STUDIES. ON POLYCHAT LARVZ. 605 
Further, this ramus, when developed in connection with the 
reproductive processes, appeared in close association with the 
sexual cells, both in regard to the time of its appearance and 
to the position of the segments which came to bear it. In 
the present species of Odontosyllis, the period of pelagic 
existence has apparently been extended from the time of 
sexual maturity till it has become life-long; and in connec- 
tion with this, the time of appearance of the special natatory 
portions of the parapodia has been thrown back to quite an 
early stage in the development of the worm. The biramous 
parapodia are, moreover, very fully developed, the dorsal 
ramus being quite separate from the ventral one distally, 
and supported by a single stout aciculum, 
The fact that the appendages with a dorsal ramus occur on 
the gonad-bearing segments, and on not more than one 
segment anterior to these, shows them to be strictly com- 
parable to the biramous parapodia developed by those Syllids 
which lead a pelagic life at the period of sexual maturity ; 
they are, therefore, only primitive in that they have reverted 
to the fundamental biramous type of Annelid appendage. 
They have not been directly derived from similar appendages 
of the primitive Polychet stock. We have here, therefore, 
an example of the precocious development of an organ in con- 
nection with the precocious assumption of a particular mode 
of life. 
The ventral cirrus is apparently not present even in the 
adult of this worm ; certain parapodia may often be seen to 
bear a conspicuous cirrus-like outgrowth of the ventral ramus, 
but in others this is shorter and proves to be simply a pro- 
longation of one of the three lobes surrounding the base of 
the tuft of sete. 
Claparéde (1863, P]. XII, fig. 12) and De Saint-Joseph 
(1886, Pl. VIII, fig. 40) both describe larve of O. gibba in 
a much earlier stage than the young Odontosyllis described 
above. These are polytrochal, and appear to show no trace 
of the dorsal lobe of the parapodium. In De Saint-Joseph’s 
larva, however, only six segments are present, and in 
