About Maximapedia

Hydromedusae

HYDROMEDUSAE, a group of marine animals, recognized as belonging to the Hydrozoa (q.v.) by the following characters.

(1) The polyp (hydropolyp) is of simple structure, typically much longer than broad, without ectodermal oesophagus or mesenteries, such as are seen in the anthopolyp (see article ANTHOZOA); the mouth is usually raised above the peristome on a short conical elevation or hypostome; the ectoderm is without cilia.

(2) With very few exceptions, the polyp is not the only type of individual that occurs, but alternates in the life-cycle of a given species, with a distinct type, the medusa (q.v.), while in other cases the polyp-stage may be absent altogether, so that only medusa-individuals occur in the life-cycle.

The Hydromedusae represent, therefore, a sub-class of the Hydrozoa. The only other sub-class is the Scyphomedusae (q.v.). The Hydromedusae contrast with the Scyphomedusae in the following points, (i) The polyp, when present, is without the strongly developed longitudinal retractor muscles, forming ridges (laeniolae) projecting into the digestive cavity, seen in the scyphistoma or scyphopolyp. (2) The medusa, when present, has a velum and is hence said to be craspedote; the nervous system forms two continuous rings running above and below the velum; the margin of the umbrella is not lobed (except in Narcomedusae) but entire; there are characteristic differences in the sense-organs (see below, and SCYPHOMEDUSAE); and gastral filaments (phacellae), subgenital pits, etc., are absent.

(3) The gonads, whether formed in the polyp or the medusa, are developed in the ectoderm.

The Hydromedusae form a widespread, dominant and highly differentiated group of animals, typically marine, and found in all seas and in all zones of marine life. Fresh-water forms, however, are also known, very few as regards species or genera, but often extremely abundant as individuals. In the British fresh-water fauna only two genera, Hydra and Cordylophora, are found; in America occurs an additional genus, Microhydra. The paucity of fresh-water forms contrasts sharply with the great abundance of marine genera common in all seas and on every shore. The species of Hydra, however, are extremely common and familiar inhabitants of ponds and ditches. i In fresh-water Hydromedusae the life-cycle is usually secondarily simplified, but in marine forms the life-cycle may be extremely complicated, and a given species often passes in the course of its history through widely different forms adapted to different habitats and modes of life. Apart from larval or embryonic forms there are found typically two types of person, as already stated, the polyp and the medusa, each of which may vary independently of the other, since their environment and life-conditions are usually quite different. Hence both polyp and medusa present characters for classification, and a given species, genus or other taxonomic category may be defined by polyp-characters or medusa-characters or by both combined. If our knowledge of the life-histories of these organisms were perfect, their polymorphism would present no difficulties to classification; but unfortunately this is far from being the case. In the majority of cases we 'do not know the polyp corresponding to a given medusa, or the medusa that arises from a given polyp. 1 Even when a medusa is seen to be budded from a polyp under observation in an aquarium, the difficulty is not always solved, since the freshly-liberated, immature medusa may differ greatly from the full-grown, sexually-mature medusa after several months of life on the high seas (see figs, n, B,C, and 59, a, b, c). To establish the exact relationship it is necessary not only to breed but to rear the medusa, which cannot always be done in 1 In some cases hydroids have been reared in aquaria from ova of medusae, but these hydroids have not yet been found in the sea (Browne [10 o]).

confinement. The alternative is to fish all stages of the medusa in its growth in the open sea, a slow and laborious method in which the chance of error is very great, unless the series of stages is very complete.

At present, therefore, classifications of the Hydromedusae have a more or less tentative character, and are liable to revision with increased knowledge of the life-histories of these organisms. Many groups bear at present two names, the one representing the group as defined by polyp-characters, the other as defined by medusa-characters. It is not even possible in all cases to be certain that the polyp-group corresponds exactly to the medusagroup, especially -in minor systematic categories, such as families. The following is the main outline of the classification that is adopted in the present article. Groups founded on polypcharacters are printed in ordinary type, those founded on medusacharacters in italics. For definitions of the groups see below.

Sub-class Hydromedusae (Hydrozoa Craspedotd). Order I. Eleutheroblastea.

II. Hydroidea (Leptolinae).

Sub-order I. Gymnoblastea (Anthomedusae) . 2. Calyptoblastea (Leptomedusae). Order III. Hydrocorallinae. IV. Graptolitoidea. V. Trachylinae. Sub-order I. Trachomedusae. 2. Narcomedusae. Order VI. Siphonophora.

