HÏdrozoa

HÏDROZOA

THE Hydrozoa form one of the three classes into which the Coelentera nematophora (distinguished from the Coelentera porifera, or Sponges) have been divided,— recognized as such in the article Coelentera, to which the reader is referred. It results from observations made by Ernst Haeckel, since that article and the article Actinozoa were penned, that the Cteηophora should not be regarded as a class equivalent to the Hydrozoa and Actinozoa, nor as a subdivision of the latter class, but that they must be considered as a peculiar modification of the medusiform Hydrozoa (see final paragraph). If this conclusion be accepted, it will be necessary to divide the Hydrozoa into two primary groups or grades, for which the names Polypo-morpha and Cteηophora are proposed.

Ι⅛e Hydrozoa correspond to the Linnaean genera Hydra, Tubularia, Sertularia, and Medusa. The name was applied by Huxley in 1856 to a group corresponding to that termed Hydromedusoe by Vogt (1851) and Medusoe by Leuckart (1853), and embracing the forms placed by Gegenbaur in his Elements of Comparative Anatomy (1878) in four classes, viz., Hydromedusoe, Calycozoa, Thecomedusoe, and Medusoe. Our knowledge of the structure and life-history of the Hydrozoa, many of which, on account of their delicacy and oceanic habits, are excessively difficult to obtain in a state fit for investigation, has greatly extended within the last five years. Whilst in the two decades preceding this period the admirable researches of Huxley, Gegenbaur, Agassiz, and Allman had brought to light and systematized a vast mass of information with regard to these organisms, the later observations of Claus, the Hertwigs, Haeckel, and Metscħnikoff, have corrected, extended, and added to their history, especially in respect of embryological and histological detail. An epitome of the present condition of our knowledge of the group is afforded by the subjoined tabular classification of its families, orders, and sub-classes.

The definition and synonymy of the divisions recognized will be entered into, after a sketch has been given of the common structural features of typical Hydrozoa. [table]

[table]

The Hydrozoa present a greater simplicity of ultimate structure than do any animal organisms possessed of as great a complexity of external form. As in all Metazoa or Enterozoa, the life cycle of a hydrozoon starts with an egg which is at first a single cell or unit of protoplasm, but proceeds after fertilization to multiply by transverse fission in such a way that the resulting cells or units are arranged in two layers, each one cell deep, disposed around a central cavity—the enteron or archenteron. The sac thus formed is known as a diblastula (figs. 1, 2, and 25). By the formation of a mouth to the sac, the enteron acquires the functions of a digestive retort in which food matters taken in at the mouth are brought into a chemical condition suitable for the nutrition of the surrounding cells. The two layers of cells (of which the outer only acquires additional layers by the division of the primary cells, and that by no means in all cases) received from Allman (Phü. Trans., 1855) the names respectively of the ectoderm and the endoderm, having previously been shown by Huxley (1849) to be the fundamental membranous constituents of which the most varied parts of the more complex Hydrozoa— such as tentacles, swimming bells, and air-bladders— are built up in the adult condition. Huxley also pointed out the identity of these membranes with the two primary layers of the vertebrate embryo. The endoderm and the ectoderm, which present themselves, as is now known, in the diblastula (or gastrula) phase of all Enterozoa, remain in Hydrozoa (and also in the allied groups of Coeleηtera) as permanently distinguishable elements of structure. This important disposition is associated with and dependent on the simple character which the archenteron or primitive digestive space retains. Into whatever lobes or processes the sac-like body may be, so to speak, moulded, whether tentacles or broader expansions, into these the cavity of the archenteron is extended in the first instance; and where the actual cavity is obliterated the endodermic cell-layer remains to represent it (Gefässplatte or endoderm-lamella, see figs. 7 and 16).

Conversely, whatever canals or spaces are discovered in the substance of a hydrozoon (excepting only the cavity of ectodermal otocysts) are simple and direct continuations of the one original enteric cavity of the diblastula, and all such spaces are permanently in free communication with one another.

The whole of the Hydrozoa seem to present a lower grade of structure than the Actinozoa, in so far as the latter, whilst retaining permanently free communication between all parts of the archenteric space, yet exhibit a differentiation of this space into an axial and a periaxial portion—a digestive tube and a body cavity. The differentiation has only to proceed a step further, namely, to the closure or shutting off of the axial from the periaxial portion of the archenteric space, and we obtain the condition which characterizes the adult forms of the Coelomata, or animals with blood-lymph space distinct from digestive canal. With the attainment of the coelomate condition, the two fundamental cell-layers, ectoderm and endoderm, which still appear in the embryo, become so far interwoven, and their products so highly differentiated, that it is no longer possible to recognize them as anatomical structures in the adult.

The only deep-seated distinction between Hydrozoa and Anthozoa (the Actinozoa being thus termed when the Cteηophora are detached from them) appears to be the particular differentiation of the archenteric space va Anthozoa which has just been noted. It is no longer possible to •separate the two groups from one another as Exoarii and Endoarii, as was proposed by Rapp (Ueber die Polypen im Allgemeinen und die Actinieη insbesondere, Weimar, 1&29) —the first term indicating the Hydrozoa as possessed of •external generative organs, whilst by the latter term the Anthozoa are pointed to as having internal generative organs. This distinction breaks down completely in the case of Lucernaria, and even in that of the so-called phanero-carpous and some other medusae which discharge their genital products by the mouth, and quite rarely by rupture of the outer body-walk The tendency to form calcareous deposits in the deep layers of the ectoderm, or mesoderm, as it has been termed, exhibited almost universally by the Anthozoa (whence the name Coralligena applied to them), is distinctive of them, though it has been shown first by Louis Agassiz, and more fully and recently by Moseley, to be paralleled among Hydrozoa, by the external calcareous deposits of the abundant and widely distributed Millepores and Stylasterids. A minute distinction between Hydrozoa and Anthozoa, which does not, however, hold good universally, is found in the form of the barbed threads ejected by the nematocysts. Instead of the complicated forms present in the latter group, the Hydrozoa are usually provided with either an unbarbed thread or one in which the barbs are confined to three at the base and a few minute barblets (fig. 5).

Fundamental Forms of the Hydrozoa.— The diblastula derived from the egg of a hydrozoon, when provided with •a mouth, may be spoken of (as are the equivalent forms in other animals groups) as a person. Either this person elongates and develops tentacles in a circlet around or near the mouth, and usually becomes fixed by the aboral pole of the sac-like body, or the sac gradually assumes the form of a clapper-bell or of an umbrella with greatly thickened handle, the mouth being placed at the free end of the handle or of the clapper, and the animal freely swimming by the contractions and expansions of the dome of the bell (disc of the umbrella). The two forms of persons are known, —the former as the “hydriform” (2, 3 in fig. 16), the latter as the “medusiform” (4, 5, 6 in fig. 16).

The HYDRIFORM persons usually occur as fixed branching •colonies or trees (figs. 36 and 37) produced by lateral budding from an original hydra-form developed from a diblastula.

The hydriform person in its most fully developed state is seen in the colonies of Tubularia. In such a colony a number of hydriform persons are united like the flowers of ⅛ plant on its branches (whence Allman’s terms hydranth, hydrophyton). Each hydriform person (fig. 35) has an elongated body with oral and aboral pole. The mouth is placed centrally at the oral pole, which is somewhat enlarged and conical At the apex of the cone, immediately around the mouth, is a circlet of small tentacles ; at the base of the cone is a second circlet of larger tentacles', the surface -of the oral cone is termed the hypostome. In other genera {e.g., Hydra, fig. 42) the smaller circle of tentacles is wanting; in others, again, the tentacles are irregularly placed and not concentrated into one circlet (fig. 38). We regard the former as the typical condition. In the hydriform persons of the Scyphomedusoe (figs. 26 and 27) the vertical axis is much shortened, the hypostome is flat, and the whole body cup-like or hemispherical.

The tentacles of the hydriform person are sometimes hollow (Hydra, Gaτveia nutans, Hydrocorallinoe), being mere prolongations of the sac-like body; but usually, though the endodermal cell-layer is continued into them, they are solid (2 in fig. 16). Very generally the tentacles of the hydra-form are indefinite in number, but in those belonging to the group of Scyphomedusoe a primary series indicating four radii (perradial) can be distinguished, to which are added four intermediate to these, marking four secondary radii (interradial), whilst eight more placed between the eight of the perradial and interradial series are known as adradial tentacles. The surface of the hydraform may be entirely naked, or encased in a horny tube (perisarc) formed by the ectoderm: this may be confined to the aboral portion of the hydranth and to the common stem which unites the persons of a colony, or it may rise up and form a cup (or hydrotheca) around the oral region of the hydranth (figs. 32 and 33).

The bodies of all hydriform persons, as well as the tentacles, are excessively contractile, and when hydrothecae are present can be withdrawn into them.

