THE CYTOGENETICS OF HYBRIDS 369 



little or no synapsis in its spermatocytes. The chromosomes divided in 

 both meiotic mitoses, giving gametes with all of the 59 chromosomes. 

 No allotetraploid type was obtained, but allotriploids resulted from back- 

 crossing the hybrid to one parent. ^^ 



From the foregoing examples it can be concluded that amphidiploidy 

 may arise if the two unlike chromosome sets brought together in a cross 

 are compatible enough to exist together in the zygote and if the chromo- 

 somes then become doubled in some manner. The role of interaction 

 of homologous elements in the origin of such alloheteroploid types is 

 conceived by Winge (1917) as follows. When two specifically different 

 chromosome sets of like number are brought together, they may some- 

 times act in complete harmony at all stages of somatic development and 

 meiosis so that a new sexually reproducing hybrid type results. If the 

 compatibility of the chromosomes is less, the chromosomes, after having 

 cooperated throughout somatic development, synapse poorly or not at all 

 in the sporocytes so that diploid spores and gametes are formed. Finally, 

 if the compatibility is insufficient for the "action in pairs" favoring the 

 development of the diploid soma, all of the chromosomes in the zygote 

 or early embryo may split, the resulting halves then acting as homologous 

 pairs in the development of the resulting allotetraploid individual. Such 

 an individual may reproduce sexually as a "permanent hybrid" if it 

 produces diploid spores and gametes after meiosis. ^^ If meiosis and 

 sporogenesis fail, the tetraploid type may reproduce by asexual means 

 alone or disappear altogether. These several conditions are observed 

 in nature, and the interpretation outlined above harmonizes well with 

 certain phenomena in apomictic diploid plants (p. 407). 



In some cases there is a doubling of the chromosomes of one parent 

 only. Thus Saccharum officinarum 9 {n = 40) X S. sponianeum cf 

 (n = 56) gave a fertile hybrid with 136 chromosomes, the set of 40 from 

 officinarum evidently having doubled in the zygote. When this hybrid 

 was back-crossed as pollen parent to S. officinarum, the progeny showed 

 about 148 chromosomes, indicating that the officinarum chromosomes had 

 again doubled (Bremer, 1928a). 



Because of their cytological and genetical constitution, alloheteroploid 

 plants may be expected to differ in many ways from autoheteroploid ones. 

 An autotetraploid, for example, may differ from its corresponding diploid 

 mainly in a quantitative way, whereas an allotetraploid usually shows 

 qualitative peculiarities of character due to the combination of unlike 

 elements. Instead of forming quadri valents as in typical autotetraploids, 

 the allotetraploid types ordinarily form bivalents like a diploid, although 



^^ Federley (1913 et seq.). See his general account (1928). 



^^ Thus an individual of hybrid constitution may arise from the union of gametes 

 which are genetically similar (cf. p. 359), but here the gametes themselves are hybrid 

 in character. 



