As regards some points in the above description, Professor Allman does not agree with Huxley, but believes, on the other hand, "that the walls of the atrium simply surround the branchial sac, without being reflected on its sides, and that the branchial sac is therefore properly within the cavity of the atrium."

In structure, the "branchial" sac is composed of a series of longitudinal and transverse bars, which cross each other at right angles, and thus give rise to a series of quadrangular meshes, the margins of which are fringed with vibratile cilia. These bars are hollow, and are really vessels, which open on each side into two main longitudinal sinuses, the so-called "branchial" or "thoracic" sinuses - one of which is placed along the haemal side of the sac, whilst the other runs along its neural aspect. The function of the entire perforated sac is clearly respiratory.

The Tunicata mostly possess a distinct heart, consisting of a simple muscular tube, which is open at both ends, and is not provided with valves. The circulation is attended with the remarkable peculiarity of being periodically reversed, the blood being propelled in one direction for a certain number of contractions, and being then driven for a like period in an opposite direction; "so that the two ends of the heart are alternately arterial and venous."

The nervous system consists of a single ganglion placed on one side of the oral aperture, between it and the anus, in all known Tunicata, except in the aberrant form Appendicularia.

The only organs of sense are pigment-spots, or ocelli, usually placed between the oral tentacles, and an auditory capsule, sometimes containing an otolith. These organs, however, are not constantly present.

With the exception of Doliolum and Appendicularia, all the Tunicata are hermaphrodite. Mr Saville Kent, however, is of opinion that Appendicularia is the free reproductive zooid of an Ascidian, bearing to the adult the same relation that the Medusae do to the Hydroid colony. The reproductive organs are situated in the fold of the intestine, and their efferent duct opens into the atrium. The embryo Tunicate (fig. 201, A and C) is at first generally free, and is mostly shaped like the tadpole of a frog, swimming by means of a long caudal appendage. In one species (Molgula tubulosa) the larval form is destitute of a tail, inactive, and amoeboid, and it almost immediately attaches itself by means of little outward processes which it develops. Several other species of Molgula are also destitute of a tail when in the embryo condition; but the embryos of those species of the genus which are fixed in the adult state appear to be provided with a caudal appendage. Lastly, the larval caudal appendage has been shown to exhibit a cylindrical rodlike body, which has been paralleled with the chorda dorsalis of Vertebrates. The body in question (fig. 200, A) is a kind of cellular rod, which agrees with the notochord of Vertebrates in giving insertion by its sheath to muscles, and which is said to hold an analogous position to the nervous system. In many cases, also, the caudal appendage exhibits in addition diverging rays, which have been compared with the rays of the tails of young fishes.

It is impossible here to enter minutely into the structures which compose the larval Tunicate, the relations of these structures to one another and to the larva as a whole, or their precise homological import. Such important theoretical conclusions have, however, been based upon the interpretation of these structures, that a few words may be directed to this point.

According to the researches of Kowalewsky and Kupffer, the larval Tunicate differs from the Invertebrata generally, and agrees with the Verte-brata in the fact that the embryo is bicavitary, the nervous system being developed in a dorsal cavity, quite separate from the cavity in which the viscera generally are formed. The axial structures of the tail, as described by the same observers, are stated to commence as a double row of quadrate cells, surrounded by oval muscle-cells, and projecting slightly into the body of the larva, having the visceral canal below, and the neural canal above. When complete, the axis of the tail is said to consist of a cartilaginous elastic axial rod, surrounded by a cellular envelope. According, then, to the views of the observers just mentioned, the axial rod of the tail of the larval Tunicate is the precise equivalent of the "chorda dor-salis" of vertebrate animals, seeing that it is dorsal in position, and that it is intermediate between the neural and visceral cavities. The perforated branchial sac being, further, regarded as a development of the anterior portion of the alimentary canal, is an additional support to this view ; as it can be compared with the dilated and perforated pharynx of the Lance-let (Amphioxus) the lowest of the Fishes.

