470 BREAKDOWN OF GENETIC SYSTEMS 



(cf. T>., 1928). (ii) All odd-numbered polyploids are numerical 

 hybrids with inherent irregularity of meiosis. Nearly all triploids 

 that have been examined, for example, form restitution nuclei 

 whether they are the product of hybridisation of two species or of 

 self-fertilising a diploid. 



High chromosome number (polyploidy) therefore favours the 

 conditions both of diploid parthenogenesis (by suppression of 

 meiosis) and apospory (by irregularity of meiosis). Taking this into 

 consideration, together with the fact that half the species of 

 angiosperms are polyploids, the relative frequency is intelligible. 



The reason for the association of a doubling of chromosome 

 number with parthenogenesis in Char a crinita (Ernst, 191 8), Artemia 

 salina (Artom, 1928 et al.), Solenohia triquetrella (Seller, 1927), and 

 Trichoniscus provisorius (Vandel, 1927) is fairly clear. Polyploidy 

 has prevented fertilisation, and therefore parthenogenesis has been 

 a condition of their survival. 



6. REPLACEMENT OF SEXUAL REPRODUCTION 



(i) The Development of Apomixis. Experiment and comparison 

 show that a failure of meiosis and the inheritance of special genetic 

 properties may in certain instances independently determine the 

 occurrence of apomixis. The problem of apomixis is the problem 

 of the part these two properties play in its origin. 



In the first place, it is clear that special genetic properties are 

 necessary, permitting the development of unfertilised eggs, 

 pseudogamously or otherwise, for parthenogenesis, and permitting 

 the development of adventitious embryos in the absence of successful 

 sexual embryos, for apospory and nucellar embryony. These 

 properties are probably characteristic of certain genera such as 

 Hier actum and Aphis. Secondly, it is necessary that unreduced 

 eggs should be produced, for diploid parthenogenesis, and 

 unsuccessfully reduced eggs, for apospory. It is this second 

 property that needs special consideration. 



For this purpose it is important to distinguish between the 

 structural conditions of the breakdown of meiosis and a genetic 

 property more clearly than has been done in the past. 



The discussion of the causation of apomixis brings into the fore- 



