370 W. J. C. Lawrence Cytologia 2, 



The association of univalents is dealt with in the discussion which 

 follows. 



Discussion 

 a) Fixation 



The quality of fixation in connection with secondary association is 

 of prime importance. How can we recognise a good fixation in plants 

 with small chromosomes ? In general terms we can say that we have 

 good fixation whenever the chromatin can be structurally related in all 

 stages of meiosis. The fixation is good if the fine details of structure are 

 clearly defined and no shrinking of the cytoplasm or general contraction 

 of the nuclear contents is found. " First " and " second " contraction 

 during prophase are the results of poor fixation. The appearance of 

 interconnecting strands at diakinesis likewise indicates poor fixation. 

 In particular no clumping of the chromosomes at metaphase should be 

 found. Now this is just the condition which might be confused with 

 secondary association, for whereas a really bad fixation may obviously 

 and excessively clump the bivalents, might not a better, but still poor, 

 fixation reproduce the grouping of chromosomes characteristic of se- 

 condary association ? 



Experience suggests that there need be little doubt of the answer 

 to this question if adequate study has been made. It, for example, the 

 disposal of the bivalents on the equatorial plate as seen in polar view is 

 such that although no bivalents are touching yet they are distributed in 

 groups which occupy the whole of the spindle then, apart from other 

 considerations, we can be fairly sure that fixation has not produced 

 clumping. 



Nevertheless, it must be insisted that secondary association cannot 

 be determined from inspection of a solitary metaphase plate. It is 

 probable that various accidents will sometimes associate bivalents, thus 

 simulating secondary association. The only proper criterion of asso- 

 ciation is the examination of a sufficient number of stages of diakinesis 

 and metaphase in order to arrive at an estimation of the constancy or 

 otherwise of the associations found. 



Variable penetration of the fixative will give reliable and unreliable 

 plates in the same section, and a given fixative may give good results 

 for one stage but poor for another. Just as in a diploid, every polar 

 view of metaphase will not show perfect dispersion of the bivalents, 

 for the same reason a parallel variation will be found in the polyploid. 



