44 Genetical and Gytological Studies on Pisum 



some non-identical parts of chromatids are associated at metaphase. 

 Where terminaKsation does not occur, identical chromatids are associated 

 with the formation of chiasmata at those points where crossing-over has 

 previously taken place. 



Examining Janssens' diagrams with a knowledge of terminalisation, 

 it is seen that they are strictly accurate only if there has been no ter- 

 minalisation. If any movement of chiasmata away from the attachment 

 constriction has occurred, the points of interchange are not at the chiasma 

 as Janssens has represented them, but at some point between the chiasma 

 and the attachment constriction. Pisum has a terminahsation coefficient 

 of •36--48, so that there is probably a shght movement of the chiasmata 

 between diplotene and metaphase. Apart from this, however, Janssens' 

 diagrams afford a very satisfactory illustration of the chromatid arrange- 

 ment in Pisum. They show clearly that where crossing-over has taken 

 place the first division is equational for some parts of the chromosome, 

 and reductional for others. It is always reductional for parts at the 

 attachment constriction. 



Another theory of crossing-over put forward tentatively by Darlington 

 (1929 c, p. 52), and later abandoned by him, has lately been presented in 

 some detail by Sax (1930). In Sax's words "The present theory of 

 crossing-over is based on the fact that at diplotene there is an exchange 

 of partners between paired chromatids at the chiasmas and that between 

 diplotene and late diakinesis there is a reduction in the number of 

 chiasmas. When extensive movement of the chiasmas is prevented by 

 the close association or coiUng of the paired chromatids, any reduction 

 in the number of chiasmas must be due to breaks in the chromatids at 

 the chiasmas so that segmental interchange occurs between two chro- 

 matids. This segmental interchange between two homologous chro- 

 matids is the cytological mechanism responsible for genetic crossing- 

 over" (Sax, 1930, p. 207). This hypothesis cannot be directly tested in 

 the Pisum case, since it has not been possible to study diplotene stages. 

 However, the metaphase stage of Pisum resembles that of other species 

 in which the earlier stages have been examined, and the chiasmata 

 have been found to be relatively stationary, e.g. Fritillaria imperialis 

 (Darlington, 1930 a), Stenobothrus (Janssens, 1924). Hence the reduction in 

 number of the chiasmata between diplotene and metaphase is negligible, 

 and not therefore sufficient to account for the "crossing-over" observed 

 in this species. 



Moreover, where the stages between diplotene and metaphase have 

 been intensively studied, a movement of chiasmata towards the ends of 



