The segmentation of the egg in the Cunanthidae has never been observed. The life histories of several members of the family, so far as known, are most remarkable. (1) Cunoctantha (Cunina) octonaria, parasitic in the bell of the Anthomedusan Turritopsis. The first noted stage (? developed from an ovum) is proboscidiform with a mouth and near the aboral pole two short tentacles. The young animal adheres by its proboscis to the margin of the bell of its host. It then migrates into the cavity of the bell and affixes itself by its tentacles, two new tentacles appearing between the two first formed, whilst the oral end of the body elongates remarkably, and may be inserted into the mouth of the Turritopsis. At the same time it developes as many as six to seven buds at its aboral pole which become larvae similar to their parent, and are detached. The ridge, from which the peripheral part of the bell is formed, appears in the same zone as the bases of the tentacles, the number of which is increased by the development of four more. The ridge grows out at its summit into intertentacular lobes, each of which bears a tentaculocyst. The larval Medusa is now set free. The oral proboscis shortens, the stomach pouches acquire their adult character, and the generative rudiments are formed.

A somewhat similar larva has been observed in the bell of Rhopa-lonema velatum. (2) Cunina proboscidea (= C. vitred). In this species, in C. rubiginosa s. rhododactyla, and apparently in C.(Cunoctantha) Kbllikeri, an asexual reproduction or sporogony occurs in both males and females, which has been accurately worked out in the first-named of the three. Immature sexual cells with a granular protoplasm wander from the sexual organs into adjacent parts. They multiply by fission. Eventually one cell ingulfs another, and the ingulfed cell divides and forms a morula. If the inclosing cell lies in the endoderm of a festoon canal further development of the inclosed cell into a ciliated lens-like body with ecto- and endoderm takes place. It escapes from the supporting cell, and is eventually converted into a Medusa. During its growth it developes buds at the aboral pole, one after another in C. proboscidea, which are detached, but may remain in connection with a stolo prolifer in C. rubiginosa. Gemmation does not appear to take place in C. Kollikeri. In the two last named the Medusa becomes a true Cunanthid though differing in the number of its segments from its parent, e. g. C. Kollikeri is eight-rayed, its young twelve.

In C. proboscidea the young sexually mature Medusa differs entirely from its parent1. (3) Cunina (Cunoctantha)parasitica. The life-history of this form is only partially known but is probably connected with a sporogony. It is parasitic in Geryonia proboscidialis s. Carmarina fungiformis2. The youngest stage observed consists of a colossal cell with pseudopodia adherent to the bell-margin of the host, and partially inclosing a morula composed of a cap of ciliated ectoderm cells and an irregular mass of entoderm cells.

1A female specimen of Geryonia hexaphylla ( = Carmarina ftingiformis, Haeckel, 'System,' p. 297) was found by Schulze to be hermaphrodite: see A. N. 41 (2), p. 404.

1See Metschnikoff, Embryologische Studien an Medusen, Wien, 1886, pp. 119, 120.

2 A sporogenetic brood seems to be capable of wandering. Metschnikoff states (op. cit. p. 122) that he has found embryoes and buds of Cunina rubiginosa (rhododactyld) in Polyxenia (Solmissus) albescens. Hence the budding larvae found in various Geryonids may belong to an identical species.

The nomenclature of the Geryonidae appears to be in much confusion. The two specific names given in the text are used by Metschnikoff for the same medusa.

The ecto- and endo-derm cells grow each in a single layer round the colossal cell, leaving a slit-liŁe or triangular opening by which it protrudes its pseudopodia. The elongate or triangular larva now swims about and settles eventually on the gastric peduncle of its host. It begins to bud, whilst the colossal cell atrophies, and is then transformed into a clavate sac, not quite \ in. long, with walls composed of an ectoderm, longitudinal ectodermal musculature, a supporting lamina, and an endoderm with muscle fibres (? circular). This sac, when full-grown, carries buds in numbers up to about Ioo, which develope into eight-rayed Cuninae, and are detached. The cycle therefore includes an Alternation of Generations, and Metschnikoff has suggested that there may be three generations, (I) a true Cunina, (2) a sac-like larva (supra), and (3) a Cunina-like Medusa with peronia but no festoon-canals.

As to the Trachomedusae, Aglaura hemistoma has an elongate larva with an axis of endoderm cells in a single row. The mouth is formed at one pole, and near it the first two tentacles, and then at right angles to them two more. In the Geryonidae mesoglaea makes its appearance as soon as the endodermal sac is complete. A patch of thickened ectoderm is then formed at one spot and the endoderm sac gradually approaches and at last comes into contact with it, becoming, sooner or later, flattened. The edge of the patch thickens, while its centre becomes thin. The primary radial solid tentacles (four in Liriope, six in Geryonid) appear in connection with the thickened edge which gives origin to the margin of the umbrella and velum, whilst the central disc becomes the ectoderm of the subumbrella and manubrium. The mouth breaks through in its centre. The interradial solid tentacles, in number equal to the radial, next appear. During growth the Medusa undergoes changes. The oral peduncle grows down, carrying the manubrium with it; the hollow radial tentacles of the adult spring from the margin of the bell, whilst the solid radial tentacles drop off; the solid interradials are lost in some genera, not in others1.

The hydroid appearance of the larva is much masked in the Geryonids by the early development of mesoglaea, in Cunoctantha parasitica by parasitism. In other instances, especially in C. octonaria, it is exceedingly well-marked, and it has a resemblance to the Actinula larva of certain Hydroidea (p. 765).