The nuclear components of the sex cells of four species of cockroaches. 509 
Statement of the Schreixer’s that sometimes more than two threads 
are involved in conjugation which, if true, would throw doubt upon 
their general Interpretation. The criticism of Meves and Gold- 
schmidt are similar to those of Fick and in addition, Meves finds 
the fibres emerging from synapsis too thick to represent simply two 
of the threads of the earlier phases, joined together longitudinally, 
while Goldschmidt adds that parthenogenetic eggs present the phe- 
nomenon of synizesis. In a recent paper, (’08 a) he describes the 
oöcyte synaptic stages of Distomum , where he finds no synapsis. 
the doubled condition of the chromosomes being due to Splitting in 
the sense of Flemming ( 80). In these cases, a complete fusion is 
believed to occur, previous to the longitudinal Splitting so that all 
individuality is lost. Boxxevie (’07, ’08), Farmer and Moore (’04), 
Allex (’04), Miyake (’05) and Cardiff (’06) agree in this conclusion. 
The present 'material seems to bear this out. 
Gregoire (’04) has suhmitted two reasons for bis belief in para- 
synapsis during the synizesis stage; 1, The more or less parallel 
arrangement of the presynaptic threads and 2, the sudden thickening 
of the synaptic threads. With respect to these arguments, Häcker 
(’07) answers that parallel arrangement is not invariable, while the 
thickening mayjust as easilybe interpreted as a Splitting, accompanied 
by growth, as in ordinary mitosis, while Boxxevie describes essen- 
tially similar conditions in Xereis, Enteroxenos and other forms where 
the so-called "heterotypical” mitosis is virtually the same as in 
ordinary somatic division. H. S. Davis, too, fails to find evidence 
for parallel conjugation during synizesis and among bis arguments 
he urges that the evidence from optical sections is not valid inas- 
much as a single loop may cut the optical plane more than twice 
and that his estimations of the number of such loops show that there 
is a smaller number of chromosomes present during these stages than 
in the spermVtogonial divisions. The observations of the present 
writer are opposed to those just given for there is no evidence of 
chromatin threads passing more than twice through a given plane and 
the number of such threads during the earlier stages is certainly 
much greater than in the later ones. In this respect, also, the 
writer must differ from those, such as Fick, Bouix and others, who 
find an increase in the amount of nuclear material during these 
stages. 
With respect to Fick’s contention that there are no definite 
threads formed during the earlier growth stages, which are composed 
