400 
AMERICAN JOURNAL OF BOTANY 
[Vol. 9. 
nucleus in diakinesis has displayed it beyond question; so that it seems 
perfectly clear that this grouping together of four chromosomes is a natural 
and regular proceeding. In the second place, a study of the various ways 
in which the bivalent rings are found linked to this circle and to each 
other in early diakinesis has made it seem very probable that here also 
we are dealing with a very definite arrangement. No more than three 
rings have ever been found attached to the circle of four. No more than 
two of these rings have been seen in a nucleus with other rings linked 
to them, and there have been no instances observed in which more than 
two rings were certainly linked together (PI. XXVI, fig. 24; PI. XXVII, 
figs. 25-27). The presence of three rings linked to the circle of four has 
been quite often seen, however, and the two two-linked chains are fre- 
quently in evidence. These facts taken together would seem to indicate 
that before the bivalent chromosomes begin to separate, there are three 
pairs or rings which are linked to the circle of four, and that two of these 
rings have linked to them in turn each another ring, making five rings in all. 
Judging from the frequency with which the various combinations occur, 
it is probable that the ring which has no other ring attached to it is most 
often the first one to break loose from the combination. Occasionally, 
however, one of the other rings will come off first. 
Such a definite linking of chromosomes is especially interesting for 
two reasons. First, it suggests very strongly that at earlier stages also 
the chromosomes may have a definite position in the spireme, so that 
when the latter condenses to form the chromosomes, they will always 
find themselves linked in the same way. Such a uniformity of linking 
as is seen in early diakinesis can be accounted for in only one way, as I 
see it, and that is by assuming that when the chromosomes are yet in 
the spireme condition they always occupy the same relative positions, and 
that the individual threads of the spireme are also definitely placed, con- 
necting with each other at the same points and interweaving in the same 
manner in all nuclei. Furthermore, it seems to me likely that this definite 
arrangement does not begin with the spireme stage, but goes back through 
synizesis into the resting condition. It may be that the chromosomes in 
somatic cells also are always arranged in the same way. 
In the second place, this uniform arrangement is interesting because 
it furnishes a very good bit of evidence for the telosynaptic arrangement 
of the chromosomes and for the univalence of the spireme thread. If the 
homologous chromosomes were arranged side by side in the spireme, it is 
difficult to see how linking could occur. It could easily be brought about, 
however, if the chromosomes were to represent segments of the whole 
spireme placed end to end. The arrangement would then be brought 
about naturally in condensation of the spireme because of the original 
position of the threads which later become the chromosomes. 
The bivalent rings usually separate from each other and from the 
