1884.] 115 [Hyatt. 



tions of the endoderm. This discrepancy may be explained as a 

 secondary mode of forming the coelomic sacs due to excessive 

 thinness of the mesoderm in the Actinozoa. Instead of being 

 steadily increased it becomes reduced in thickness, disappears 

 almost completely and is reinforced by fresh epithelial cells aris- 

 ing from the ectoderm and endoderm, which form the secondary 

 layers of the mesoblast or mesoderm of the Hertwigs (Jena. 

 Zeitschr., vol. xiv, p. 19.). The wall of the archenteron, there- 

 fore, grows inwards and indirectly forms the coelomata by 

 building the septal divisions, instead of growing outwards at spe- 

 cial places and forming these pouches by the direct method as in 

 the Porifera. Thus no pores appear in these young forms, so far 

 as known, and they are certainly not essential structures in the 

 Ooelenterata. The structure of the archenteron and of the larva 

 also, previous to the ingrowth of the mesenteries, and the 

 extremely thin or absent mesoderm makes a close comparison 

 possible between this form and the gastrula and Ascon stage of 

 Sycandra. These facts and the thin middle layer and absence 

 of diverticula in the adults of the acoeloinate Ascones plainly 

 indicate that we cannot assume the former existence of a porif- 

 erous ancestor with a thick middle layer or mesenchyme, but 

 must necessarily picture the common ancestor of Ooelenterata 

 and Porifera as a triploblastic gastrula with a very thin mesen- 

 chyme and no pores. The assimilative and feeding functions of 

 the endoderm in Actinozoa appear in strong contragt with the 

 transference of the first functions to the mesenchyme in Porifera, 

 but it must be recollected that this middle layer is very 

 likely equally inactive as compared with the endoderm in the 

 Ascones; and whether the next stage of the middle layer in 

 Actinozoa is formed by epithelial layers split off from the ecto- 

 derm and endoderm, or whether it is developed out of the pre- 

 viously existing mesenchyme, it is evident that this layer has 

 not the primitive importance in the Actinozoa and Echino- 

 dermata that it has in the sponges, and consequently more work 

 is thrown upon the endoderm. The conclusion seems almost 

 inevitable, therefore, that the antimera of Actinozoa and the 

 ampullae of sponges, though formed from the same layer, are not 

 homogenous but homoplastic parts arising independently in the 

 two types from an ancestor which did not possess them. 



