377 
Pollen Development in Lactuca. 
place. In Fig. 51. is shown a late prophase in which the chromosomes are 
just being drawn into the equatorial plate. Nine bivalents are present, but 
two of them are coalesced end to end. There are therefore eight bodies to 
be seen. Fig. 56 also shows eight, the end to end attachment of two being 
clear. Cells in the intermediate condition with eight £ chromosomes ’ appear 
to be quite as frequent as those with seven. 
Bands of darkly staining material frequently occur on the spindle, and 
are seen connecting some of the chromosomes in Figs. 51 and 54. In 
Figs. 55 and £ 6 the whole of the spindle is embedded in a mass of denser 
and more deeply staining cytoplasm. This feature was found to be typical 
of the wild L. muralis , but these two cells were the only ones found to 
exhibit the condition in the lettuce rogue and it has not been seen in the 
type. Fig. 55 represents an equatorial view of the metaphase spindle in 
which the nine chromosome bivalents are clearly separate. 
Figs. 52.53,54,57 show clearly the point of attachment of the spindle fibres, 
which appears to be terminal in every case. In the last figure the separation of 
the chromosomes has already begun. All indication of their bivalent nature 
has at this time completely disappeared. The line of demarcation between 
the two halves is usually obliterated during diakinesis. This may be partly 
attributed to the swelling action of the fixing fluid, but even apart from 
this the two halves of a bivalent must come into very close relationship, 
especially during the heterotypic metaphase. Again, the coalescence of 
bivalents at this time cannot be attributed to the fixation, for it occurs 
in cells where the chromosomes are otherwise loosely grouped. It represents 
rather a marked tendency to coalescence in which certain of the bivalents 
appear to be concerned. It is not a mere clumping of the whole chromo- 
some group. The chromosomes are riot sufficiently distinguishable at this 
period to make certain whether two particular pairs of bivalents only are 
concerned, but the observations lead us to consider that the shorter bivalents 
are usually concerned, though figures such as 51 and 56 indicate that 
bivalents of intermediate or maximum length may be involved. 
Figures such as 51 and 56 make it certain that the coalescence is an 
end to end one, at least in the cases which can be clearly determined. We 
have seen relatively few heterotypic telophases in which the chromosomes 
could be counted exactly, but these all showed nine chromosomes, indicating 
that the coalescence is a temporary one on the heterotypic spindle. 
These coalescences on the heterotypic spindle are much more numerous 
than during diakinesis, and frequently two occur in the same cell. From 
this it follows that while the fusions during diakinesis probably persist into 
the heterotypic mitosis, fresh coalescences also arise during that division. 
The possible genetic significance of these phenomena will be considered 
later. 
Fig. 57) which represents an early anaphase, shows the bivalent 