The ectoderm or outer cell-layer furnishes the protective and contractile tissues of the hydra-form. Very usually it is not more than one or two cells deep, and is separated from the endoderm by a structureless lamella of firm consistence. In. Hydra large cells of the ectoderm (neuro-muscular cells of Kleinenberg) bound the external surface (fig. 3) and give off horizontal muscular processes which lie side by side on the structureless lamella— forming thus a deep muscular coat, the fibrous elements of which are not independent cells. In larger species some of the fibres may become separated from the tegumentary or superficial cells, and acquire the character of independent nucleated corpuscles (Hydractinia, Van Beneden). No nervous elements nor sense-organs occur in any hydra-form (except perhaps the Lucernamoe). In Antennularia some ectoderm cells are amoebiform, and project processes which change shape (nematophυrs). Tactile hairs (palpocils), however, occur on the ectodermal cells, and the solid tentacles are essentially tactile organs. Placed in and between the large cells of the ectoderm (Hydra, Cordylophora, Allman, Kleinenberg, F. E. Schulze) are small nucleated cells which become converted into vesicles containing a three-barbed (figs. 4 and 5) or simple filament (nematocysts). These are frequently grouped on the surface in wart-like processes or “batteries.” Nematocysts also are found in the endoderm; but it is probable that their presence there is due to their having been swallowed.

The endoderm is usually but one cell deep, and lines the entire cavity of the body starting from the margin of the mouth. In the region of the body proper, and in hollow tentacles, the cells are ciliated (fig. 4). In this region they are concerned in the secretion of digestive fluids and in absorption, and sometimes contain coloured granules (hepatic?). Allman found in Myriothela (Phil. Trans., 1875) that the endoderm cells project processes like the pseudopodia of Protozoa, and suggests that solid food particles are incepted by them. T. J. Parker has published similar observation on Hydra (1880). In the solid tentacles the endodermal cells are greatly modified, forming a kind of skeletal tissue, each cell recalling by its vacuolation and firm cell-wall the characters of vegetable parenchyma (fig. 6). In the stems of Siphoηophora endoderm cells give origin to muscular processes like those of the ectoderm (Claus). This latter fact has a morphological significance which cannot be too gravely estimated.

Generative products are not developed by any hydriform persons (excepting the Lucernaria?) , the sexual process being carried on by a distinct set of buds developed on the sides of hydriform persons. These buds either become medusiform persons, or are degenerated representatives of such persons (sporosacs) (figs. 17 and 18). Even the fresh-water Hydra (fig. 42) does not appear to be an exception to this generalization. The single egg cell of Hydra projects at the breeding season in an ectodermal covering, as a wart, from the lower part of the body. A conical eminence or two nearer the mouth contains the spermatozoa. Each ovarium and each spermarium represents an aborted generative person. According to Kleinenberg the egg-cell and the sperm-cells are both derived from the ectoderm. The Lucernarioe develop internal generative organs (fig. 19) which correspond closely with those of the medusiform persons of the group Scyphomedusoe (see below), with which they are classified. Both ova and testis are endodermal in origin in Lucernaria and in the medusiform persons of the Scyphomedusoe, whilst they appear to be ectodermal in origin in the complete medusiform persons of Hydro- medusoe, though in the degenerate medusiform persons known as sporosacs they may either or both have an endodermal origin.

Medusiform persons usually present themselves as isolated free-swimming individuals, but like hydriform persons they have the power of producing new persons by budding (figs. 44, 45, and 46), which may become detached or may remain connected with the primary person (fig. 57) to form a freely swimming colony (Siphonophora) comparable to the fixed colonies of hydriform persons. Medusiform persons are often produced as the immediate result of the development of the diblastula without any intermediate hydriform phase (Pelagia among Scyphomedusoe, Tracho- medusoe, Narcomedusoe, and probably some A nthomedusoe and Leptomedusoe), but quite as frequently originate as lateral buds upon the body-walls of hydriform persons (figs. 34, 37, and 43), or of other medusiform persons (see below), or as metameric fission-products of hydra-forms. The typical medusa-form is ,a hemispherical cup (the nectocalyx, or umbrella, or disc), from the centre of which rises up a cylindrical or conical process (the manubrium, erroneously ρolypite) at the summit of which is the mouth (4, 5 in fig. 16). Four perradial (see above for use of this term) tentacle-like lobes very commonly surround the mouth, or numerous small tentacles (fig. 58), whilst the margin of the disc is beset with tentacles four in number, or a multiple of four (sometimes six, or one only, or indefinite). The aboral pole is dome-like, and is never attached except in those forms which take their origin as buds on a hydriform colony when the connexion exists at this point. The tentacles are, as in the hydriform persons, some solid, some hollow; both occur in the same individual.

The body is not so completely hollowed out as in the hydriform persons. The mouth leads into a straight tube (the stomach) which occupies the axis of the manubrium, and expands at its insertion into the disc. The disc, even when thick and fleshy, is not fully excavated by the enteric cavity. In young forms the cavity does occupy it right up to the margin, but gradually the lumen disappears (fig. 29), leaving a series of canals and a continuous plate of endoderm (fig. 7) formed by the coalesced walls of the space (the endoderm-lamella of the Hertwigs, see Organismus der Medusen, 1878; the vascular-lamella of Claus, “Polypen und Quallen der Adria,” Wiener Denksch., 1878). The peripheral portion of the lumen of the original enteric cavity forms the ring-canal, which runs all round the margin of the disc, and is continued into the hollow tentacles. The lumen is further retained at intervals in the form of radiating canals connecting the axial enteric cavity with the ringcanal. These may be perradial, interradial, and adradial (see above as to tentacles of hydra-form), and may branch dichotomously in the disc or form networks.

The medusae are thicker and more fleshy to the touch than are the hydra-forms, and are at the same time transparent. This is entirely due to the enormous development of a structureless substance between ectoderm and endoderm,, corresponding to the “Stutz-lamella” or structureless lamella of the hydra-forms. (See figs. 49 and 51, representing sections of Carmarina and of Cunina.) The remarkable development of this substance in a hyaline condition has led to the description of canals and spaces where none exist—the supposed spaces being really occupied by this hyaline substance. F. E. Schulze’s statements as to extra-enteric spaces in Sarsia are thus explained—and more decidedly the supposed circular and longitudinal canals attributed by some authors to the scyphistoma phase of Discomedusoe. In the same manner (according to Claus) Allman’s observations on Siephanoscyphus are reconciled with those of F. E. Schulze on Spongicola— clearly the same form. Siephanoscyphus is devoid of either circular or longitudinal canals, and though it has four remarkable ridges on the enteric wall like those of the scyphistoma of Scyphomedusce (see fig. 26) stands in all probability very close indeed to the Tubularian genus, Berigonirnus.

In a large number of medusa-forms the hyaline gelatinous substance is structureless, but in many of the larger Scy- phomedusoe it is occupied by in-wanderingamoeboid cells derived from the endoderm and by fibrous trabeculae (fig. 8).

The wandering endodermal cells are nutrient in function, and represent so far isolated elements of the enteric canal system.

The medusiform person is fundamentally adapted to swimming movements. The muscular fibres are mostly transversely striated, and are as a rule outgrowths of super ficial ectoderm cells as in Hydra (fig. 9), (though in some cases distinct cells); they are confined to a sheet spread on the oral face only of the disc or swimming-bell (sometimes called sub-umbrella), to the extensile manubrium and tentacles, and to an inwardly directed flap of the margin of the disc known as the velum ( Ve in 4 of fig. 16), which is present in those medusae that are not flattened but conical (bell-like). The muscular fibres on the oral face of the disc and on the velum have a circular direction, interrupted in some cases by radial tracts. The direction of the swimming movements is obvious from this arrangement.

. The velum is not a constant element in the medusa’s disc; it serves to contract the space by which water is expelled from beneath the bell in the act of swimming.

All fully-developed Hydromedusoe possess the velum, but only a few of the Scyphomedusa? (Charybdcea), In the former the endoderm plate (vascular lamella) is not continued into it; in the latter vessels of the enteric system are present in it (fig. 21), and, being probably morphologically distinct, it has been here termed the “pseudo-velum.”

