BREAKDOWN OF SPINDLE 409 



are separated at anaphase is perhaps due to its developing too soon. 

 In Kniphofia Nelsonii (Moffett, 1932 a), a series of abnormalities 

 occur during or after meiosis, apparently according to the time at 

 which spindle degeneration sets in. This probably varies with the 

 position of the flowers in the spike. In the extreme case the 

 chromosomes form, instead of two nuclei at the end of the first 

 division, a number of scattered nuclei which develop separately. 

 If the first division proceeds regularly the same abnormality may 

 befall the second division, and where this also is regular a wall may 

 fail to form between two or all four of the daughter-nuclei. In 

 consequence, these, or the numerous irregular nuclei, often fuse 

 during the first mitosis of the single or double giant pollen grain. 

 Side by side with these abnormalities normal tetrads are formed. 

 Variation in the abnormality presumably depends on differences in 

 developmental conditions. 



In Datura and Allium the breakdown sets in at a definite period 

 and inhibits the second division so that unreduced pollen grains 

 are formed. Even here the result is not entirely regular. Normal 

 second divisions may occur. The highest degree of precision is 

 found as we shall see in parthenogenetic organisms in the sup- 

 pression of either the first or the second division, and there is no 

 reason to doubt that selected mutations have in these cases developed 

 an analogous genotypic control. 



Genotypic control of the spindle and of its relation with particular 

 chromosomes has played an important part, not only in the evolution 

 of parthenogenesis, but also in that of sex-chromosomes (q.v.). 

 Probably there are no flowering plants which do not exceptionally 

 fail to give four reduced nuclei at meiosis, under such conditions that 

 they survive in nature. Levan (1935 h), for example, finds giant 

 pollen-grains in nearly all species of Allium. Several workers have 

 by special treatment succeeded in suppressing the first or second 

 division regularly, even in some cases reducing the pairing of the 

 chromosomes in otherwise normal organisms (Table 65). Whether 

 pairing can be seriously reduced by external means without killing 

 the cells before meiosis is complete is, however, doubtful. 



These observations show that the mechanism of the spindle is, 

 in its origin, genetically and physiologically independent of chromo- 



