Andrews, Karyokinesis in Magnolia aud Liriodendron. 139 



at the middle of each segment. Contrary then to the views 

 of Belajeff^), Atkinson'^) and others, — I have not seen, 

 after a careful study of many preparations a single indication 

 of what we would term a transverse division. Guignard^) 

 has also raised the objection tliat the regularity of the chronio- 

 somes in the V- and U-shaped forms is exaggerated, but in the 

 species of Magnolia in question the length of the legs in tliese V- 

 and U-shaped forms is quite constant. The only difference in 

 shape observed after the metaphase was a slight inequality in the 

 diameter of the legs (Fig. 3). When the Separation of the 

 daughter chromosomes is completc and they are at their respective 

 poles they still give the exact appearance of V- or U-shaped 

 figures. It was again observed in this stage (Fig. 5) as in the 

 one previously described (Fig. 3) that the chromosomes were 

 thickened considerably at the point of bending of the U's. 



After the chromosomes have reached the poles the connecting 

 fibres persist a while and then gradually disappear. The chromo- 

 somes soon lose their identity entirely when the daugther nucleus 

 passes into the resting stage which is complete and endures for 

 some time. No longitudinal Splitting of the daughter chromosomes 

 was observed in either the meta-or anaphase. The formation of 

 the division wall, after both the first and second mitosis in the 

 pollen-mother cells of Magnolia occurs in a peculiar way. A cell- 

 plate is not immediately formed after nuclear division. At the 

 close of the first mitosis the nuclear spindle increased somewhat 

 in width by the addition of new fibres and at the same time 

 there accompanied it a marked inward growth of the outer proto- 

 plasmic layer of the cell. This indentation always occurred at 

 the centre of the cell and spindle and continued to constrict the 

 cell until not infrequently it equalled a fourth of its diameter (Figs. 

 1, 8). No instances were found when it formed a complete partition 

 wall bj itself. The distance of inward growth and width of the 

 isthmus varied considerably in different cells and even in different 

 sides of the same (Fig. 8). Many preparations were examined 

 but the wall was never formed betöre the nuclear division or more 

 often tili near its close (Figs. 7, 8). A cell plate was finally 

 formed in the isthmus connecting the daughter protoplasts thus 

 completing cell- division. The formation of the cell plate of the 

 second mitosis takes place in about the same way as described 

 for the first division. Guignard*) has observed almost the 

 same method of cell walls formation in the two species Magnolia 

 Gulan and M. Soulangeana he has studied. Very soon, however, 

 the nuclear membrane reappears plainly marking of the cavity 



') Ueber die Reductionstheilung des Pflanzenkernes. (Berichte der 

 deutschen botanischen Gesellschaft. Bd. XVI. p. 27. 1898.) 



^) Studies on reduction in plants. (Bot. Gaz. Vol. XVIII. July 1899.) 



^) Les centres cinetiques chez les vegetaux. (Annales des Sciences 

 Naturelles. Botanique. Serie VIII. Tome V.) 



*) Les centres cinetiques chez les vegetaux. (Annales des Sciences 

 Naturelles. Botanique. Serie VIII. Tome V.) 