Unlike the hydra-forms, the medusa-forms of Hydrozoa possess in addition to the tentacles highly-developed senseorgans and ganglionic nerve-centres and nerves. The senseorgans appear to be either eye-spots, or else otocysts, or to combine the functions of both. In addition to these are olfactory tracts or pits connected with the preceding. The sense-organs are placed along the margin of the disc (hence called marginal bodies), and are of three kinds:— (1) ocelli—rounded pigment spots, rarely provided with a lens {Lizzia) (fig. 10), always placed at the base of a tentacle or in the radius of one on the oral surface {Lizzia), entirely ectodermal in origin; (2) vesiculi or otocysts—formed (as discovered by the Hertwigs, 1878) by an invagination of the ectoderm (fig. 11) containing concretions and hair cells; either open or entirely closed, generally numerous, and placed between tentacles, sometimes at the bases of tentacles ∖ θbelia) ; (3) tentaculocysts—which are reduced and modified tentacles; into them alone of the three kinds of mar ginal bodies do the endoderm and, in the more complex, the enteric canal system enter (figs. 12, 13, and 30). The endodermal sac forms the axis of the tentaculocyst, its cells secrete crystalline concretions, and it functions as an otocyst; pigment spots, which may have cornea, lens, and retina well developed, are formed sometimes to the number of six (Charybdoea) on the ectoderm of the tentaculocyst (fig. 13). The olfactory sense-epithelium (fig. 14) is either distributed in a continuous band on the margin of the disc (Hydromedusoe, discovered here by the Hertwigs), or it is confined to deep pits (foveae nervosae) from each of which a tentaculocyst arises (discovered in the Scyphomedusae independently by Schäfer and Claus). With some exceptions, medusae provided with ocelli are destitute of vesiculi, which alone occur in the vesiculate Leptomedusoe. Tentaculocysts characterize to the exclusion of the ocelli and vesiculi the Trachomedusae and Narcomedusae among Hydromedusoe and all the Scyphomedusae, except Lucernaria, where they are replaced by “colleto-cystophors.”

The nervous system has only recently been correctly recognized in medusae, though seen by Agassiz as long ago as 1849, and described both by Fritz Müller and Haeckel in certain forms {Geryonidoe) more recently (1860). It differs remarkably in the two great groups into which the Hydrozoa are divisible. In the Scyphomedusae there is no continuous nerve-centre, but around and about each tentaculocyst nerve-fibres and cells are grouped in such a way as to divide the disc into zones of nerve supply corre- ≡ponding to the number of tentaculocysts (usually eight).

Both the Hertwigs {Nerven- Sy stem der Medusen, 1878) and Ei.ner {Die Medusen, 1879) entirely missed in their researches the large nerve-fibres and prominent ganglion cells (fig. 15) which were discovered by Professor Schäfer of University College, London {Phil. Trans., 1879), in the Scyphomedusae. The writer can confirm Schafer’s observation of the existence of such fibres and ganglion cells in the region of the circular muscular zone on the oral face of the disc of Aurelia, immediately beneath the flattened epithelium of the ectoderm. Professor Claus of Vienna has independently described { u Polypen und Quallen der Adria,” 187,8) similar nerve-cells and fibres in Chry-sαora and Charybdaea. Professor Schäfer failed to ascertain satisfactorily the origin and termination of the fibres, which appear, however, to originate in superficial ecto dermal cells (“sense-epithelium”) in the neighbourhood of the tentaculocysts and in the cells of those organs, and to terminate without any plexiform connexion with one another in the muscular fibres. Eimer has described very abundant and excessively fine fibres, often moniliform, which extend from epithelial cells in the neighbourhood of tentaculocysts and form a network traversing the gelatinous substance of the disc in every direction. This observation, though supported by the fact that such fibres are indicated by the extended experimental investigation of Eimer and of Romanes (Eimer, Die Medusen·, Romanes, Phil. Trans., 1876, et seq. ∖ is not confirmed by other observers, and the fibres described are regarded as skeletal tissue. If Eimer’s fibres do not exist, the muscular tissue of the medusae must be regarded as acting to a large extent independently of nerve-control; and this is borne out by Claus’s observation of the absence of sense-organs and nerve-fibres from the swimming-bells of the Siphoηophora (compound medusae). In the Hydromedusoe the nerve ganglion cells are grouped in a continuous ring around the margin of the disc, separated horizontally into an inferior and superior portion by the insertion of the velum. The difference in the form of the nervous system has led Eimer to propose the names Cycloneura for the Hydromedusoe and Toponeura for the Scyphomedusae. Amongst the latter, however, Charybdaea, having a continuous velum like Hydromedusoe, has also a continuous nerve-ring.

Comparison and Relations of Hydriform and Medusiform Persons.— A simple shortening of the vertical axis, and a widening of the hypostome, with obliteration of the lumen (but not of the cells) of the endoderm over a considerable region of the disc thus produced, suffice to convert the hydraform into the medusa-form. This change of proportion made (fig. 16), the sense-organs of the medusiform person have to be added, and the change is complete. Thus it becomes clear that we have to deal with one fundamental form, appearing in a lower, fixed, nutritive phase and a higher, locomotor, generative phase in the two cases respectively.

The phylogeny of the Hydrozoa and the historical relationship of the two phases (hydriform and medusiform) appears to be as follows.

A two-cell-layered sac-like form, with mouth and with or without tentacles, was the common ancestor of Hydrozoa, Anthozoa, and Sponges. The particular form which the proximate ancestor of the Hydrozoa took (1 in fig. 16) is most nearly exhibited at the present day in Lucernaria and in the scyphistoma larva (hydra-tuba) of Discomedusae. It was a hemispherical cup-like polyp with tentacles in multiples of four, with four lobes to the wide enteric chamber. This polyp, after passing a portion of its life fixed by the aboral pole, loosened itself and swam fιeely by the contractions of the circular muscular fibres of its hypostome (sub-umbrella), and developed its ovaria and spermaria on the inner walls of the enteric chamber. This ancestor possessed, like its descendants, a very marked power of multiplication, either by buds or by detached fragments of its body. Accordingly it acquired definitely the character of multiplying by bud-formation during the earlier period of its life; each of the buds so formed completed in the course of time its growth into a free swimming person. We must suppose that the peculiarities of the two phases of development became more and more distinctly developed, the earlier budding phase exhibiting a more elongated form and simple enteric cavity (hydra-form), which subsequently became changed in the course of the ontogeny (development of the individual) into the umbrella or disc-like form, with coalesced enteric walls and radial and circular surviving spaces (medusa-form). And now the ancestry took two distinct lines, which have given rise respectively to the two great groups into which the Hydrozoa are divisible—the Scyphomedusoe and the Hydromedusoe. In the one set the hydriform persons of a colony, instead of each becoming metamorphosed into a medusiform person, proceeded each to break up into a series of transverse divisions; each division became a medusiform person, and was liberated in its turn as a free swimming organism (figs. 26 and 27). We must suppose that this process began historically by the outgrowth of new tentacles around the point where the disc of a person fully transformed from the hydriform to the medusiform phase was loosened in its attachment and about to separate from the colony. The “hastening of events,” a well-known feature of organic growth-sequences, would complete the development of the newly sprouting person before the loosened medusa had got well away, and so on with a third, fourth, and even with twenty such successive buds. The separation of the adult form from its fixed larva by fission has been justly compared by Louis Agassiz to the separation of the Comatula from its pentacrinoid larval stalk. If the stalk could only produce new Comatuloe, the analogy would be complete. Lucernaria is in the same way comparable with the stalked crinoids, being an adult form which retains the characters exhibited by the immature phases of its congeners.

The Scyphomedusoe do not, however, all exhibit a hydriform phase, and a production of medusae by the

“strobilation” or “metamerizing” of a scyphistoma. Some of them {Pelagia) “hasten events” so far that the diblastula never fixes itself, but becomes at once a single medusa, the hydriform phase of the ontogeny being altogether omitted. Certain peculiarities of the medusa’s structure, above all the possession of gastral filaments (solid filaments like tentacles projecting in four interradial groups near the genitalia into the enteric cavity), serve to unite Pelagia, which has no larval stage, and Lucernaria (which is always of intermediate character between hydra-form and medusa-form) with the numerous species which develop by the strobilation of hydriform larvae.

The second line of descent which has given rise to those Hydrozoa known as Hydromedusoe not only acquired at the start a different mode of producing medusiform persons, but the medusiform persons acquired characters differing from those of the Scyphomedusoe in important (but not fundamental) features. The larval stage in this series developed the property of budding to a very great degree, so as often to form fixed tree-like colonies of considerable size. Then the transformation of the identical colonyforming persons into free-swimming persons was finally and definitively abandoned, and only a late-appearing set of buds proceeded to complete the typical changes and to become medusae. The earlier-produced buds were thus arrested in development, and became specially modified for the purposes of a fixed life as members of a colony. Thus they acquired the elongate form and the sporadic position of the tentacles which we see in some hydriform persons of the Hydromedusoe group (figs. 38 and 40), and were adapted to nutrition solely (hence the term trophosome applied by Allman to such colonies). The characters of the mature generative person, with its power of detachment and free locomotion, being confined to the later buds borne on the sides of the hydriform persons or on special portions of the colony, we find that the former became more and more specialized as sexual medusiform persons in proportion as the latter became specialized as asexual hydriform persons, and thus it is that we have the remarkable phenomenon of hydriform colonies, developed from the eggs of medusae, producing as it were crops of medusae (figs. 34 and 37) which detach themselves and swim away to deposit their eggs (alternation of generations). The Hydromedusoe never produce medusae by strobilation or transverse division of a hydriform person, although in rare cases the cicatrix left by a detached medusa-bud has been observed to sprout and produce a hydriform person. Neither medusiform nor hydriform persons of the Hydromedusoe series ever have gastral filaments (unless they are represented by the “villi” of the Siphonophora described by Huxley, Oceanic Hydrozoa), whilst the medusa-forms always possess a velum and a comparatively simple set (four, six, or eight) of radiating canals in the disc, the remains of the enteric lumen.

