ON THE DEVEL0FM15NT OF SAGITTA. 389 



Sagitta are similar in structure and formation to Cha)topod 

 set^_, which, however, are absent in Polygordius. 



One weak point in the above argument is that it is based 

 almost entirely on the anatomy of the adult, and that no 

 account is taken of embryology. There can be no doubt, 

 J^pwever, that the development of Sagitta is very much 

 abbreviated, as is indicated by the fact that after the 

 ccelomic and enteric cavities have been formed they are 

 closed by the compression of the whole animal, and only open 

 again several days later. Similarly, the reduction of the 

 eudoderm to a simple septum, the formation of the ventral 

 ganglion so greatly out of proportion to its future size, and 

 other considerations all point to the abbreviated development. 

 This, however, does not dispose of the objection that the 

 development of the Cha^tognatha and Annelida is of a 

 fundamentally different type. The production of the meso- 

 derm in the one case by archenteric diverticula, in the other 

 from pole-cells, shows a wide distinction, though it is pos- 

 sible that the difference is less fundamental than at first 

 appears. However, apart from the earliest stages, the 

 development of Sagitta differs very widely from that of the 

 typical ^innelid. There is never a stage which resembles the 

 Trochophore with its large segmentation cavity, its sense- 

 organ, cilia, and head-kidneys, all of which are conspicuously 

 absent in the young Sagitta. But in addition to the differ- 

 ences in embryology, there is grave reason to believe that 

 many of the apparent similarities between the Chffitognatha 

 and the Annelida are superficial, and do not indicate true 

 affinity. The theory is based on the supposed homology of 

 the three segments of Sagitta with Annelid somites, but this 

 homology is by no means certain. Metameric segmentation 

 has arisen in the animal kingdom in very various ways and 

 independently in different groups, and it is very probable 

 that segmentation apparently similar in character may arise 

 in a variety of ways. In examining the embryological 

 origin of the segments of Sagitta, the first thing that is 

 noticed is that they arise at veiy different times ; the head 



