92 



cells in a more primitive condition. Though a pity, this view 

 was wrong. 



Perplexing positions occupied by the embryosacs like the one 

 just described have the advantage of keeping the observer 

 continuously awake ; less nice places are occupied f . i. by those 

 marked R. M. in fig. 57 PI. X. 



This curious arrangement of the embryosacs, which offers 

 nothing extraordinary after it has once been explained, has 

 caused a great deal of perplexity at the beginning of the in- 

 vestigation. 



h. Internal changes. 



The young embryosac contains but one nucleus. This is 

 divided into two, those into two again and so forth at least 

 that is the way which is indicated by most authors. As however 

 there exists a fundamental difference between the micropylar 

 and the chalazal end of the embryosac of Gnetum Gnemon L., 

 it is very possible that the nuclei found in the upper half are 

 all descendants of the one daughternucleus of the original 

 embryosacnucleus, while those in the lower half are descended 

 from the other daughternucleus. The persistent absence of nu- 

 clear figures in my material of Gnetum Gnemon, which no col- 

 lecting at the most different hours of the day or night seems 

 able to remedy [a fact about which Karsten complains also], 

 has made it impossible to follow this question up, yet it is a 

 point of primary importance. I can but hope that in future 

 some lucky chance may procure a material which will allow 

 to settle this point. 



After a while the nuclei are all arranged in a comparatively 

 thin layer of protoplasma around an enormous central vacuole 

 compare f. i. fig. 27 PI. IV. After a while again the contour 

 of the embryosac takes somewhat biscuit-like shape arising 

 from the fact that a constriction takes place somewhere below 

 the middle (fig. 28 PL IV). This contraction persists as is seen 

 from (fig. 29 PI. IV, fig. 33, 34 PI. V, fig. 36 PI. V, fig. 43 PL VII, 

 fig. 45, PL VII fig. 46, 48 PL VIII, fig. 53, 54, 55 PL IX), until 



