THE SPERMATOGENESIS OF LEPIDOSIREN PARADOXA. 29 
Bearing on the Conditions in Lepidosiren upon the Nature 
op the Copepod Type op Tetrad. 
In regard to the simple Copepod type of tetrad, apart from 
the fact that Lerat describes its formation according to the 
parasyndetic scheme (though Matschek in his more recent 
work denies this), it now seems very unlikely that it is the 
transverse and not the longitudinal slit which represents the 
point of contact of the conjugants. Nevertheless, in spite of 
the fact, already mentioned, that he has demonstrated that 
the transverse joint plays no part in either division, Matschek 
still maintains that it is an indication of an end-to-end, but 
permanent, junction of pairs of chromosomes. Hacker has 
quite recently (1910) re-stated his views regarding the meaning 
of the Copepod maturation divisions, in accordance with the 
latest work on this group. He thinks that the numerical 
reduction is brought about by the end-to-end junction of 
pairs of chromosomes, of which the transverse joint is the 
outward sign. Since the pairs of chromosomes are divided 
in each division in a plane running longitudinally through 
both chromosomes, it follows that each gametic nucleus gets 
the full number of chromosomes, joined however in pairs to 
form the apparently reduced number. This pairing is a per- 
manent one, and the fusion of the conjugants is presently 
complete, as shown by the fact that the transverse joint 
gradually disappears. 
If it be granted that the transverse constriction of the 
quadripartite chromosomes of the metaphase of the first 
meiotic division in Lepidosiren (i. e. the constriction parallel 
with the spindle axis, which is the same thing as the 
transverse constriction of each diakinetic univalent) is homo- 
logous with the transverse joint of the Copepod tetrad, it 
would be impossible to agree with Hacker’s scheme. There 
seems no reason to doubt the homology. The resemblance is 
close and the behaviour identical, i . e. appearing in diakinesis 
and persisting through the maturation divisions without acting 