The complete differentiation of hydriform and medusiform persons existing on one and the same colony having been attained in the Hydromedusoe, further changes of a most remarkable character were brought about in some of the descendants of these forms. The condition which we have so far noted is perpetuated at the present day in Bougainvillia (Budendrium), Campanularia, and a vast number of the so-called hydroid polyps; others have undergone further adaptational changes. We have to notice at least four important additional modifications independent of one another.

(1.) r Biιe hydriform stage was suppressed altogether, and, as in some Scyphomedusoe, so here too the diblastula developed directly into a medusa {Trachomedusoe, Narco- medusoe, and probably some Leptomedusoe like Thaumantias and Aeguorea, and some Anthomedusoe like Oceania and Turritopsis).

(2.) The medusiform persons being early produced did not separate themselves from the colony, but the whole colony became free (if it ever were fixed), the medusiform persons carrying the hydriform persons away with them. Thus the highly differentiated swimming and floating colonies of the Siphonophora originated.

(3.) The medusiform persons ceased to detach themselves from the fixed hydriform persons or colonies, and developed the ova and sperm within themselves, whilst still small in size and attached to the hydriform stock. Having once abandoned the detached, free-swimming life, the medusae underwent in different genera a varying amount of degeneration and atrophy, of which we have in existence all possible degrees, leading from the fixed “phanerocodonic gonophors” (Allman, bell-like genital buds) of many Siphonophora through the “adelocodonic gonophors” (genital buds with the bell no longer open but closed by the union of the margins of the disc) of Cordylophora to the sporosacs of Hydractinia, and even to the simple genital warts of the little degenerate Hydra viridis of fresh waters (see fig. 17, and explanation). By this process a large num ber of Hydromedusas (figs. 35, 38, 39, 40, and 42) have lost all evidence of the real characters of their medusa-forms, just as others have suppressed the evidence of their hydra-forms by direct development from the egg; and inasmuch as both these processes take place in genera having the closest affinity with genera in which both hydra-form and medusa-form are fully preserved, it is not possible to erect groups similar to the Haplomorpha of Carus or the Monopsea of Allman for their reception. The difficulty of classification is, however, rendered very great, fora double system becomes necessary, which shall deal with the characters of hydriform and medusiform persons in parallel equivalent series. The difficulty is considerably enhanced when we find that identical medusa-forms may spring from unlike hydra-forms, and, conversely, that closely allied hydra-forms may give rise to very different medusa-forms. The character first noticed by Rapp as distinguishing the hydroid polyps from the coral-polyps, namely, that of developing their genitalia as external bodies (Exoarii) instead of internally (Endoarii), is seen by the considerations just adduced to be fallacious. The Hydromedusae, it is true, often (not always) develop their generative products from the ectoderm, and the genitalia frequently project as ridges and discharge themselves directly to the exterior in this division. The Hydromedusas contrast in this respect with the Scyphomedusa? and Anthozoa, which develop their genitalia from the endoderm, and are (to use Rapp’s terms) Endoarii whilst the former are ExoarU. But the bodies mistaken for external generative organs by Rapp and other early observers in many hydroids, and in Hydra itself, are aborted degenerate medusae.

(4.) A further set of changes, which have affected the original hydriform colonies and their medusa-buds so as to produce new complications of structure among the Hydromedusas, are summed up under the head of “polymorphism.” The differentiation of hydriform and medusiform persons is a case of dimorphism; a further distribution of functions, with corresponding modification of form, gives us “polymorphism.” Polymorphism is unknown in the Scyphomedusa?, and it is chiefly confined to two groups of Hydromedusae (the Hydrocorallinas and the Siphonophora). In the hydriform colonies of Hydractinia (one of the Gymηoblastea-Aηthome- dusoe) the outer hydriform persons of the colony (fig. 39) differ in form from the rest, and have wart-like tentacles. In the same genus, and also in many Calyptoblastea, the hydriform persons which are destined especially to give origin to medusa-buds are devoid of tentacles and mouth, and are known as blastostyles (Allman), (fig. 43). In Hydro- corallinoe (fig. 53) elongated hydriform persons (dactylozooids) with no mouth and sporadic tentacles are set in series around a central short mouth-bearing person (gastrozooids) forming the “cyclo-systems” of Mr Moseley (figs. 52 and 55). In the Siphonophora, in addition to nutritive (hydriform) persons and generative (medusiform) persons, there may be rows of swimming-bells (medusae devoid of mouth and of genitalia), covering-pieces (flattened medusae), and tentacle-bearers (hydriforan persons with one long highly-developed tentacle), (see figs. 56 and 57).

Hypothesis of the Individuation of Organs.— The building up of complex individualities, such as a hydrozoon colony, a flowering plant, or a segmented worm or arthropod—in any one of which a number of common units are repeated, but with varied form and function in each part of the compound body—is generally admitted to be explicable in two ways, and which of the two explanations may be adopted in any one case must depend on the ultimate inference from a wide series of observations. The first hypothesis, which undoubtedly applies to the ordinary hydriform colonies of Hydrozoa, to the segments of Tcenia, and to plants formed by the repetition of phyllomes, is that an original unit like those which constitute the composite organism has freely budded, and repeated its own structure in the well-marked units which remain conjoined to form an aborescent or linear aggregate. This is “eumerogenesis,” and such aggregates may be termed eumeristic. By a division of labour and consequent modification of form among the units of a eumeristic aggregate, such an aggregate may (in the course of phylogeny) acquire varied shape and definite grouping of its constituent units, and a high specialization as an individual. The high degree of individuation which may be thus attained is due to the more or less complete synthesis of a eumeristic colony. The more highly individuated Chaetopods and Arthropods are synthesized linear colonies. The cyclo-systems of the Hydro· corallinae are undoubted examples of synthesized colonies. The second hypothesis is one which is applicable to cases which, in the absence of special evidence to the contrary, might be regarded as highly synthesized colonies. According to this second hypothesis, such highly individuated composite organisms have not (in their phylogeny) passed through a eumeristic phase in which the units were well developed and alike, but the tendency to bud-formation (whether lateral, linear, or radial) has all along acted concurrently with a powerful synthetic tendency, so that new units have from the first made but a gradual and disguised appearance. This is “dysmerogenesis,” and such aggregates as exhibit it may be called dysmeristic. In dysmeristic forms the individuality of the primary unit dominates from the first, and the merogenesis (segmentation or bud-formation) can only show itself by partially here and more completely there compelling (as it were) the organs or regions of the body of the primary unit to assume the form of new units. The arms of star-fishes are, when we consider them as derived from the antimera of a Holothurian, explained as examples of dysmerogenesis. So, too, the series of segments constituting a leech, and probably also the segments of a vertebrate. Eumerogenesis and dysmerogenesis are only variations of one process, merogenesis, and no sharp line can be drawn between them. Individuation may appear at any period in the phylogeny of a eumeristic aggregate and synthesize its units. On the other hand, individuation is more or less completely dominant throughout the history of a dysmeristic aggregate, and is gradually broken down as a more and more complete analysis of the primary unit into new units is effected. It will be observed, however, that in dysmerogenesis, the form which individuation tends to preserve is that of the primary unit (notably the case in leeches as compared with the ameristic flukes), whereas when we have eumerogenesis followed by synthesis the resulting form-individuality is something absolutely new. Thus, using the terms eumeroraorph and dysmeromorph, we have—(1) synthesized eumeromorph simulates normal dysmeromorph; (2) analysized dysmeromorph simulates normal eumeromorph.

Whether the fixed hydriform colonies of the Hydrozoa, with their more or less complete medusiform buds, and further, the floating colonies of Siphonophora, with their polymorphous units, are to be regarded as synthesized eumeromorphs or as dysmeromorphs, more or less analysed, is perhaps still open to discussion. The former view (that adopted here) is that held by Allman (Monograph of the Tubulariaη Hydroids, 1874), by Leuckart (1851), by Gegenbaur (Grundriss, 1874), by Claus (Grundzüge der Zoologie, 1876), and by the Hertwigs (Organismus der Medusen, 1878). On the other hand, Huxley (Oceanic Hydrozoa, 1856), formerly Gegenbaur (Zur Lehre der Generationswechsel, 1854), and, more recently, Ed. Van Beneden (“De la distinction originelle du testicule et de l’ovaire,” Bull. Acad. Boy. Belg., 1874) have held that the medusiform person is a generative wart which has gradually assumed the characters of a bud, and that the various phases presented by it in different genera are so many more or less successful strivings after complete assumption of the hydra-form (from which the medusa-form is thus secondarily derived). Similarly the variously modified units of the siphonophorous colony have been regarded as the organs of a parent unit which have each more or less completely acquired the form of that parent unit, or, in other words, the colonies in question have been held to be dysmeromorphs. Recently ascertained facts as to the polymorphism of HydrocoraUina>, but more especially the demonstration of the identity of structure of the medusae of the Scypho-medusan and Hydromedusan groups, and, further, the mode of development of the Scyphomedusoe from the scyphistoma and the relations of the generative products to the enteric cavity, combine to render the view that the polymorphous and dimorphous colonies of Hydrozoa are synthesized eumeromorphs more probable, in the judgment of the present writer, than that which would explain them as dysmeromorphs.

