtless awaits the preparation of more numerous 

 sections and the comparisons they may permit. In some of 

 the sections presumably cutting the upper half of the pro- 

 embryo, as already hinted, there is a suggestion if not a clear 

 indication that the mass of proembryo tissue was either less 

 dense in its central regions, or, that there was actually present 

 a small central cavity. This important point, which would 

 indicate a fundamental agreement with the existing cycads, 

 cannot be so readily settled as yet, since in no instance has a 

 longitudinal section been cut from a proembryo as well pre- 

 served as the two shown in the plate. In one longitudinal 

 section showing the lower two-thirds of a seed it is clear that 

 the lower half of the nucellus was closely filled by the typical 

 large undifferentiated cells making up the mass of the pro- 

 embryo. In another longitudinal section, the superior end of 

 the nucellus is seen to extend well into the tip of the seed, 

 which is quite filled with the characteristic large-celled pro- 

 embryo tissue. Unfortunately the middle region is in this 

 instance not conserved. 



There is nowhere a distinct indication of the presence of 

 endosperm, or of any differentiation of the large-celled tissue 

 filling the nucellar cavity, into an inner and outer zone. The 

 proembryo tissue appears to be homogeneous throughout, except 

 in one instance where some more elongate cells appear to rest 

 against the nucellar wall. It is, however, to be constantly 

 borne in mind that it is necessary to amplify the series of 

 sections. Structure will be found in many instances illustrat- 

 ing not only all the features of the proembryo, but in all prob- 

 ability the other stages of development, including possibly the 

 early stages of embryo formation ; although it may be years 

 before all the facts are learned, since it is so often the .fortu- 

 nate exceptional section which tells the story and yields the 

 reward for the cutting of sections where preservation proves 

 less clear. 



Meanwhile it is possible in the light of these newly discov- 

 ered proembros to make several highly interesting comparisons 

 with existing gymnosperms. The proembryo was a term first 

 used by Treub* in describing the embrogeny of Gyoas. In 

 * Ann. Jard. Bot., Buitenzorg, ii, 1881, and iv, 1884. 



