306 BOTANICAL GAZETTE [may 



ment of the egg takes place. Strasburger regards the parthenoge- 

 netic tendency of Eualchemilla as associated with excessive muta- 

 bility, which has weakened sexuality so that the process of fertilization 

 is being displaced by apogamy. 



Winkler (75, 76) reports an interesting case of apogamy in 

 Wikstroemia indica. He observed that seeds are produced apoga- 

 mously, in spite of the fact that some pollen matures. The apogamous 

 development of the embryo was proved by castration experiments. 

 He describes a peculiar phenomenon in the cells of the tapetal layer, 

 which usually contain two to six nuclei. These nuclei fuse into a 

 huge nucleus, whose mitotic figure often shows over 100 chromosomes; 

 but usually he counted 52 chromosomes. Although most of the 

 pollen does not reach maturity, a tetrad division with reduction is 

 present in sporogenesis, and in the heterotypic figure there are 26 

 bivalent chromosomes. The micropyle of the ovule is closed by the 

 elongation of the inner wall of the ovary during the formation of 

 embryo sac, which undoubtedly may have some relation to the apoga- 

 mous development of the embryo. A megaspore mother cell in this 

 form becomes directly an embryo sac, with the entire suppression of 

 tetrad division. His material was not sufficient for a cytological 

 study of the mitoses, and consequently he was unable to determine 

 the entire absence of a reduction division in the embryo sac, but it 

 seemed very likely that the egg which develops parthenogenetically 

 may retain the sporophytic number of chromosomes. 



Rosenberg (59) presents the result of cytological studies on six 

 species of Hieracium. He took up the species of Hieracium in which 

 apogamy was proved by the experimental studies of Ostenfeld (52), 

 and traced out their nuclear details. In Hieracium excellens the 

 nucleus of the pollen mother cell, after it has come out from synapsis, 

 presents a heterotypic figure with 17 bivalent chromosomes, but 

 often with an irregularity in the number of the bivalent and univalent 

 ones. When daughter halves of the bivalent chromosomes separate 

 and become grouped to form the daughter nuclei, the univalent ones 

 are left behind in the cytoplasm, as he had already observed in Dro- 

 sera (58). In H. flagellare a normal heterotypic mitosis takes place 

 in the pollen mother cell, the reduced number of chromosomes being 

 21. In these two species, the embryo sac develops after a tetrad divi- 



