THE SPERMATOGENESIS OF LEPIDOSIREN PARADOXA. 39- 
Summary. 
The different generations of cells composing the germinal 
epithelium resemble those often described in other forms, 
especially in the Amphibia. Very little arrangement of the 
different generations in different parts of the testis could be 
observed. 
The somatic number of chromosomes is thirty-eight. One 
pair of these is conspicuously larger than the rest. 
The reduced number of chromosomes in the bouquet stage 
appears to be arrived at by a parallel conjugation in the 
early prophase according to von Winiwarter’s scheme. 
In strepsiuema, which synchronises with the onset of 
synizesis, the conjugatits separate except at their ends, to 
form very long-drawn-out rings. 
During synizesis and diakinesis the rings break into their 
constituents, and the somatic number of univalent chromo- 
somes is again obtained, the “ homologous ” chromosomes 
being often widely separated from each other. 
During diakinesis each univalent becomes divided by a 
transverse constriction, which probably corresponds with the 
apices of the V’s of the pre-meiotic chromosomes, and also 
with the transverse division of the Copepod type of tetrad, 
which cannot therefore be taken as indicating the point of 
junction of two chromosomes united end to end. As in 
Copepods, the transverse constriction is not the division 
plane in either mitosis, but disappears during anaphase II. 
After the dissolution of the nuclear membrane “ homo- 
logous ” chromosomes are seen to approach each other, and 
join together a second time to form the rings or modifications 
of them found in metaphase I. 
The first maturation division separates entire “homologous” 
chromosomes. 
There is no resting stage between the two divisions. In 
the second division the chromosomes divide longitudinally, 
forming “ tetrads,” etc., very like those of metaphase I. 
A resting or semi-resting stage may be intercalated iuto 
