The whole of the skeletal parts occupy a position between the epiblast and hypoblast, and are exactly homologous in this respect with the skeleton of the Alcyonidæ. By Lacaze Duthiers they are however believed to originate in the hypoblast, but from the observations of Kowalevsky there be little doubt that they arise in the connective tissue between the two embryonic layers which is probably epiblastic in origin. A peculiar larva, probably belonging to the Actinozoa, has been described by Semper[83]. It has an elongated form and is provided with a longitudinal ridge of cilia. There is a mouth at one end of the body and an anus at the opposite extremity. The mouth leads into an oesophagus, which opens freely into a stomach with six mesenteries. In the skin are numerous thread-cells. A mesotrochal worm-like larva, also provided with thread-cells, and found at the same time, was conjectured by Semper to be a younger form of this larva. [83] "Ueb. einige tropische Larven-formen." _Zeit. f. wiss. Zool._, vol. XVII. 1867. Ctenophora. The ovum of the Ctenophora is formed of an outer granular protoplasmic layer and an inner spongy mass with fatty spherules. It is enveloped in a delicate vesicle, the diameter of which is very much greater than that of the contained ovum. This vesicle appears to be filled with sea-water, in which the ovum floats. Fertilized ova may usually be easily obtained by keeping the captured adults in water from 12-24 hours. The two main authorities on the development of these forms (Kowalevsky, No. 147 and 178 and Agassiz, No. 172) are unfortunately at variance on one or two of the most fundamental points. It seems however that the embryonic layers are formed by a kind of epibolic gastrula; while the true gastric cavity, as distinct from the gastrovascular, is formed by an invagination, and deserves therefore to be regarded as a form of stomodæum. [FIG. 82. FIVE STAGES IN THE DEVELOPMENT OF IDYIA ROSEOLA. (After Agassiz.) The protoplasmic layer of the ovum is represented in black.]