HOW MEIOSIS GOES WRONG 405 



Partial Failure of Spindle and Cell-wall Formation at First or 

 Second Div^ision. 



Kniphofia Nelsoni .... Mofifett, 1932. 



Cannabis sativa (with environmentally 



controlled dioecism) . . . Breslawetz, 1935. 



Smith. 1935. 

 Beadle, 1933 {^)* 

 V. Berg, I934-* 

 Dobzhansky, 1934. 



Impatiens pallida 



Zea Mays (variable sterile) 



Triticum turgido-villosiim 



Drosophila psendo-ohscura 



Lathyrus odoratus (race with delayed first 



metaphase and complete terminalisation) Faberge 1936; Upcott, 1936, 



2. (a) Total Failure of Second Division Spindle. 



Datura Stramonium (stock derived from 



radium-treated plants) . . . Satina and Blakeslee, 1935. 



Allium, schcenoprasum (?^ + 3) . . ) t - i, 



A 4 /^ I \ h Levan, 19350. 



A. nutans {^x -f 2) . . . . j ^^ 



SphcsrocarpHS Donnellii (possibly). . Allen, 1935. 



I. or 2. Failure of Pairing at First or Second Division induced by 

 Treatment {cf. Table 65). 



(ii) Failure of Metaphase Pairing. In all cases of the genotypic 

 reduction or suppression of metaphase pairing there is some 

 pachytene pairing, except in certain parthenogenetic organisms 

 where, as we shall see, all relationship with meiosis, except position 

 in the life-cycle, is lost. The circumstances in which pairing fails 

 are clearly defined only in the " asynaptic " strains of Crepis and 

 Zea, in both of which metaphase pairing is variable and the frequency 

 of chiasmata varies from normal to none at all. In these plants 

 pachytene pairing seems to be fairly complete as a rule, but at 

 diplotene the chromosomes gradually fall apart. Their separation 

 is delayed, especially in Crepis, by the untwisting of the relational 

 coiling that has developed during pachytene. From this account 

 it might appear that the abnormality began with the failure of 

 chiasma-formation in a previously normal pachytene nucleus. But 

 it is found that where failure of chiasmata is absolute, pachytene 

 is also very defective, e.g., in Hevea. Where it appears complete, 

 chiasmata are always formed in a proportion of nuclei — a proportion 

 subject to great local variation. Thus in Crepis one bivalent may 

 have three chiasmata, while none of the other chromosomes in the 

 same nucleus have any : they are unpaired as soon as they are 

 uncoiled. 



* A tripolar spindle is sometimes formed at the first metaphase. 



