1912] STEVENS—HETEROSTYLOUS PLANTS 299 
dark staining bodies in the nuclear reticulum as small nucleoli, and 
thinks that conditions similar to those in figs. 30, 32, and 33 repre- 
sent a fusion of some of these smaller nucleoli with the large one. 
As it is impossible to arrange these stages in Houstonia with any 
certainty, there is no proof as to what actually takes place. The 
appearance strongly suggests a ‘‘budding off” of material which 
is caught up by the nuclear reticulum; and the presence of numerous 
dark staining bodies in the reticulum at the time of synapsis (figs. 
34 and 35) have been held to show that such is actually the case. 
There is no proof, however, that the dark staining bodies which are 
present in the reticulum at this time have any connection with the 
nucleolus. In fact, a series of somatic stages, taken from the 
rapidly growing tissue of a young ovule (figs. 36-42), seems to 
show that the dark staining masses appearing in the reticulum pre- 
vious to the formation of the spindle have no connection, at least 
directly, with the spherical bodies observed near the nucleolus, 
but that they are chromosomes which become differentiated from 
the nuclear reticulum during the prophase. 
SYNAPSIS.—Synapsis is characterized by a crowding together at 
one side of the nucleus of the entire nuclear reticulum (figs. 34 and 
35). No structure can be made out in this mass except that it 
consists of a number of dark staining bodies in a much lighter, 
rather indefinite network. Synapsis is of considerable duration, 
and during this period the nucleus undergoes a marked increase 
in size (compare figs. 34 and 35) and the cell becomes rounded 
(fig. 43). 
SPIREM STAGE.—The chromatic mass comes out of synapsis in a 
series of thin loops (figs. 43 and 44). Each loop apparently con- 
sists of a single thin thread which does not take the chromatic stains 
uniformly throughout its length, but shows numerous dark staining 
bodies, connected by paler linin portions. As these loops shorten 
and thicken they stain more uniformly (fig. 45). 
The synaptic knot loosens very slowly, and even at the period 
of greatest thickening of the loops there is still a considerable por- 
tion of the nuclear mass, which does not show any definite structure, 
surrounding the nucleolus (fig. 45). This condition can be explained 
only on the supposition that some parts of the nucleus often pass 
