PARTHENOGENESIS 437 



although there were no archegonia. The failure of reduction was 

 evidently due, as in Marsilia, to the formation of restitution nuclei 

 following imperfect pairing of the chromosomes at the first meiotic 

 division (cf. Rosenberg, 1927, 1930). 



[d) Apospory with apogamy [Athyrium Filix foemina var. 

 clarissima Jones, Farmer and Digby, 1907 ; Nephrodium pseudo-mas, 

 Rich. var. cristata, Digby, 1905). In this type both the alternating 

 generations arise from unspecialised cells by purely vegetative 

 processes. 



These abnormalities may be variously combined with one another 

 in the same individual (Digby, 1905). 



(iii) Flowering Plants (Angiospermae). In the angiosperms the 

 same types of abnormality occur as in the ferns. Relatively few 

 of the great number of forms which are known to be apomictic 

 from genetic experiment have yet been studied cytologically. 



The following classification applies to the ways in which an 

 embryo may develop asexually in the angiosperms ; it is not a 

 final classification of the habits of species, for it is often found that 

 when sexuality fails it is replaced in several competing ways in the 

 same species. 



{a) Diploid Parthenogenesis. In the sexual relatives of most 

 angiosperms developing in this way the egg-cell with its eight 

 nuclei arises from the end one of the four cells produced by meiosis. 

 This is the " normal " type. A " S cilia " type, where it arises from 

 two of the four, is found in the sexual relatives of the partheno- 

 genetic Erigeron, and a " Lilium " type, where it arises from aU four 

 together, ,is found in the relatives of Balanophora. In many 

 sexual forms behaviour is variable and in an aberrant sexual form 

 of Leontodon the four cells sometimes separate and then later their 

 wall breaks down and they join to form one embryo-sac (Bergman, 



1935). 



Evidently the cells that are going to contribute to the embryo-sac 

 are distinguished from their sisters at the time when the first division 

 occurs without forming cell-walls and this time is often variable. 



With parthenogenesis it was formerly believed that an 

 " Alchemilla " type corresponding to the normal sexual forms 

 might stiU occur. If an unreduced embryo-sac were produced in 