The term “merogenesis,” and its subordinate terms, “eumerogenesis, dysmerogenesis,” &c., are applicable to units of the first order, namely, cells, as well as to the “persons” which are built up by them. Ordinary celldivision is an example of eumerogenesis; free-formation of nuclei, as in the fertilized ovum of Arthropods, is dysmerogenesis. A syncytium is usually a synthesized eumeromorph, but may be a dysmeromorph.

Definition of the Hydrozoa.— The Hydrozoa are Coelentera ηematophora, distinguished from the fellow-group Anthozoa (the name applied to Actinozoa when the Ctenophora are removed from them) by not possessing the latter’s constant and sharp differentiation of the arch-enteric cavity into axial digestive and periaxial septate portions, usually by a simpler form of nematocyst, and generally by lower histological differentiation?

The following is a brief summary of the chief characters of the larger divisions of the Hydrozoa :—

Sub-class I. Scyphomedusae. —These are Hydrozoa which in the adult condition always have four or eight interradial groups of gastral filaments (“pha-cellae” of Haeckel) (figs. 16 (6), 23,and 26). Thegeni-talia (ovaria and spermaria) are developed from endoderm, and are always interradial (in the four radii formed after the first four). The hydra-form is not a “hydroid,” but a short polyp with broad hypostome—the “scyphistoma,’’which gives rise to medusa-forms by transverse fission (strobilation), or itself develops genitalia (Lucernariae). Combined visual and auditory organs in the form of modified tentacles (tentaculocysts) to the number of four, eight, or more occur on the edge of the disc (except in Lucernariae, where they are represented by the “colleto-cystophors”). The medusa-form in some cases develops from the egg without the intermediate scyphistoma-stage (Pelagia, CharybdoeaT). The edge of its disc is provided with lappets, which cover the sensorial tentaculocysts (hence Steganophthalmia of Forbes), and is not provided with a velum (hence “Acraspeda ” of Gegenbaur), excepting the rudimentary velum of Aurelia (fig. 31) and the well-developed vascular velum (pseudo-velum) of Charybdoea (fig. 21). There is no continuous marginal nerve-ring (except in Charybdoea), but several separate marginal nerve centres (hence Toponeura of Eimer). The diblastula in all cases, as yet observed, is formed by invagination, the blastopore closing up (Balfour).

The binary division of the Hydrozoa was established by Eschscholtz {System der Acalephen, 1829) whose Discophorae phanero- carpoe correspond to the Scyphomedusae, whilst his Discophorae cryptocarpae represent the Hydromedusae. The terms point to distinctions which are not valid. In 1853 Kolliker used the term Dis cophoro, for the Scyphomedusae alone, an illegitimate limitation of the term which was followed by Louis Agassiz in 1860. Nicholson has used the term in the reverse sense for a heterogeneous assemblage of those medusae not classified by Huxley as Lucernaridoe, nor as yet recognized as derived from hydroid trophosomes. This use of the term adds to the existing confusion, and renders its abandonment necessary. The term Discomedusae was used for the Scyphomedusae by Haeckel in his Generelle Morphologie (excluding Charybdoea) — whilst Carus {Handbuch, 1867) confines the term “ Medusae ” to them alone, which is objectionable, since it belongs as justly to the Hydromedusae. Forbes’s term for them, Stcganoph-thalmia, indicates a true characteristic, failing only in the Lucer- narioe, but its complementary term Gymnophthalmia is inaccurate. Similarly the terms Acraspeda and its complement Graspcdota are inacceptable. Eimer has proposed to use the terms Toρoneura and Cycloneιιra for the two divisions—but Charybdoea appears to break down this division as so many others. The old term Acalephae, which is retained by Gegenbaur in its proper sense for all the Coelentera nematophora, is used as the designation of the Scypho- medusoe alone by Claus {Grundzüge der Zook, 1878), which cannot fail to produce confusion. The term Liicernaridae, proposed so long ago as 1856 by Huxley {Med. Times and Gazette), most truly indicates the relationships of these organisms which he was the first to recognize, but it seems desirable to restrict this term to the limited order in which Lucernaria is placed, and to employ for the larger group— Scyphomedusae —a term which is the true complement of the convenient name assigned to the other division of Hydrozoa, viz., Hydromedusoe.

Order 1. Lucernaria?, — Scyphomedusae devoid of tentaculocysts, with the aboral pole of the body produced into au adhesive disc by which the organism (which possesses the power of swimming by contraction of the circular muscular zone of the hypostome) usually affixes itself. The enteric cavity is divided into four perradial chambers by four delicate interradial septa. The genitalia are developed as four-paired ridges at the sides of the interradial septa on the oral wall of the chambers (fig. 19). No reproduction by fission nor “alternation of generations” is known in the group. At the edges of the disc capitate tentacles are developed in eight adradial groups; between these are modified tentacles in some genera,—the marginal anchors or colleto-cystophors. The canal system which has sometimes been described in them is a product of erroneous observation. A very few genera and species of this order are known. They may be justly called the coenotype of the medusae (James Clark), and their relationship to the free swimming forms may be compared, as was done by L. Agassiz, to the relationship of the stalked Crinoids to such forms as Comatula. Three species are not uncommon on the British coasts.

By Milne Edwards the animals forming this group were termed Podαctinαriα and associated with the A nthozoa. By Leuckart they were termed Calycozoa ; it is only of late that the closeness of their relationship to the Scyphomedusae has been fully recognized, though long since insisted on by Huxley and by James Clark. Haeckel in his new system of the medusae {Sitzungsbcr. der Jenaische Gesellschaft fiir Medicin und Naturwiss., July 26, 1878) adopts for them the term Scyphomedusae in allusion to their permanently maintaining the distinctive features of the scyphistoma larval form of the Acrasρed<Λ the term which he adopts from Gegenbaur for our Scyphomedusae.

Order 2. Discomedusae. — These are Scyphomedusae developing as sexual rnedusiforrn persons by transverse fission from a scyphistoma, or else directly from the egg. They have eight tentaculocysts, four perradial, four interradial, and sometimes accessory ones (adradial). Four or eight genital lobes (ovaria or spermaria or hermaphrodite) are developed from the endoderm forming the oral floor of the central region of the enteric cavity, which is produced into a corresponding number of pouches. The mouth is either a simple opening at the termination of a rudimentary manubrium (sub-order Cubostomoe), or it is provided with four or eight arm-like processes (sub-orders Semostomae and Rhizostomae). In the sub-order Rhizostomae (fig. 24, a), the edges of the oral opening fuse together at an early age and leave several sucker-like secondary mouths, which were formerly mistaken for independent persons. The central enteric chamber is continued through the disc by a complicated often reticulate system of radiating canals, which excavate the endoderm lamella.

In the Semostomoe and Rhizostomoe (not in the Cubostomoe) four remarkable (respiratory) sub-genital pits (fig. 28) are hollowed out in the gelatinous substance of the sub-umbrella (oral face of the umbrella). These do not communicate, as has been erroneously supposed, with the genital organs, the products of which normally are evacuated by the mouth. In the Tetragamelian Rhizostomoe these pits remain distinct from one another as in Semostomoe, but in the Monogamelian Rhizostomoe they unite to form one continuous sub-genital cavity placed between the wall of the enteric cavity and the polystomous oral disc. The common English forms, Aurelia, Chrysaora, and Cyanoea, are types of the Semostomoe, the somewhat less common Rhizostoma of the Monogamelian Rhizostomoe, whilst Nausithoe and Discomedusa represent the simple Cuoostomoe.

The writer has adopted the term used by Haeckel for this order, and is indebted to his preliminary notices of a large work on the Medusae, now in the press, for outlines of the classification and definitions which have been introduced with modifications in relation to these and the other Medusae. The term Discοphοrα is used by Claus (Gτwηdzuge} for the Biseomedusae. It is quite clear from the varied and inconsistent use by different authors of that term, and also of the terms Acalephae and Medusae, that they must be ejected altogether from use in systematic treatises.

The structure of the common Aurelia aurita and its development have recently formed the subject of investigation by Claus, Eimer, and others. As the current accounts in text-books are very inadequate, a short sketch of the morphology of that form is appended here.

