GENOTYPE AND STRUCTURE 



399 



determined by centrosomes which thus keep pace in division with 

 the chromosomes. The spermatids are originally tetraploid through 

 double non-reduction and by three polymitotic divisions become ^2x. 

 The two abnormalities considered together show the independence 

 of the spindle development from chromosome division on the one 

 hand, and from cell division on the other. They indicate by their 

 consistency that some body like the centrosome, visible or invisible, 

 must always be responsible for its perpetual renewal. 



2. THE BREAKDOWN OF MEIOSIS 



(i) Evidence of Genotypic Control. We have seen how chromo- 

 some pairing may fail in genotypically and environmentally normal 



( } J ♦ I ^ J 



Fig. 122. — First metaphase and anaphase in two sister seedlings 

 of Loliiim perenne [n = 7) showing genotypic control of chromo- 

 some size. X 2000 (Thomas, 1936). 



conditions as a result of structural or numerical abnormahty, 

 especially in hybrids. We shall later consider the results of this 

 failure in detail. But we must first find out in what ways and to 

 what extent the course of meiosis may be affected by genotypically 

 abnormal conditions. The evidence by which we can decide that 

 genotypic abnormality is concerned and structural abnormality is 

 not, is of various kinds. 



First, there are true breeding organisms whose abnormality is 

 inherited on crossing as a mendelian character (e.g., Zea, Datura, 



