99 



believe that we have to deal with rudimentary archegonia; 

 able to undergo some few cell division (fig. 56 PL X) all 

 though they are never fertilised. 



The projection of fig. 63, PI. XI contained such an arche- 

 gonium-like body also, but the sections obtained through it 

 were not clear enough to be sketched. 



These archegonium-like bodies are by no means common, 

 yet they are not very rare either; after having once noticed 

 them I at first felt satisfied that the tubes approaching them 

 (fig. 63, PI. XI) were the poUentubes ready for impregnation; 

 we now know that these are the proembryo's. 



RELATION BETWEEN THE PROEMBRYO'S AND THE STERIL 

 PART OP THE EMBRYOSAC. 



Fig. 63 PI. XI shows very nicely the mode of approach of 

 the proembryo's; they try to push in between the coneshaped 

 projection of the prothallium and the nucellar tissue (cp. also 

 fig. 53 PI. IX, fig. 58, 60 PI. X) etc. If they succeed in 

 this they run over the surface of the coneshaped proces until 

 they reach its broad base at which point they penetrate into 

 the prothallium itself (fig. 61 PI. XI) or run over the surface 

 of it for some while yet before entering it. After this stage 

 has been reached the prothallium enlarges considerably, des- 

 troying all the nucellar tissue and obliterating the fertil part 

 of the sac, subsequently the fruit drops and the development 

 is at an end for the present. The dropping fruit consequently 

 contains besides the envelops of the flower nothing but a pro- 

 thallium on which a number of proembryo's have become parasitic. 



The formation of the embryo's proper occurs during the 

 sprouting of the seed and has been described by Bower, (see 

 historical review). 



Theoretical conclusions. 



It seems to me that the various investigations concerning 

 the groupe of the Gnetaceae prove them to be of very ancient 