From the egg, according to the researches of Claus (whose figures, here reproduced, refer more especially to the closely allied genus Chrysaora, up to the compl e-tion of the scyphistoma), a single-cell-layered blastula develops which forms a diblastula by invagination (fig. 25, A, B, C). The orifice of invagination closes up, and the ciliated “planula” (as this stage used to be termed in all Coelentera), after swimming around for a time, fixes itself, probably by the blastoporal pole. The true mouth then forms by inruption at the opposite pole. Two tentacles now grow out near the mouth opposite to one another (fig. 25, D), and are followed by two more (fig. 26), these indicating the four primary radii of the body which pass through the angles of the foursided mouth, and are termed perradial. Meanwhile the aboral pole narrows and forms a distinct stalk, which in Chrysaora secretes a horny perisarc (fig. 25, D). Four new tentacles, those of the intermediate or secondary radii, now appear between the first four, and are termed interradial. At the same time four longitudinal ridges grow forward on the wall of the enteric cavity (fig. 26). These interradial ridges have sometimes been erroneously described as containing each a longitudinal canal connected with a circular canal at the base of the tentacles. They are in reality solid, as is the margin of the hypostome from which the tentacles spring. It is in connexion with these four ridges that the gastral filaments will subsequently appear, as also the genital organs either along their middle line or adradially to them. The ridges correspond to the mesenteries of the Anthozoa. Eight additional tentacles placed one on each side of the perradial tentacles (or of the interradial, according as we may choose to regard the matter) next appear, and are distinguished as αdradial. All the tentacles reaching an equal size, we obtain the appearance seen in fig. 26, when the young scyphistoma is looked at from above. Looked at from the side, with its wide hypostome and short vertical axis, the scyphistoma differs widely from an ordinary hydra-form, and approaches the medusa-form, to which its four longitudinal gastral ridges further assimilate it. The little creature is now about an eighth of an inch in height; in other genera, but not in Chrysaora, it may now multiply by the production of a few buds from its fixed basal disc. After nourishing itself for a period, and increasing to four or five times the size just noted, the vertical axis elongates and a series of transverse constrictions appear on the surface, marking off the body of the scyphistoma into a series of discs (figs. 26 and 27), each of which by the development of tentacles and completion of the constriction will become a separate medusa (in its young state called “ephyra ”). The tentacles of the Aurelia and the structure of the margin of its hypostome are very different from those of the scyphistoma. They are exhibited in their earliest condition (when the Λwe‰-medusa is first liberated from its attachment and is an ephyra) in fig. 26, E, F. The margin of the hypostome is drawn out into eight arms (which are not to be confused with tentacles); the end of each arm is bifid, carrying a pair of lappets—the marginal lappets which persist in the adult (see figs. 30 and 31). Between the lappets is placed a short and peculiar tentacle, the tentaculocyst or sense-organ. The eight arms of the disc and their tentaculocysts are perradial and interradial. As the organism grows, a set of eight adradial tentacles appear in the notches between the eight arms, but never attain any relatively large size in Aurelia. The asteroid arm-bearing character of the margin of the disc is soon obliterated by the relative growth of the intermediate adradial areas, which become quite filled up, so that in the adult the tentaculocyst is carried in a notch instead of on a prominence, and is concealed by the two lappets (figs. 28 and 30). The margin of the disc between adjacent pairs of lappets gives rise to a fold which grows inwards (toward the mouth) during an early stage (fig. 31), and numerous small tentacles (the fringe) appear along the margin of the disc, which soon equal in size the first adradial tentacle. The ingrowing fold is the velum or “pseudo-velum,” and never increases in size, so that in the adult it is not observ able. The tentacles also remain very small and fine in Aurelia, forming a continuous fringe along the edge of the disc, interrupted only by the eight notches for the tentaculocysts (fig. 29),

The sixteen tentacles of the scyphistoma are necessarily attached to the most anterior of the pile of medusae; they atrophy, but to what extent they may be metamorphosed to form the parts of the ephyra or young medusa has not been determined. The scyphistoma, having given rise to its pile of ephyrae, may (in some genera, AureliaX) redevelop its own kind of tentacles below the constriction marking off the last ephyra. Hence scyphistoma tentacles appear sometimes at the top and sometimes at the bottom of the pile, which has led to diverse accounts of the mode of development of the ephyrae.

Whilst changes are going on in the configuration of the margin of the disc of an ephyra on its way to the perfect form of the adult Aurelia, the enteric cavity has also undergone most important changes. Foremost in importance is the development of a single gastral filament on each of the four gastral ridges which necessarily are present in the transverse slice (so to call it) of a scyphistoma, which becomes an ephyra (fig. 26). These rapidly increase in number as the ephyra grows. Further, the enteric cavity at first follows the outline of the ephyra, sending a process into each arm, but then by adhesion of its walls is converted into a four-lobed central chamber, a marginal canal, and an endoderm lamella. A system of canals, the arrangement of which is seen in figs. 29 and 31, subsequently opens out again certain lines and tracts of the conjoined endoderm walls.

In the adult Aurelia we find the mouth surrounded by four large arm-like ρerradial processes (figs. 25 and 29) (not tentacles), and leading through a short manubrium into a flattened four-lobed chamber, the lobes being interradial, and having on their oral floor numerous gastral filaments (rich in thread cells) (6 in fig. 16). Each pouch or lobe gives off a canal, which runs towards the circular canal at the margin of the disc, but breaks up into three or four secondary canals on its way. Between the pouches come off eight other “radiating” canals (adradial), which do not branch, but go straight to the circular canal.

The oral floor of the concavity of each lobe of the enteric cavity is occupied by a horse-shoe-shaped frill (fig. 29, ov), either testis or ovary (the sexes being in separate individuals). The open arms of the horse-shoe are turned towards the centre of the disc, and the folds of the genital frill are so deep as to show themselves on the outer ectodermal wall of the disc. Here, however, there is a very remarkable arrangement, which has rarely, if ever, been correctly described and figured in our common Aurelia. The gelatinous substance of the disc is hollowed out on that part of the oral face corresponding to the position of the genital frills, so as to form four separate extensive pits or chambers. Each of these sub-genital pits has in Aurelia a small round opening on the oral face of the disc (fig. 28, GP), but is otherwise entirely closed, having no communication with the genital tissues, from which it is separated by a delicate layer of ectoderm (6 in fig. 16). The pits probably serve to admit water for respiratory purposes into close proximity with the genital tissues.

The whole enteric surface, including canals, is ciliated, whilst the ectoderm is not ciliated, but provided with groups of nematocysts.

The tentaculocyst in the adult Aurelia is relatively an extremely minute body, completely hidden by the two large marginal lappets (fig. 30, T). Above it (that is, on the aboral surface, as the Aurelia swims) is a deep pit (A), Schafer’s fovea nervosa superior, sunk in a sort of bridge which connects the two lappets and overhangs the tentaculocyst. A similar pit (the fovea inferior) exists on the oral surface. These have been recognized by Claus, Eimer, and the Hertwigs as olfactory organs. The tentaculocyst is seen in section in fig. 30 (right-hand figure), which exhibits its central cavity continuous with the enteric cavity, its ectodermal pigment spot (eye), and its endodermal mass of concretions (auditory organ).

The chief muscular mass of Aurelia, except that of the oral arms, is a circular zone on the oral face of the disc. The muscular fibres are not distinct cells, but transversely-striated processes of the epidermic cells (epidermo-muscular cells) (fig. 9). In the “arms” of other medusae, and presumably of Aurelia, the muscular fibre is formed by independent nucleated cells (fig. 9).

The nerve-epithelium from the olfactory pits of Aurelia is drawn in fig. 14. Starting from this and from the cells of the tentaculocysts are nerve-fibres, which spread themselves on the surface of the circular muscular zone in the neighbourhood of the tentaculocysts, and these are connected each and separately with large isolated nerve-ganglion cells (fig. 15). The nerve-fibre is continued beyond the cell, and in some instances has been traced into a broadened expansion lying on a muscular fibre (Schäfer). The nerveganglion cells lie very superficially immediately below the flat epithelium of the body surface and between it and its muscular processes.

Tbe ova and spermatozoa of Aurelia develop in the genital frills from endoderm cells in separate individuals. They pass to the exterior through the mouth.

Order 3. Conomedusoe, — Scyphomedusoe with only four tentaculocysts, and these perradial. A broad velum (so-called pseudo-velum) of complete circular form is present, differing from that of the Hydromedusoe in the fact that it is penetrated by canals of the enteric system (Charybdoea). The whole umbrella is bell-shaped. The genital organs are four pairs of lamelliform ridges (fig. 22) which are attached to the four narrow interradial septa that divide the large enteric cavity of the umbrella into four perradial gastrocanal pouches. The lamelliform genital glands hang freely in these pouches. At the edge of the umbrella are four interradial lappet-like prolongations of the gelatinous substance of the disc, which support each a long tentacle (fig. 20). The nerve-ring is complete, like that of the Hydromedusoe.

