es oF 
= 
1907] GATES—POLLEN DEVELOPMENT AND MUTATION 93 
certainty, but it appears probable that it is less. BEER (3) states that 
the reduced number of chromosomes in O. longiflora is seven; hence 
these two species probably agree in the number of their chromosomes. 
Frequently the chromosomes in the late prophase show evident pairing, 
but it is usually difficult to represent this accurately in a drawing, 
because the members of a pair generally lie in different planes. 
However, indications of such a pairing are shown in figs. 21 and 22. 
The positions of the chromosomes make it certain that this is an 
apposition and pairing of whole chromosomes and not a splitting. 
This is evident, for the chromosomes are completely separated when 
first seen in prophase and are only arranged in pairs later, on the 
equatorial plate. 
Fig. 22 shows another prophase with ten chromosomes, no hetero- 
chromosomes, and a nucleolus in which the stain has been with- 
drawn from its outer shell except on one side. This condition of 
the nucleolus is common enough to cause one to suspect that it pos- 
sesses some significance. Possibly the side of the nucleolus in which 
the amphipyrenin retains the stain is the one which was earlier 
attached to the nuclear wall. However I have no observations to 
prove this. The nucleolus now rapidly loses its staining power, and 
in the late prophase, after the chromosomes are formed, frequently 
stains only with orange G in an iron-haematoxylin-orange stain. It 
gives no indications of loss of shape or size before it disappears, 
and I think the most probable supposition is that it is quickly dissolved 
in the cytoplasm when the nuclear membrane breaks down. The 
evidence is clear that it disappears about this time, and it is very 
rarely seen as late as metaphase. I may state here that the stages 
of its reappearance in the reconstructed daughter nuclei at the end 
of the heterotypic mitosis are in the reverse order of its disappearance. 
Soon after the nuclear walls appear around the daughter nuclei, 
when the nuclear sap has increased in quantity, the nucleoli appear 
(fg. 42) as pale yellow-staining bodies, usually several in number 
and frequently attached to the nuclear wall. It is impossible to 
believe that these bodies are reconstituted from fragments of the 
original nucleoli thrown out into the cytoplasm during the first mitosis 
and collected again into the daughter nuclei, as described by ScHarr- 
NER (26) for the megaspore divisions in Lilium philadelphicum. For 
