414 BREAKDOWN OF GENETIC SYSTEMS 



kind depending on the proportion of chromosomes that are paired 

 and are still able to behave normally, and on the time when the 

 centromeres of the unpaired chromosomes divide in relation to the 

 time of development of the spindle. Let us take first the extreme 

 case of the haploids or hybrids or genotypic forms with no pairing 

 at all. The univalent chromosomes come to lie on the spindle more 

 or less at random. Some of the chromosomes may then move to the 

 equator and begin to divide as at mitosis. In Triticum turgido- 

 villosum (4x) all of the chromosomes may do so. In T. monococcum- 

 dicocciim (3%) only a proportion do so (Mather, 1935). 



The same is true of one of the Drosophila pseudo-obscura hybrids. 

 In such cases the first division is normal. But since the chromo- 

 somes do not divide twice the second division is abortive. The 

 chromosomes do not form a metaphase plate and a single nucleus 

 is reconstituted. It may happen, on the other hand, that no7te of the 

 chromosomes reach an equatorial position at the first division. A 

 single restitution nucleus may then be re-formed and the second 

 division is usually successful. The intermediate condition, in which 

 some of the chromosomes make a plate at the first division, may 

 lead to the formation of a restitution nucleus at one or both divisions. 

 Thus one, two, three or four nuclei may result at second telophase. 



In the PygcBra hybrids with complete failure of pairing, all the 

 chromosomes divide at both divisions. Regular diploid gametes 

 are then formed in fours. This regular double division of the 

 chromosomes is not known in any plant. 



When none of the univalents move to the equator at the first 

 division, the spindle may stretch as it does between separating 

 anaphase chromosomes. But, instead of separating, it then merely 

 disperses the undivided chromosomes. Being scattered in this 

 way they may form two daughter nuclei or several— the number 

 being limited by the number of chromosomes and the size of the 

 cell (cf. Fig. 127). Each nucleus then goes through a second 

 division which is successful so far as the two main daughter nuclei 

 are concerned. 



When a moderate number of bivalents or multivalents are 

 formed the case is different. The spindle does not stretch until 

 the bivalents are divided and the regularity of their first and second 



