426 WILSON. [Vol. VI. 



I may call attention finally to the frequency with which the 

 three-segmented larva occurs among annelids, — a fact which 

 almost tempts one to compare it with the Crustacean Naupliiis 

 or the three-segmented Pycnogonid larva. In Nereis the three 

 segments are well developed by the middle of the third day, 

 while the fourth segment does not appear until about the 

 twelfth day. During the whole of this period the larva con- 

 tinues to swim, though latterly much less actively than at first. 

 The same pause after the appearance of the first three seg- 

 ments takes place in many other forms, and it is not easy to 

 avoid the conclusion that this curious analogy with the Nauplius 

 is more than accidental. 



VIII. Axial Relations of the Nereis Trochophore. 



The axial relations of the Nereis larva are of great interest 

 and seem to afford new evidence on a number of important 

 questions. At first sight they appear to present a number of 

 features which are irreconcilable with those of the typical trocho- 

 phore {e.g. of Polygordiits). When carefully studied, however, 

 these peculiarities are found not only to be easily explicable, 

 but even to afford new data for a consideration of the axial rela- 

 tions of the trochophore in general. In the fully established 

 trochopore (Diagram V, A, B) the mesoblast-bands (dotted) 

 extend straight forwards from the pigment-area {p. a.), at the 

 centre of the lower hemisphere, at right angles to the plane of 

 the prototroch. The subsequent elongation of the body takes 

 place in the direction of the mesoblast-axis, m ; i.e. parallel to 

 the mesoblast-bands ; and the anus is formed in the centre of 

 the space originally occupied by the pigment-area. The pig- 

 ment-air a, therefore, represents the posterior extremity of the adult 

 body. 



Thus far the larva agrees precisely with the typical trocho- 

 phore of Polygordins, Eupomatns, or Hydroides. As soon, how- 

 ever, as the ventral plate is considered, we encounter an 

 apparent contradiction. For the neural axis, n (which coincides 

 with the long axis of the ventral plate, of which the neural 

 plates form a part), is parallel to the prototroch, and hence at 

 right angles to the mesoblast-axis. For reasons given at p. 436, 

 it is impossible to doubt that the posterior teloblasts of Nereis 



