954 Davis.—A Comparison of the Reduction Divisions of 
but certain important differences in our conclusions, evident from the text, 
will be noted. The following account will be written with the view to 
comparison with the foregoing description of Lamarckiana. 
Presynapsis. As in Lamarckiana , the daughter nuclei of gigas follow¬ 
ing the last mitosis in the archesporium (PL LXXII, Fig. 46) show clearly 
the sporophytic chromosomes, but the large number, twenty-eight, is 
naturally not easy to count with accuracy. As the nuclei increase in 
size a delicate reticulum is developed, the material of which is apparently 
drawn from the chromosomes since the latter decrease in size and finally 
can be recognized only as deeply staining granules (Figs. 47 and 48). 
These are the chromatic bodies and correspond to the similar structures 
in Lamarckiana. 
The large number of the chromatic bodies, together with the difficulty 
of distinguishing them in these stages from other granules in the nucleus, 
makes accurate counts impossible, but there can be no doubt that the 
chromosomes are represented in the nuclei by deeply staining bodies upon 
the nuclear reticulum in such stages as are shown in Figs. 47 and 48. No 
evidence was discovered that the chromatic bodies are grouped in pairs and 
such relations to one another as appeared were rather those of an end-to-end 
arrangement on the strands of the network. 
As the nuclear reticulum develops further (Figs. 49 and 5 °) the chro¬ 
matic bodies become so lost in the thicker strands of the network that they 
can no longer be recognized. As in Lamarckiana the development of the 
reticulum is a gradual process involving the thickening of what were at first 
relatively few delicate threads (Figs. 47 and 48), and the development of 
many others until the nucleus finally becomes filled with a rather dense 
reticulum (Fig. 50) in which lie one or more nucleoli. In later stages 
and just previous to the beginning of the synaptic contraction, the net¬ 
work becomes more like a loosely tangled group of threads (Fig. 51), and 
this condition is generally more clearly shown in gigas than in Lamarckiana , 
biennis , or grandiflora . 
Synapsis. The approach of synapsis is indicated here, as in other 
species of Oenothera , by a contraction of the nuclear reticulum away from 
the nuclear membrane (Fig. 53), and this contraction continues until almost 
all of the threads in the network are drawn together into a close mass, the 
synaptic knot (Figs. 53 and 54 ), generally at one side of the nucleus and 
near the nucleolus. The appearance of the synaptic knot is similar to that 
of Lamarckiana (compare Fig. 53 with Fig. 9); it consists of a contracted 
intricately coiled mass of threads whose arrangement cannot be followed, 
but from which occasional loops and strands extend into the nuclear cavity. 
The form of the knot may be almost spherical as in Fig. 53 or more 
irregular as in Fig. 54. As in Lamarckiana , biennis , and grandiflora the 
writer was unable to discover any relations between the threads other than 
