TETRAD AND Empryo-sac MITOSES 15 
ends of the chromosome pairs do not invariably fuse even when 
they take the form of rings. With the approach of metaphase the 
two members if separated become laid together. The fixation of 
the fibers then takes place somewhere between and not at the ends 
of the chromosomes, so that the possibility of the occurrence of a 
transverse splitting of the chromosomes at this time, as suggested: 
most recently by Schaffner * for Erythronium, seems to me to be 
excluded. 
If the ends of the chromosome pairs always fused completely 
it would probably be impossible to tell whether the breaks during 
anaphase occurred at the points of fusion or somewhere between 
these, When, however, such a fusion does not take place the 
simpler and more obvious interpretation does not involve the idea 
of a reducing division. 
The separation of the daughter chromosomes takes place with 
extreme regularity but curiously enough with quite constant differ- 
ences in the two genera. In Asperula the ends of the daughter. 
Vs separate simultaneously and the chromosomes take with little 
variation the form of a conventional heart. This is to be explained 
by the shortness of those bodies and the point at which the fibers 
are inserted (pl. 9, figs. 3, 22). 
In Crucianella, on the contrary, the ends of the V-formed seg- 
ments scarcely ever separate at the same time. The ones which 
remain united the longer are attached to each other with so great 
a degree of firmness that it is only after the greater mass of each 
daughter chromosome reaches about half the distance to the 
spindle poles that they finally separate (pl. 17, figs. 2, 3, 16, 17, 
18, 39). The forms taken by the separating chromosomes under 
these conditions have received special study, and all of them sug- 
gest very definitely the behavior of plastic masses under tension, 
` Examination of the detailed representations on pl. 77 (fig. 3, a— 
f; fig. 18, a-g) will reveal this point quite clearly. „Especially in- 
structive are such forms as shown in fig. 78, a, as they bear on 
the question of transverse division. Were the chromosome here 
breaking transversely and not undergoing separation at united 
ends, it is difficult to see why such a bulge at the point of separa- 
* A Contribution to the Life History and Cytology of Erythronium. Bot. Gaz. 
: 369-387. 1901 