The views sketched out above, though accepted and endorsed by many high authorities, have not been allowed to pass without question. The opinions originally put forward by Kowalewsky as to the essential identity of the developmental processes of the Tunicates and the Vertebrates have been attacked, and many of the most important of his alleged facts have been denied by such well-known authorities as Mecznikow, Donitz, Giard, Von Baer, and Reichert. Without entering into this controversy further, it will be sufficient to enumerate the more important points which the researches of the observers just mentioned would seem to render more or less probable:

1. The axial rod of the larval Ascidian is ventral in position, and cannot, therefore, be homologically compared with the "chorda dorsalis" of the Vertebrate embryo. (Von Baer.)

Fig. 201.   Development of Tunicata. A, Larva of Botryllus violaceus, greatly magnified : a Processes for attachment; b Mass of primitive cells from which the digestive organs are developed; c Circlet of eight cellular outgrowths; d Eye spot; e Entrance to the branchial sac ',/ The external structureless

Fig. 201. - Development of Tunicata. A, Larva of Botryllus violaceus, greatly magnified : a Processes for attachment; b Mass of primitive cells from which the digestive organs are developed; c Circlet of eight cellular outgrowths; d Eye-spot; e Entrance to the branchial sac ',/ The external structureless "test; " g Large nucleated cells forming the sheath of the central axis (eight rows of these cells are present). B, A portion of the tail, highly magnified; h Central axis (f and g, as before). C, Another larva of the same, viewed from the side, and highly magnified, showing the superior and inferior fin-like prolongations (p p) of the "test," with raylike striae (the other letters as before). D, Diagrammatic cross-section of the tail, showing the position of the fins (p p), and the relations to one another of the central axis (h), the intermediate cellular sheath (g), and the external structureless test (f). (After Reichert.)

2. The embryo of the Tunicates is not really bicavitary, and the nervous system is not developed in a chamber separate from and lying above the visceral cavity. (Donitz, Reichert.)

3. The nerve-ganglion of the Tunicates is placed upon the ventral surface of the larva, and does not, therefore, correspond with the cerebrospinal nervous system of the Vertebrates. (Von Baer.)

4. The tail of the larval Ascidian is a purely provisional organ.

5. The axial structures of the tail (fig. 201, A, B, and D) consist of a central homogeneous, structureless, and elastic rod, surrounded by a sheath of large nucleated cells, and the former is not primitively composed of cells. (Reichert.)

6. The perforated branchial sac is not a dilated pharynx, but appears to correspond rather with the branchial chamber of the Lamellibranchiata.

From the above, therefore, the general conclusion may safely be drawn, that we are not at present in a position to give a dogmatic account as to even the most important phenomena in the development of the Tunicates. We may further conclude that the view that the development of the Tunicates is fundamentally identical with that of the Vertebrata, is as yet not to be regarded as scientifically proved.

Amongst the Salpians a species of alternation of generations has been observed. A solitary Salpian produces long chains of embryos, which remain organically connected throughout their entire life. Each individual of these associated specimens produces solitary young, which are often very unlike their parents, and these again give rise to the aggregated forms.*

The Tunicaia are often spoken of as exhibiting three main types of structure, which give origin to as many sections, known respectively as the solitary, the social, and the compound forms. In the "solitary" Tunicaries (fig. 200, C), the individuals, however produced, remain entirely distinct, or, if not so primitively, they become so. In the " social" Ascidians, the organism consists of a number of zooids, each like a solitary Ascidian, produced by gemmation and permanently connected together by a vascular canal, or "stolon," composed of a prolongation of the common tunic, through which the blood circulates. Finally, in the "compound" forms (fig. 200, B) the zooids become aggregated into a common mass, their tests being fused together, but there being no internal union. The Botrylli, which are familiar examples of the compound Tunicates, form semi-transparent masses, often of brilliant colours, attached to various submarine objects, and consisting of numerous zooids arranged in star-shaped groups. They are almost always "very small, soft, irritable, and contractile, changing their form with the slightest movement" (Stark). The atrial apertures of all the zooids of each stellate system open into a common central cloaca (fig. 200, B.)