There is now no doubt that Charybdoea, which has been placed in each of the two large divisions of the Ηydrοzοa, must be classed with the Scyphomedusoe. The recent investigations of Claus (Arbeiten aus dem Zool. Institut zu Wien, Bd. i. Hft. ii., 1878), as well as those of Haeckel and Fritz Müller, lead to this conclusion. The term Conomedusoe is adopted from Haeckel, who places here, besides Charybdoea and Tamoya, other forms, a fuller description of which maybe expected in his forthcoming System der Medusen. In many respects—its quadrangular form, its marginal lappets, its broad enteric pouches in place of fine canals, its vascular velum, and its highly complicated tentaculocysts (fig. 13, B)— Charybdoea is peculiar. The simplicity of the enteric system and the arrangement of the genital glands brings it near to Lucernaria. The existence of four interradial groups of gastral filaments, and the disposition of the paired genital glands at the sides of the interradial septa, determine its position to be among the Scyphomedusoe. Its development is not known. Figs. 20 to 23 illustrate the structure of Charybdoea.·

Order 4. Peromedusoe, — Scyphomedusoe with four interradial tentaculocysts. The enteric system consists of three divisions,—an aboral main stomach with four interradial gastral ridges and filament groups; a mid-stomach, which communicates by means of four perradial slits with a very large ring-sinus (occuping two-thirds of the umbrella); and thirdly, an oral portion or pharynx, with four wide perradial pouches. The genital organs are four pairs of sausage-shaped interradial ridges lying on the oral floor of the ring-sinus.

This is a new group founded by Haeckel, of which we have at present no further details.

Sub-class II. Hydromedusae. —These are Hydrozoa devoid of gastral filaments; the sexual persons are always medusiform, the genital glands are developed sometimes from ectodermal cells, sometimes from endoderm, and are always perradial (in the radii of the first order). The medusiform persons always possess a muscular non-vascular velum (hence Craspedota) and a complete nerve-ring (heuce Cycloneura of Eimer). The marginal sense organs are either ocelli or otocysts or tentaculocysts. The diblastula, in all cases as yet observed, is formed by delamination-(Balfour). The sexual medusiform persons may develop directly from the egg, but more usually the egg gives rise to a hydriform person—the hydroid—which differs from a scyphistoma in its elongate vertical axis, the indefinite number (often also position) of its tentacles, and its frequent formation of a colony of large size by lateral budding. By lateral budding (not by metameric fission) medusiform persons which alone develop sexual glands are produced on the hydriform colonies; these may separate from the colony, or may be retained in a more or less degenerate form adherent to it, as generative buds or warts.

The medusiform persons of this group are the Discophoroe crypto· carpae of Eschscholtz, the Graspedoia of Gegenbaur (1854), and the Hydromedusida of Kolliker (1853)—the last two authors at that time separating the hydriform persons as Hydroidea. Louis Agassiz (1860) includes both sets of persons under the term Hydroida (together with Lucernaria), which also is the term adopted by Allman in his beautiful monograph (1871-74). J. V. Carus, amending the limitations given by Carl Vogt, was the first to use the term Hydromedusae in the sense here adopted (Handbuch der Zoologie, 1863), and it is now employed in the same sense by Gegenbaur (Elements of Comparative Anatomy, London, 1878), namely, to embrace both the cryptocarpous medusae of Eschscholtz and the hydroids related to them. The term Hydromedusae is used unwisely by Claus (Grundzüge d. Z.) for the whole group of Hydrozoa. It has been the practice of some authors to give a double classification of the group—one based on the characters of the medusiform persons, the other on that of the hydriform persons. In the present article a double name will in some cases be assigned to a group— but the attempt is made to bring both sets of persons under one system.

Order 1. Gymnoblastea-Anthomedusoe. — These are Hydro- medusae which all, as far as is known, pass through a hydriform phase, but in which the medusiform persons may either reach full development or exhibit the extremest degeneration (Hydra). The ectoderm of the hydriform persona may secrete a horny tubular protective case (perisarc), but this does not form cups for the reception of the tentacular crown nor cases enclosing groups of medusiform buds (gonangia). The fully-developed medusiform persons never possess otocysts nor tentaculocysts, but always ocelli at the base of the tentacles. The latter are usually four or six, corresponding to the same number of simple radial enteric canals, but may be more numerous or reduced to one or to two; rarely they are branched (Cladonemd).

The sexual glands are placed in the wall of the manubrium, either equally distributed all round it or in four separate perradial groups, which are often divided into eight adradial groups by the perradial longitudinal muscles.

This is a very well defined group, since the Gymηοblastea of Allman, based on the characters of the hydriform persons—also known as Tubularioe and Gymm>tοka— correspond exactly with the Anthomedusae of Haeckel’s new system. Hydra is included here, though placed in a separate order by Allman. Some of the leading forms of hydriform and medusiform persons are given in the cuts (figs. 34 to 42). The greatest range in the amount of degeneration of the medusiform persons is seen even in genera of the same family— e.g., Turris and Clava —the former producing free medusae, the latter sessile sporosacs. The Oceanidae of Gegenbaur (excluding the Williadoe, which Haeckel assigns to the next group) correspond on the whole to the medusa-forms of this order.

Order 2. Calyptoblastea-Leptomedus<x.— These are Hydro- medusoe of which the hydriform phase is known in a large number of cases, whilst of others only the medusaforms are known; none are known to develop directly from the egg to the medusa-form. As in the preceding group, the medusiform persons may reach full development or exhibit themselves as degenerate sexual sacs on the hydriform colonies. The ectoderm of the hydra-forms always secretes a perisarc which forms a cup-like protection (hydrotheca) to the tentacle-crown, and which also encloses the group of medusa-buds in peculiar horny cases (gonangia). The fully-developed medusiform persons (fig. 47) either have no otocysts, but only ocelli (Ocellatoe), or they have otocysts (fig. 11) (ectodermal sacs), four, eight, or over a hundred, not homologous with tentacles, and sometimes in addition ocelli (Vesic ulatoe). The radial enteric canals are usually four or eight in number, but may be more numerous, whilst the marginal tentacles of the disc are either few or very numerous. The genital glands always are placed in the course of the radial canals of the disc (not in the manubrium), and stand out as groups of wart-like processes on the sub-umbrellar surface (fig. 43). Their mode of discharge is uncertain.

The Cαlyptoblαsteα of Allman, Skenotoka of Carus, and Campanu∙ lariae of authors form a well-marked group of hydroids which, when they give rise to free rnedusae, give rise to those termed Lcptomedusae by Haeckel, corresponding to the Thaumantiadae and Eucopidae of Gegenbaur’s system. The calyptoblastic hydroid Leptoscyphus, which, according to Allman, gives rise to a 2λ'zziα-like medusa (Antho∙ rnedusae), is the only recorded exception to this correspondence. The Aequoridae and other medusae of similar structure have not been traced into connexion with any hydriform trophosome, but we are not justified therefore in concluding that they develop directly from the egg without hydriform phase. The chief point distinguishing the Lcptomedusae as a lot from the Anthomedusae is the development of the generative bodies in the radial canals. This position is similar to that occupied by the same organs in Trachomedusae and Scyphomedusas. Allman, however, considers the genital glands of the Lcptomedusae, not as mere glands like those of Aurelia or Charybdoea, but as a series of buds—a generation of aborted medusae or sporosacs. In consequence he terms the medusa of the Lcptomedusae a blastocheme (or bud-producer), as distinguished from a gonocheme (or genital-producer). In support of this view,

Allman (Monograph, 1874) adduces the various remarkable cases of production of buds by medusae which have been recorded (fig. 44, 45, 46), and, further, the very striking similarity between the structure of a lobe of the genital gland of Obelia and a sporosae such as we find in Hydractinia. It seems necessary to accept Allman’s view on this matter, unless we are prepared to abandon the homology of sporosacs with medusae in the case of hydriform persons.

The colonies of hydriform persons of the present group differ inter sc according to the arrangement of the cups or hydrothecae. In Plumularidae they are sessile, and all on one side of a branch; in Sertularidoe they are sessile, and alternately placed on either side; in Campanularidae each cup is raised on a pedicel or stalk. The medusiform persons sometimes remain abortive and sessile in their gonangia.

Order 3. Trachomedusae, — Hydromedusae which have as sense-organs tentaculocysts. The otoliths (fig. 12) are formed by endodermic cells as in Scyphomedusoe, and ocelli may or may not be present on the tentaculocyst.

The genital glands have the form of wide outgrowths or lamelliform enlargements in the course of the radial canals (figs. 48, 49). No hydriform phase is known in any member of this group, and one at least (Geryonid) has been observed to develop from the egg directly into the medusaform.

Order 4. Narcomedusae. — These have the same characters as the Trachomedusae, excepting that the genital glands are in the wall of the manubrium or in pocket-like radial outgrowths thereof (figs. 50 and 51). Further, the marginal tentacles of the disc possess peculiar “roots,” which can be traced upwards into the gelatinous substance of the body. No hydriform phase has been observed in this group, whilst Aegina and Aeginopsis have been shown to develop directly from the egg to the medusa-form.

