112 MEIOSIS IN DIPLOIDS AND POLYPLOIDS 



bivalent forms are to be found which correspond with different 

 stages of terminahsation, and it is evident from observations on 

 Rosa (Erlanson, 1931 c) that this is due to the process being 

 interrupted before completion by the onset of metaphase, after which 

 it proceeds no further. 



We have as the extreme type of terminahsation organisms in 

 which interstitial chiasmata have never been observed at metaphase, 

 e.g., Rhoeo discolor (Fig. 53, D., 1929 c ; Roller, 1932 c). 



Secondly, we have organisms in which a proportion of interstitial 

 chiasmata are found, e.g., Circotettix, Matthiola incana and Rosa. 



Finally, we have organisms in which terminal chiasmata occur 

 fairly frequently, but since there is no appreciable reduction in the 

 number of chiasmata between diplotene and metaphase the 

 occurrence of terminal chiasmata is presumably due to the 

 terminahsation of only the distal chiasma, accompanied no doubt 

 by slight movement of the proximal ones — e.g., Stenobothrns, 

 Mecostethus, Fritillaria imperialis and F. Meleagris (Figs. 25-28). 



All chromosome association at m^etaphase that is derived directly 

 from pachytene association may therefore be regarded as derived 

 merely by the formation and movement of chiasmata. Pairing 

 that is not derived from pachytene association will receive special 

 consideration (Chs. IX and XII). 



The table classifying bivalent types is based on these considera- 

 tions. The original distribution and later movement of the 

 chiasmata is inferred where, as in so many organisms, it has not 

 yet been observed, from the form of the metaphase chromosomes. 

 This kind of inference has been made in (Enothera (D., 1929 a and c), 

 where at the time interstitial chiasmata had never been observed. 

 Later work has shown such inferences to have been correct. 



The classification enables us to examine on a broader basis the 

 conditions affecting terminahsation and localisation. 



Thus it will be seen that the first two classes with minimum 

 terminahsation have all large chromosomes. The fact that amongst 

 these are also all those with localisation of chiasmata is not sur- 

 prising, since this depends on delay in pairing of parts of chromo- 

 somes such as can only occur where a great length is to be paired. 

 The fact that all types with the minimum terminahsation have 



