Delamination is constant (with the doubtful exception of some Tubularidæ) amongst the Hydromedusæ and Siphonophora. It is perhaps in the main characteristic of the Actinozoa.
Invagination by embole takes place, so far as is known, constantly amongst the Acraspeda and frequently amongst the Actinozoa; and an epibolic invagination is characteristic of the Ctenophora.
If confidence is to be placed in the recorded observations on which this summary is founded, and there is no reason why in a general way it should not be so placed, the conclusion is inevitable that of the above modes of development the one must be primitive and the other a derivative from it, for, if this conclusion be not accepted, the absolutely inadmissible hypothesis of a double origin for the Cœlenterata would have to be adopted.
Two questions arise from these considerations:—
(1) Which is the primitive, delamination or invagination?
(2) How is the one of these to be derived from the other?
There is a great deal to be said in favour of both delamination and invagination; but it will be convenient to defer all discussion of the question to the general chapter on the formation of the layers throughout the animal kingdom.
The hypoblast cells are often filled with yolk material, and secondary modifications are thus produced in the development. The most important examples of such modifications are found in the Siphonophora and Ctenophora.
In the simplest forms amongst the Hydrozoa there is no trace of a third layer or mesoblast. The epiblast is typically formed, as was first shewn by Kleinenberg, of an epithelial layer and a subepithelial interstitial layer of cells. The cells of the former are frequently produced into muscular or nervous tails, and those of the latter give rise to the thread-cells and generative organs and in some cases to muscles[85]. In many cases, amongst all the Cœlenterate groups, and constantly amongst the Ctenophora the epiblast is simplified and reduced to a single layer. The hypoblast undergoes in most cases no such differentiation but simply forms a glandular layer lining the gastric chamber and its prolongations into the tentacles; but in the Actinozoa it appears to give rise to muscles, and strong evidence has been brought forward to shew that in some groups it gives rise to the generative organs.
Between the epiblast and hypoblast a structureless lamella appears always to be interposed.
In many Cœlenterata further differentiations of the epiblast are present. In many forms the layer gives rise to a hard external skeleton. This is most widely spread amongst the Hydrozoa, where in the majority of cases it takes the form of the horny perisarc, and in the Hydrocoralla (Millepora and Stylasteridæ) of a hard calcareous skeleton. The skeleton in these forms, though closely resembling the mesoblastic skeleton of the Actinozoa, has been shewn by Moseley ([164]) to be epiblastic.