Timherlake — Sivarm-S pores of Hydrodictyon. 507 



An explanation of the meclianics of the process of cleavage 

 described above is by no means obvious. The best known 

 cases of cell formation from large coenocytic cells in plant 

 tissues are those of the formation of a layer of endosperm cells 

 from a large multinucleate! mother cell and those cases of spore 

 formation in certain fungus sporanges as described recently by 

 Harper. The Hydrodictyon cell has a striking superficial 

 resemblance to the multinucleate endosperm mother cell in 

 that in both cases there is a relatively thin layer of protoplasm 

 containing numerous nuclei and surrounding a large central 

 vacuole. There is also somewhat of a resemblance in the 

 process of cleavage in that a somewhat regular succession of 

 stages m.ay be observed in passing from one end of the cell to 

 the other. (See Strasburger, 23.) But at this point the re- 

 semblance ceases, and it may serve to bring out more sharply 

 the problem to be solved in case of Hydrodictyon if the differ- 

 ences in the actual process of cleavage in the two cases are 

 pointed out. In the first place in the embryo sac the cleav- 

 age is very clearly in direct connection with the nuclei tlie po- 

 sition of any one part of the cleavage plane being detei^ 

 mined jointly by the two nuclei that are connected by the fibres 

 in which that part of the cell plate is being formed. To be 

 sure the planes may not always be so arranged as to cut out 

 uninucleate pieces as Strasburger has shown for Corydalis 

 and other forms, but in these cases it is none the less true that 

 where the division does take place the cleavage plane is just 

 as distinctly determined by the nuclei in pairs. In Hydro- 

 dictyon as previously sho^vn, no such relation of cleavage 

 planes to nuclei is at all evident, with the possible exception 

 of those cases where the constriction cuts in between two re- 

 cently formed nuclei. Again, in the embryo sac the process 

 is clearly initiated and practically completed in the midst of 

 the protoplasm, the two boundary membranes being simply 

 divided into portions that form the outer and inner portions 

 of the membranes of the new cells, the lateral membranes be- 

 ing formed entirely anew from the spindle fibres, while in 

 Hydrodictyon the process is clearly indicated on either sur- 

 face by the limiting membranes and its completion depends 

 6 



