THE TIME'LIMIT 



99 



i.e., at some point between the ends of the chromosomes, not at the 

 ends (Figs. 25-28). They vary in number from one to as many as 

 thirteen in the pair of chromosomes or bivalent, but their range of 

 variation is characteristic for each bivalent of an individual under 

 constant conditions. 



This variation has certain characteristic properties in all 

 organisms. Its curve is unimodal. In some organisms where the 

 range of size is small, the mean number of chiasmata per bivalent is 

 approximately proportional to the length of the paired chromosomes. 



But in many species where there is a big discrepancy between 

 the sizes of the chromosomes, the smaller ones have a higher pro- 

 portional frequency {e.g., Stenobothrus parallelus, D. and Dark, 1932 ; 

 Hyacinthus amethystinus, D., 1932/, cf. Sato. 1934, and Ch. VII). 



This is not true of organisms in which new small chromosomes 

 have suddenly arisen. Thus in Fritillaria imperialis and in Trades- 

 cantia new small fragments fail to be paired at metaphase in a 

 proportion of cells owing to the failure of chiasmata to be formed. 

 It seems therefore that the regular pairing of very small chromo- 

 somes that are part of the characteristic complement in Stenobothrus 

 and elsewhere is a derived property permitting pairing of short 

 chromosomes without an unduly high frequency of chiasma 

 formation in the long ones. 



(iii) Localisation of Pairing. In organisms with very large 

 chromosomes, pairing at pachytene is sometimes incomplete, owing, 

 as we have seen, to the time-limit for pairing. In Fritillaria, 

 Mecostethus, and elsewhere, the pairing usually begins near the 

 centromere, and it is in the distal parts that pairing fails. Chiasmata 

 are formed in the paired parts and are thus localised near the 

 centromere (D., 1935 b ; White, 1936 a). In the extreme case one 

 chiasma is found in every bivalent, long or short. The numbers, 

 however, vary in different nuclei and a detailed study of the dis- 

 tribution of the chiasmata shows us indirectly how pairing varies in 

 different chromosomes, in different nuclei and in different species. 

 The chief conclusions to be drawn are that shorter chromosomes 

 often have an advantage over long ones in the time of pairing and 

 the ends of chromosomes have an advantage over the middles unless 

 the centromere is median. Thus chiasmata are sometimes found 



4—2 



