206 Mr. A. Hyatt ow the 



in spite of the opportunity offered by the increasing thickness 

 of the mesenchyme, is very significant. It is not yet esta- 

 blished that the mesenchyme does receive some additions in 

 course of its growth from the endoderm and ectoderm, but, so 

 far as the histology is now understood, it is doubtful. 



In other words the Porifera are intermediate with regard to 

 structural composition between primitive larval individuals, 

 like the free larvse of all colonial types, and the differentiated 

 colonies which arise from such primitive individuals after 

 they become attached, as in the Hydrozoa. They contain all 

 the elements necessary for the formation of complicated 

 colonies ; but in consequence of the less differentiation of the 

 mesenchyme their primitive individuality is maintained and 

 the processes of budding take place internally and externally 

 without perfect correlation. That is, the exterior has out- 

 growths and so has the archenteron, but these are not strictly 

 coincident and produce true buds only in forms with thin 

 mesenchyme. 



The evidence in favour of the opinion that the diverticula 

 or ampullge are strictly homologous with the archenteric 

 diverticula of all other animals is very strong. The young 

 have no diverticula until the ampullinula is formed, and this 

 correlates with the absence of these organs in the adults of 

 the lowest type, Ascones. These facts among sponges seem 

 to be in accord with the history and development of the diver- 

 ticula among Hydrozoa and Actinozoa, and lead to the con- 

 clusion that in all of these three types the diverticula are 

 homoplastic organs, and not found in the lowest forms of 

 these groups or in the early stages of development of the 

 normal forms. 



The considerations we have presented above have therefore 

 a direct application to the results of the work done of late 

 years by Semper, Dohrn, and others in tracing the origin of 

 the Vertebrata to some worm-like type. The whole of this 

 evidence hangs necessarily upon the probability that the somites 

 of the embryo of Am^Moxus imply descent from a segmented 

 animal ; whereas, if we are correct, exactly the opposite view 

 may be considered as the more probable ; and the very close 

 comparisons made by Semper between what he considers 

 homogenous organs and parts in Vertebrata and Vermes can 

 only be considered as evidence of the production of homoplastic 

 effects by means of similar modes of growth and the similar 

 habits of elongated and necessarily bilateral animals. 



We have objected to the theory that the Vertebrata may be 

 considered as descended from a (Jcelenterate ancestor, because 

 the actinostome probably arose independently and very late 