The two orders Trachomedusae and Narcomedusae are established by Haeckel in his new “system” for the peculiar forms classed by Carus as Haplomorpha, and by Allman as Monopsea. These latter names have reference to the fact that no hydriform phase is known to occur in the life-history of these organisms, a fact which is not peculiar to them, and, if it should prove to be not universal amongst them, would by no means invalidate their claim to a distinct position on the grounds afforded by the characters above given. They are remarkable for a certain hardness and stiffness of the gelatinous substance of the disc, or at any rate of the cellular axis of the tentacles, on accout of which the orders are contrasted by Haeckel as Trachylinoe with Anthomedusae and Leptornedusae, which are termed Leptolinoe ; a curious parallelism as to the position of the genitalia exists between Anthomedusae and Narcomedusae on the one hand and Leptornedusae and Trachomedusae, on the other. The orders present a very high degree of development, both in coarser and histological differentiation. At one time it was supposed, in accordance with Haeckel’s observations, that Geryonia (Carmarina, fig. 48), one of the Trachomedusae, gave rise by buds from its enteric walls to young Cuninoe (Narcomedusae, fig. 50), but this has been explained by the observations of Franz Schulze and of Uljanin as due to parasitism, young Cuninoe in the condition of ciliated Blanuloe entering the mouth and enteric chamber of the Carmarina. The same explanation probably applies (Claus) to the Supposed internal buds of Cunina observed by Gegenbaur, Fritz Müller, and Mβtschnikow. The process is sufficiently remarkable according to the last observer, for the first generation of buds produce a second generation by external gemmation, before attaining the characters of the parent CuniiM. The anatomy of these forms is fully given in Haeckel’s memoirs in the Jcnaische Zeitschrift, vols, i. and ii., 1864-66; also further details as to Carmarina are given in Eimer’s Medusen, 1878.

Order 5. Hydrocorallinae. — These are Hydrcymedusae in which the hydriform phase forms large colonies, presenting a copious calcareous deposit in the ectodermal tissue (coenosteum of Moseley), leaving only the hydranths or tentacular region free from such hardening. The medusiform persons are, at present, only known in the degenerate form of sporosacs, which occupy cavities (ampul lae of Moseley) in the hardened base of the colony {Stylasteridce). No such cavities have been detected in others (Milleporidae), which may, therefore, give rise to complete medusiform persons.

In all a marked polymorphism has been observed (fig. 53), consisting in the differentiation of longer tentacle-like persons (dactylozooids) and shorter mouth-bearing persons (gastrozooids). The persons of both kinds are either scattered irregularly or the dactylozooids are arranged around the gastrozooids in cyclosystems of greater or less definiteness, or in distinct rows (fig. 55). The position oι these two kinds of hydriform persons is marked by definite groups of pits (cyclosystems) in the dried calcareous skeleton of the colonies, which simulate the calycles of the stony corals (Anthozoa}.

Louis Agassiz was the first to recognize the true nature of the Millcporidae, and his imperfect observations have been fully confirmed and greatly extended by Mr Moseley (PhU. Trans., 1878) who added the Stylasteridoe previously regarded as Anthozoa to the category of calcigenous hydroids, and founded the order of Hydrocorallinoe. The Stylasteridoe differ from the Millcporidae in possessing a calcified axial style at the base of the dilated portion of each gastrozooid, and further in the ascertained development of sporosacs, and in the greater complication of their cyclosystems. These forms are abundant in tropical seas, and contribute with the Anthozoa and Corallines to the formation of coral reefs. Allopora and Stylaster occur off the Norwegian coast. The woodcuts illustrating the structure of this group are borrowed from Mr Moseley’s Notes of a Naturalist on the “Challenger.”

The nearest allies of the Hydrocorallinoe are such polymorphic Gymnoblastea as Hydractinia (fig. 39); the definite division of labour and the polymorphism in the former, together with their calcigenous peculiarity, entitle them to rank as a distinct order.

Order 6. Siphonophora.— These are Hydromedusoe in which hydriform persons alone (Velella) or hydriform persons and sterile medusiform persons are united, under many special modifications of form, to constitute floating colonies of very definite shape and constitution. In addition to these are developed medusiform sexual persons which usually are sporosacs and only exceptionally attain full development so as to be liberated from the colony as free-swimming medusae {Velella, as Chrysomitra' i Phy salia, only liberating female medusae). The medusiform persons, where sufficiently developed, exhibit the velum characteristic of Hydromedusoe-, the larger mouth-bearing nydriform persons, which are sometimes the only representatives of their kind, are remarkable for differentiation into four regions,— a proboscis, a stomach, a basal ring, and a short stalk on which the single tentacle of great length is situated (fig. 56, f). In the suborder Physophoridoe (fig. 57, C) the persons are united by a short or long and spiral stem, terminated at one end by a flask-like air-sac (pneumatocyst); below the air-sac a biserial or multiserial range of swimming-bells (nectocalyces = medusae with suppression of manubrium, tentacles, and sense-organs) are placed. Covering pieces (hydrophyllia, reduced medusae) and dactylozooids are affixed to the succeeding region of the stem, and alternate in definite order with the mouth-bearing hydriform persons (polyps or nutritive persons) and generative medusiform persons. In the sub-order Physalidoe the stem is converted into an air-sac, enormously enlarged, and the nectocalyces and hydrophyllia are absent. In the sub-order Calycophoridoe the air-sac is not developed, the nectocalyces are in a biserial group, or reduced to two or to one. Dactylozooids are wanting. The modified persons (appendages, Huxley) arise from the stem in groups, and can be withdrawn into the cavity of a swimming-bell (fig. 57, B).

Each group consists of a nutritive person, with long tentacle, of generative medusoids, and usually also an umbrellashaped or funnel-like covering piece. The latter separate in some Dijjhyidoe, and lead an independent life as Eudoxioe.

In the suborder Discoidoe the stem is converted into a flattened disc with a system of canalicular cavities. Above this lies the air sac, a flattened reservoir of cartilaginous consistence. The hydriform persons depend from the disc, centrally a large nutritive person surrounded by smaller similar persons carrying at their bases the generative medusoids -, near the edge of the disc are dactylozooids. The medusoids develop into complete medusiform persons, and develop the genital products after liberation from the colony, when they are known as Chrysomitra.

The Siphοnοphοraaλcme, amongst the colonies formed by Ηydrο~οa, exhibit a high degree of division of labour and consequent individuation. The mode of origin of such colonies has been discussed above. The locomotive habit, as contrasted with the sessile habit of other colonies, is no doubt correlated with the sharply defined individuality which they attain (compare Cristatclla among Pοlyzοa). Velella and Physalia are occasionally seen on the southern and western shores of England, but as a rule the Siphonophora are met with only in the open ocean and in the Mediterranean. By some authorities the Siphonophora are assigned a distinct position among the Hydrozoa, side by side with the Hydromedusoe and Scyphomcdusoe ; their interpretation as floating colonies of Hydromedusoe, an interpretation necessitated by the structure of their medusiform persons, forbids their separation from that group.

Fossil Hydrozoa. —The researches of Moseley have necessitated a redistribution of the group of Anthozoa known as the Tabulata. Among these appear to be a few Hydro- corallinoe, which occur in the fossil state. The Palaeozoic forms known as graptolites are by some authors assigned to the Hydrozoa, but the grounds for placing them in this position are very slight, owing to the imperfect nature of the remains. A discussion of the small amount of structure which they present would be out of place here.

Remarkable Scyphomedusoe have been obtained from the Solenhofen slates (Jurassic); excepting these, no noteworthy extinct Hydrozoa are known (see Haeckel in Zeitsch. wist.

Zooh, vols, xv., xix., and Jenaische Zeitsch., vol. viii., 1874).

Relationship of the Ctenophοra to the Hydrozoa.— The remarkable medusa-form recently described by Haeckel (Sitzung sb er. Jenaische Gesellsch., 1878) as Ctenaria cteno-ρhοra, and classed by him amongst the Anthomedusoe, seems to furnish a very direct transition from the structure of a medusa to that of such a ctenophor as Cydippe (Pleurο- brαchia). The woodcut and appended explanation (fig. 58) copied from Haeckel’s memoir will render the relations of the two forms clear. Ctenaria has the margin of its disc narrowed so as to give the organism a spherical form. The approximated margins bound an orifice leading to the sub-umbrella space. This orifice corresponds to the so-called mouth of a Cydippe. Further, Ctenaria has two, and only two, long-fringed tentacles, like those of Cydippe, and each springing from a pocket as in that genus, and on the surface of its spheroidal umbrella eight rows of differentiated ectodermal cells, which though not ciliated correspond closely in position with the eight ctenophoral ambulacra of Cydippe. The disposition of the enteric canalsystem of Ctenaria is, as shown in the cut, also transitional in the direction of Cydippe. Apart from the existence of Ctenaria, the homologies suggested by Haeckel between Hydromedusae and Ctenophora are such as to commend themselves very strongly to acceptance. (e. r. l.)