The sexes are separate, and the sexual glands consist of simple or branched sacs derived from the epidermis (?), opening by external pores, and placed in lobes of the body-walls, which are arranged in a single series to the outer side of the branchial pores, and extend behind the branchial region to a variable distance. The larva of B. Kowalewskii is a cylindrical organism with an anterior tuft and a posterior ring of long cilia. The body is covered with short cilia, and the collar is early separated off from the proboscis and trunk by two constrictions (Bateson). In other species (? all) the larva is known as Tornaria. It has cilia arranged in a prae-oral, and a longitudinal post-oral or oblique dorso-ventral band, and in one or two, posterior rings. Two eye spots lie at the anterior extremity. It resembles very closely the Bipinnaria larva of Asteroidea.

The various species of Balanoglossus differ from one another in minor details. The development of B. Kowalewskii has been accurately investigated by Bateson. It may be noted that the proboscis gland originates, like the mesoblast, from the wall of an archenteric outgrowth, and that there is neither stomo- nor procto-daeum.

Bateson has proposed to class the Enteropneusta under the name Hemi-Chordata with the Chordata. Of the points to which he draws attention, the gill-slits formed as outgrowths from the digestive tract, together with their skeleton and blood-supply, the origin of the mesoblast from enterocoelic pouches, the presence of an anterior pouch, which is cut off from the archenteron and opens externally, are undeniable resemblances to Amphioxus; and it is possible that the backward growth of the collar over the gill-slits, slight as it is, may be comparable to the epi-pleural folds of that animal. But it must be borne in mind that the anterior enterocoelic pouch is divided in Amphioxus into a left and right half, the former of which is converted into a sensory organ, opening into the oral cavity. In Balanoglossus Kiipferi, instead of one pore there are two, which lead into the interspaces of the cells filling the proboscis cavity, and this may be an original feature. The resemblance, therefore, between the structures is possibly only a general one.

As to the other points Bateson mentions, it may be noted, relative to the notochord of Balanoglossus, (I) that it is below, i.e. ventral, to the main dorsal blood-vessel, whereas in all Chordata it is above, i. e. dorsal, to it; (2) that the histological changes undergone by its cells may be simply correlated with its function as a supporting structure, and indeed Spengel states that in the species examined by him, the cells retain their cylindrical form and cilia, and that the organ is regenerated with the proboscis after amputation. In the nervous system the presence of the diffused sub-epidermic network of nerve fibres, of a peripharyngeal band and ventral cord, are points of unlikeness to which attention should be paid; and, it may be added, the canals occurring in the dorsal cord, can scarcely be paralleled with the neural canal of typical Chordata1. Nor can much weight be laid on the absence of organs of special sense, or of specialised excretory organs. They are features which are as likely as not due to the mode of life of the animal.

Tornaria requires a fresh examination. The anterior enterocoelic pouch is represented in it by a diverticulum, which opens on the dorsal surface, and is eventually shut off from the archenteron, and its cavity is said to become the cavity of the proboscis. The latter represents an overgrown praeoral or prostomial lobe, and it has at its apex two eye-spots, resting on an epiblastic thickening, which should by rights be the supra-oesophageal thickening (Balfour), but it seems to disappear without leaving a trace. Balfour regarded Tornaria as intermediate in structure between the Echinoderm larva and the Trochosphere, resembling the former in shape, in the longitudinal band of cilia, the origin and structure of the water-vascular vesicle (= anterior enterocoelic pouch), and in the formation of the body-walls from archenteric diverticula; whilst Trochospheral characters are seen in the prostomial eye-spots, the contractile band from the eye-spots to the oesophagus, the two posterior ciliated rings, and terminal anus. It may be questioned which is the more primitive form, Tornaria or Bateson's larva.

In one feature the latter appears to differ very markedly from Tornaria, viz. in the dorsal anus, which becomes terminal by the atrophy of the post-anal portion of the tail.

Bateson, Q. J. M. xxiv. 1884; ibid. xxv. Suppl. 1885; ibid. xxvi. 1886. Spengel, Mitth. Zool. Stat. Naples, v. 1884.

Tornaria, Spengel, Tagebl. d. Natf. Versaml., Miinchen, 1877; Agassiz, Mem. American Acad, of Arts and Sciences, ix. 1873 (or analysis by Perrier, A. Z. Expt. ii. 1873)

Spengel is stated to be preparing a monograph of the genus in the 'Fauna and Flora of the Gulf of Naples.' New species have been recently described by Marion, C. R. 101, 1885, and by Koehler, C. R. 102, 1886.

For relations to Chordata, see Bateson, op. cit. supra, and Id. 'The Ancestry of the Chordata,' Q. J. M. xxvi. 1886; cf. Koehler, Z. A. ix. 1886.

1An invagination in the development of the ventral cord possibly takes place in some Chae-topoda, and it is difficult to assign any other origin to the canal which traverses the cord in the Cephyrea chaetifera. It is doubtful whether such an invagination per se can be held to have a phylogenetic significance. The point can only be determined by future investigation.