484 THE AMERICAN NATURALIST. (Vou. XXXIX. 
botanists of the Carnoy Institute, Grégoire and Berghs. ‘To 
the writer the conclusions of the second school seem better 
founded and we shall present them first. Allen’s last paper 
(:05) gives the most complete statement of their interpreta- 
tions. 
Allen's conclusions (:03, :05) are based on the study of the 
pollen mother-cell of Zilium canadense and his account of 
synapsis in this form is of great interest for the simplicity of 
his explanation of the events of this phenomenon and their sig- 
nificance. The nucleus of the young pollen mother-cell follow- 
ing the last mitosis in the archesporium and previous to synapsis 
contains a network of large irregular masses connected by fibers 
of varying thickness. The irregular masses, which probably: 
contain both chromatin and linin, are derived from the chromo- 
somes of the previous mitosis but these structures cannot be 
recognized in the resting nucleus. Nucleoli are present among: 
the irregular masses or chromatin knots but are readily dis- 
tinguished from them. As the nucleus grows larger the chroma- 
tin knots become more widely separated, but synapsis does not 
occur until it has reached its full size. 
During synapsis the reticulum becomes transformed into a 
definite spirem. The fibers connecting the chromatin knots 
increase in length and become more uniform in thickness while 
the knots become less conspicuous as though their material were 
drawn out along the fibers. The fibers of the reticulum are now 
seen to arrange themselves in pairs and a general contraction 
of the network takes place which is the beginning of synapsis. 
Allen believes that this contraction is associated with the approx- 
imation of the fibers. The contracting network occupies at first 
the center of the nucleus but later moves to the periphery where 
the „nucleoli may be found flattened against the membrane. 
There is now a continuous spirem in the nucleus, plainly com- 
posed of two slender threads lying side by side and probably ` 
with no free ends. These two threads often run closely parallel, 
sometimes loosely twisted about one another, sometimes in con- 
i pads e “oes it sometimes rather widely separated. 
lens a UM e nue of the thread is not due to a 
ependent threads are developed indepen- 
