Both Agassiz and Metschuikoff justly insist upon the correspondence in development of the lateral gastric diverticula of the Echinopaedium with the gastrovascular canal system of the Ctenophora ; and, on the ground of this resemblance, the former refers the Echinoderms to the Itadiata, retaining under that Cuvierian denomination the Acalephce (Coelenterata) and the Echinodermata. But this arrangement ignores the real value of his own discovery, which shows that the Echinoderms have made a great and remarkable step, in passing from their primarily ccelen- terate stage of organisation to their adult condition. And it further ignores the unquestionable fact, admirably brought out by the same excellent observer s recent investigations into the development of Balanoglossus, that the Echino- piedium is almost identical in structure with the young of animals, such as the Gephyrea and Enteropneusta, which are in no sense radiate, but are, eminently, bilaterally symmetri cal. In fact, the larva of Balanoglossus (the sole repre sentative of the Enteropneusta), was originally described by Miiller, under the name of Tornaria, as an Echinoderm larva; and was subsequently more fully examined by Prof. Alex. Agassiz, who also regarded it as an unquestionable Echinoderm larva ; and it is only recently that it has been proved, partly by Metschnikoff and partly by Agassiz himself, to be the larval form of Balanoglossus. In Balanoglossus, as in the Echinoderms, it appears that saccular outgrowths of the intestine give rise to the perivis- ceral cavity and its walls ; and, if such be the case, the mesoblast will be chiefly, if not wholly, represented by diverticula of the alimentary canal. Thus in the Chcetog- natha and Eckinodermata, and possibly in the Enteropneusta, the perivisceral cavity is a portion of the alimentary cavity shut off from the rest ; and, in contradistinction to the Schizoccela, in which the perivisceral cavity is produced by a splitting of the mesoblast, they may be said to be En- ierocoela.
If we endeavour to determine the place of the three remaining great groups of animals, the Mollnsca, the Tuni- cata, and the Vertebrata, obstacles arise, firstly, from a want of sufficiently exact knowledge respecting the Mol- lusca; and secondly, from the difficulty of interpreting certain well-ascertained facts in the Vertebrata.
That the Mollnsca, including under that name the Poly-oa and Brachiopoda, as well as the higher Mollusks, are closely allied to the Annelida, is readily demonstrated. The known forms of Brachiopod, Lamellibranch, Pteropod, and Gasteropod larvae all have their parallels among Annelidan larva?. The Folyzoa are closely allied with the Gephyrea and Rotifera; and a Mollusk may be said to be a few-segmented annelid with a mantle. But whether the perivisceral cavity is developed in the annelidan or in the echinoderm fashion is not yet clear. In the Polyzoa, the evidence is at present insufficient to justify any conclu sion. In the Brachiopoda, there is some ground for think ing that the perivisceral cavity is formed in the same way as in Sagitta and the Echinodermata ; while, in the Lamelli- branchiata and Odontophora, there is every reason to believe that the perivisceral cavity is formed by splitting of the mesoblast, or that they are schizocoelous.
In the lowest Tunicata, represented by Appendicularia, the recent investigations of Fol have shown, that, in the adult, the body proper is formed almost exclusively by an ectoderm and endoderm, which proceed directly from the epiblast and hypoblast of the embryo. It is only in the caudal appendage that a distinct mesoblast is represented by the notochord and the muscles. The blood channels correspond with the blastococle, and the
- house" in which these singular animals shelter themselves
is a cuticular secretion, representing the cellulose coat of the higher ascidians. The Appendicularice have no atrium, or at most only rudiments of it, hence the branchial clefta open directly on the haemal aspect of the body, which corresponds with the ventral face of a vertebrate animal. In all other Tunicata, an atrial cavity is formed by involu tion of the ectoderm, which thus gives rise to a cavity on each side of the branchial sac, into which the branchial clefts of the adult open ; and a thick cellulose cuticula, into which cells from the ectoderm usually wander, invests the exterior of the body. The "atrial tunic," or invagi- nated layer of the ectoderm, is reflected, as a visceral layer, over more or less of the outer surface of the alimentary canal, and, as a parietal layer, over more or less of the inner surface of the body wall j and the space between the two (the blastocoele) becomes converted into the blood passages. Thus, such an ascidian resembles a vertebrated animal, not only in the manner in which its nervous centre is deve loped, but in the fact that it possesses an atrial cavity, which singularly resembles the pleuroperitoneal chamber of a vertebrate. For this cavity is bounded externally by the atrial tunic and the integument, which correspond with the somatopleural layer of the mesoblast and the epiblast of a vertebrate embryo ; and it is bounded, internally, by the atrial tunic and the epithelium of the alimentary canal, which, to the same extent, correspond with the splanch- nopleure and the hypoblast. The primitively double atrial aperture has its parallel in the peritoneal openings which persist in many Vertebrata.
Thus the ascidian has no " perivisceral cavity" formed by splitting of the mesoblast, nor has it any "perivisceral cavity " formed by diverticula from the alimentary canal. It is neither enterocoelous nor schizocoelous, but what, at first sight, resembles a perivisceral cavity is formed within the body by involution, and the ascidian may therefore be said to be epiccelous. If the alate prolongations of the body which lie at the sides of the branchial apertures, in Balanoglossus, were to enlarge and unite round the anus so as to leave but a relatively small opening between their edges, the cavity so formed would answer to the atrial chamber of an ascidian.
appears to be formed by the splitting of the mesoblast into two layers, a splanchnopleure and a somatopleure, and, there fore, seems at first to correspond with the perivisceral cavity of the Annelids and Arthropods. But what is now known of the structure and development of the lowest and most embryonic of known Vertebrata, Ampkioxus, throws very great doubt upon this interpretation of the facts. One of the most singular of the many peculiarities of Ampliioxus is the fact that the branchial clefts open, not on the exterior of the body, as in all other Vertebrata, but into a chamber with a single external aperture, which, on the one hand, curiously resembles the atrium of an ascidian ; while, on the other, it is undoubtedly homologous with the pleuroperitoneal cavity of the higher Vertebrata. Now Kowalewsky s investigations have shown, that, at first, the branchial apertures of the embryo Ampkioxus open upon the exterior of the body, but that, after a time, a process of the wall of the body, on the dorsal side of the branchial apertures, grows down over them, and, uniting with its fellow in the median ventral line of the body at all points, except at the abdominal pore, gives rise to the outer wall of the pleuro peritoneal cavity. Thus the lining of that cavity, like the atrial tunic of the ascidian, is a derivative of the epiblast ; and A mph ioxus is epicoelous. As it can hardly be doubted that the somatopleure of Amphioxus is the homo-
logue of the somatopleure in the higher Vertebrata,\i becomes- ↑ It must be recollected that the pericardium is also origiually a part of this c.ivity, and that in some fishes, e.g., the Rays, it never becomes completely shut off from it.
