I93I The secondary association of chromosomes 363 



It has long been known that homologous chromosomes may often 

 be found side by side in somatic metaphases. In the Diptera this 

 behaviour is normal and it has been observed in a number of plants. 

 If the affinity of homologues is evidenced by association at mitosis, 

 then a 'priori there is no reason why we should not expect to find it 

 at meiosis, especially if the conditions favour association. 



That the conditions do favour association at pro-metaphase has 

 already been shown, but apart from this, the general configuration of 

 the chromosomes at meiosis should not be overlooked. 



At mitosis they are usually elongated and moreover no stage is 

 found comparable to pro-metaphase of meiosis ; on the other hand the 

 highly, condensed, spherical shape of the chromosomes at meiosis 

 (especially in polyploid dicotyledons) permits of easier movement and 

 give a greater chance of association along the whole length of the 

 chromosome. 



In this respect it is noteworthy that secondary association is only 

 found in plants with small chromosomes, a fact which may be at- 

 tributed 1) to the difficulties of intimate association between chromo- 

 somes of large size owing to the interpolation or proximity of masses 

 of non-homologous chromatin, 2) and lack of space for free move- 

 ment. 



In the evidence from the literature to be dealt with later in this 

 paper reference is made to the similarity of size of associated bivalents 

 in not a few of the drawings. It has frequently been stated and by 

 numerous workers that size differences in polar view are of question- 

 able value. This statement however needs qualification. It is obvious 

 of course that if the bivalent is inclined to the angle of vision the size 

 of the chromosome will seem larger on casual inspection, but it is nearly 

 always possible to prove such an inclination by careful focussing up 

 and down. Again if a chromosome is not round in transverse section 

 but flattened in one direction then in side views its apparent size will 

 be different according to which side is presented to the eye. (cf. Law- 

 rence 1931, pp. 295). In this case comparison of several side and polar 

 views will usually reveal such non-spherical chromosomes. 



Differences of size in polar view at metaphase and, in particular, 

 at anaphase, may be important. As has been shown, bivalents with a 

 single terminal chiasma seem invariably to separate first at anaphase, 

 whilst those with two terminal chiasmata are always slower. This 

 may be ascribed to the greater length of chromatin which has to sepa- 

 rate at disjunction in the latter type. Bivalents with sub-terminal or 



