50 COMPARATIVE ANATOMY 



Comparing the different members of the Annelids with one 

 another, one cannot fail to be struck by the general uniformity 

 of structure throughout the group, a uniformity which extends 

 even to developmental details. Metameric segmentation of 

 the body is characteristic of the phylum, and the nervous 

 system, the ccelom and its relation to the nephridia are 

 remarkably similar in the different classes into which it is 

 divided. It is significant, too, that the trochosphere larva is 

 of frequent occurrence, being absent only in those classes 

 which, like the Oligochseta, have taken to a terrestrial or fresh- 

 water habit of life, or, like the Hirudinea, are semi-parasitic. 

 There are weighty reasons for believing that animal life was 

 first developed in the sea, and then spread to fresh waters and 

 dry land. If this belief is well founded, special interest attaches 

 to the trochosphere larva so frequently interpolated in the life- 

 histories of marine annelids. It is regarded by many authors 

 as the representative of the ancestral form from which all the 

 Annelida (and, as we shall see, other groups of animals) have 

 sprung, its absence in terrestrial and fresh-water forms being 

 explained by the unsuitability of the conditions of life to such 

 a form of larval existence. How far the typical trochosphere 

 may be representative of the actual Annelid ancestor is, of 

 course, a question to which we are unable to give a positive 

 answer. It must be remembered that there are many forms 

 of polychaete trochospheres differing from one another in more 

 or less important details, especially in the arrangement of the 

 ciliated rings. Among so many forms it is impossible to fix 

 with certainty upon one and say, " This is the representative 

 of the Annelid ancestor." Moreover, it is certain that the 

 trochosphere, leading a free existence, and competing with 

 numbers of its own and other kinds for the means of existence, 

 must have been modified in the course of ages by the action of 

 natural selection. Hence we can only say that existing trocho- 

 spheres enable us to form a general idea of the ancestral form 

 from which all the Annelida are derived, but we must not regard 

 that form as having had an exact resemblance to any one 

 trochosphere. But the development of the worm from the 

 trochosphere proceeds on so nearly exactly the same lines in 

 all cases, or, when the trochosphere form is masked or sup- 

 pressed, the development offers such obvious resemblances to 

 the course followed e.g. in Polygordius that we are able to 



