of Amphioxus--in which the actual condition which I had supposed to exist in the Vertebrata was shown to occur, namely, the formation of the mesoblast as paired pouches in which a narrow lumen exists, but is practically obliterated on the nipping-off of the pouch from the archenteron, after which process it opens out again as coelom" (pp. 16-18). The enterocoelic theory was taken up by O. and R. Hertwig as an essential part of their _Coelomtheorie_.[447] In a lengthy series of monographs these workers made a comparative study of the mode of formation of the middle layer, and arrived at a coherent theory of its origin. They distinguished in the middle layer two quite distinct elements, the mesoblast proper, formed by the evagination of the walls of the archenteron, and the mesenchyme, formed by free cells budded off from the germ-layers. The following passage gives a good idea of their views and of the phylogenetic implications involved:--"Ectoblast and entoblast are the two primary germ-layers which arise from the invagination of the blastula; they are always the first to be laid down, and they can be directly referred back to a simple ancestral form, the Gastraea; they form the limits of the organism towards the exterior and towards the archenteron. The parietal and visceral mesoblast, or the two middle layers, are always of later origin, and arise through evagination or plaiting of the entoblast, the remainder of which can now be distinguished as secondary entoblast from the primary. They form the walls of a new cavity, the enterocoel, which is to be regarded as a nipped-off diverticulum of the archenteron. Just as the two-layered animals can be derived from the Gastraea, so can the four-layered animals be derived from a Coelom form. Embryonic cells, which become singly detached from their epitheliar connections we consider to be something quite different from the germ-layers, and accordingly we call them by the special name of mesenchyme germs or primary cells of the mesenchyme. They may develop both in two-layered and in four-layered animals. Their function is to form between the epithelial limiting layers a secreted tissue (_Secretgewebe_) or connective tissue with scattered cells, which cells can undergo, like the epithelial elements, the most varied modifications.... This secreted tissue in its simple or in its differentiated state, with all its derivatives, we call the mesenchyme" (p. 122). The important point for us is that, just as all Metazoa were considered by Haeckel to be descended from the Gastraea, so all Coelomati were held by the Hertwigs to be derived from an original coelomate _Urform_. In both cases an embryological archetype becomes a hypothetical ancestral form.