APOMIXIS 319 



time of maturation of the egg and the entrance of the male gametes. 

 It is also possible that there may be two embryo sacs in the same 

 ovule. The egg of one of these may be fertilized and give rise to a 

 normal diploid embryo, while that of the neighboring sac is stimu- 

 lated to develop without fertilization. A second source of origin 

 cf haploid embryos may be some cell of the embryo sac other than 

 the egg. Usually this is a synergid, but rarely even antipodal cells 

 may give rise to embrycs. The third possibility is the origin of the 

 embryo from a male gamete. This may be due either to a degenera- 

 tion of the egg nucleus, so that the male nucleus alone is left to 

 function, or to failure of pollen tube to open, so that one or both of 

 the male gametes begin to develop in situ. 



RECURRENT APOMIXIS 



In the second or recurrent type of apomixis the embryo sac is 

 diploid and may arise either from a cell of the archesporium or from 

 some other cell of the nucellus. The distinction between the two 

 forms, known as generative and somatic apospory, is somewhat 

 artificial, as it is frequently difficult to say whether the cell in ques- 

 tion belongs to the archesporial tissue or to the "somatic" tissues 

 of the nucellus. 



Since there is no reduction in the chromosome number, the first 

 division of the initial cell is mitotic, semiheterotypic, or pseudo- 

 homo typic. The mitotic division requires no explanation except 

 that in apomictic plants there is a retardation of the prophase and 

 the resting nucleus shows an intense hydration of the chromosomes. 

 They do not exhibit either the pairing or the marked contraction 

 characteristic of meiosis, and at anaphase the two halves of each 

 chromosome move apart to the poles (Fig. 182M-R). The semi- 

 heterotypic division begins like a meiotic prophase, 8 but a normal 

 metaphase plate is not organized; instead the chromosomes are 

 widely scattered on the spindle. In the anaphase there are numer- 

 ous laggards, and eventually a nuclear membrane is laid down in 

 such a manner as to incorporate the entire chromosome complement 

 within a single "restitution nucleus" (Fig. 182A-G). This nucleus 

 is somewhat dumbbell-shaped in the beginning, but it soon becomes 



8 A pairing of the chromosomes and the occurrence of "bivalent-like formations" 

 have been noticed in some cases of semiheterotypic division, but chiasmata are rare 

 or absent (see Gustafsson, 1946). 



