466 Dr. C. Semper on the Identity in Type 



tral) median line, Ibut in a line exactly corresponding to an axis 

 which, in the form of an irregularly cellular cord, lies close 

 beneath (resp. above) the rudiment of the central ganglion. 

 This axis is comparable to the notochord. The muscle-plate 

 bends outwards from it in a cardiac direction (towards the 

 dorsum) round the heart and alimentary canal, and also in a 

 neural (ventral) direction round the central nervous system. 



This is the type of the Vertebrates. In Nais, just as in 

 them, a cellular cord indicates an axis, from which the animal 

 muscle-plates gradually envelop the alimentary canal on the 

 one side, and on the other the central nervous system deve- 

 loped out of the ectoderm. 



3. It is well known that every complete zooid of a chain of 

 Naids is developed by the coalescence of a body part, which 

 first appears, with a later-appearing cephalic part ; the latter has 

 usually only four (at most six), but the former from 9-24 seg- 

 ments. In both parts these segments appear according to the 

 laws of annelid -segmentation ; the first body-segment is in- 

 variably the oldest, and it coalesces with the fourth and youngest 

 cephalic segment. This difference in the formation of cephalic 

 and body-segments is here extremely sharply defined ; it 

 appears also in the larvee of marine Annelids ( Terehella ac- 

 cording to Milne-Edwards) , and reminds one of the analogous 

 but less clearly marked condition in the Vertebrates and 

 Arthropods. In both groups several new cephalic segments 

 (which are much yovmger than many of the body-segments) 

 interpolate themselves between the oldest body-segment and 

 the oldest cephalic segment or segments ; in both regions seg- 

 mentation begins in front and ends behind; so that here, as in 

 the Annelids, the youngest cephalic segment is next to the 

 oldest body-segment. 



4. In the cephalic part, the brain of the zooid does not origi- 

 nate in a dorsal medullary plate overlying the alimentary 

 canal, but it is developed by a division of the anterior end of 

 the ventral cord and the upward growth of the two halves of 

 the oesophageal ring around the gullet. In this growth the 

 two lateral ganglia chiefly participate, with, perhaps, a part 

 of the central one (it was not possible to determine this with 

 certainty in the specimens, requiring much difficult treatment, 

 which I have yet examined), and finally also some secon- 

 dary structures. 



There appear, namely (even, as it seems, in the forms with- 

 out eyes), either laterally or rather towards the ventral side, 

 two sense-plates, which unite with the oesophageal ring before 

 the latter has lost its cellular structure. Possibly (or even pro- 

 bably), therefore, three different cell-groups take part in the foi-- 