Sub-order I. Chondrophorida. ,, 2. Calycophorida. 3. Physophorida. 4. Cystophorida.

Organization and Morphology of the Hydromedusae. As already stated, there occur in the Hydromedusae two distinct types of person, the polyp and the medusa; and either of them is capable of non-sexual reproduction by budding, a process which may lead to the formation of colonies, composed of more or fewer individuals combined and connected together. The morphology of the group thus falls naturally into four sections (i) the hydropolyp, (2) the polyp-colony, (3) the hydromedusa, (4) the medusa-colonies. Since, however, medusa-colonies occur only in one group, the Siphonophora, and divergent views are held with regard to the morphological interpretation of the members of a siphonophore, only the first three of the above sub-divisions of hydromedusa morphology will be dealt with here in a general way, and the morphology of the Siphonophora will be considered under the heading of the group itself.

i. The Hydropolyp (fig. i) The FIG. i. Diagram of a typical Hydropolyp.

a, Hydranth ; b, Hydrocaulus; c, Hydrorhiza ; /, Tentacle; ,.

ps, Perisarc, forming in the general characters of this organism region of the hydranth are described above and in the acuporhydrotheca(/,0, articles HYDROZOA and POLYP. It which, however.is only j s rarely free, but usually fixed and found in polyps of the incapable of locomotion. The foot order Calyptoblastea. , by which it is attached often sends out root-like processes the hydrorhiza (c). The column (b) is generally long, slender and stalklike (hydrocaulus). Just below the crown of tentacles, however, the body widens out to form a " head," termed the hydranth (a), containing a stomach-like dilatation of the digestive cavity. On the upper face of the hydranth the crown of tentacles (/) surrounds the peristome, from which rises the conical hypostome, bearing the mouth at its extremity. The general ectoderm covering the surface of the body has entirely lost the cilia present in the earlier larval stages (planula), and may be naked, or clothed in a cuticle or exoskeleton, the perisarc (ps), which in its simplest condition is a chit i nous membrane secreted by the ectoderm. The perisarc when present invests the hydrorhiza and hydrocaulus; it may stop short below the hydranth, or it may extend farther. In general there are two types of exoskeleton, characteristic of the two principal divisions of the Hydroidea. In the Gymnoblastea the perisarc either stops below the hydranth, or, if continued on to it, forms a closely-fitting investment extending as a thin cuticle as far as the bases of the tentacles (e.g. Bimeria, see G. J. Allman [i], 1 pi. xii. figs. I and 3). In the Calyptoblastea the perisarc is always continued above the From Allman's Gymnoblastic Bydroids, by permission of the Council of the Ray Society.

FIG. 2. Stauridium productum, portion of the colony magnified; p, polyp; rh, hydrorhiza.

hydrocaulus, and forms a cup, the hydrangium or hydrotheca (h, t), standing off from the body, into which the hydranth can be retracted for shelter and protection.

The architecture of the hydropolyp, simple though it be, furnishes a long series of variations affecting each part of the body. The greatest variation, however, is seen in the tentacles. As regards number, we find in the aberrant forms Protohydra and Microhydra tentacles entirely absent. In the curious hydroid Monobrachium a single tentacle is present, and the same is the case in Clathrozoon; in Amphibrachium and in Lar (fig. n, A) the polyp bears two tentacles only. The reduction of the tentacles in all these forms may be correlated with their mode of life, and especially with living in a constant current of water, which brings foodparticles always from one direction and renders a complete whorl or circle of tentacles unnecessary. Thus Microhydra lives amongst Bryozoa, and appears to utilize the currents produced by these animals. Protohydra occurs in oysterbanks and Monobrachium also grows on the shells of bivalves, and both these hydroids probably fish in the currents produced by the lamellibranchs. Amphibrachium grows in the tissues of a sponge, Euplectella, and protrudes its hydranth into the canal-system of the sponge; and Lar grows on the tubes of the worm Sabella. With the exception of these forms, reduced for the most part in correlation with a semi-parasitic mode of life, the tentacles are usually numerous. It is rare to find in the polyp a regular, Cnrvmnrt>ha symmetrical disposition of the tentacles fXrm n<-r"on as in the medusa. The primitive number of four in a whorl is seen, however, in Stauridium (fig. 2) and Cladonema (Allman [i], pi. xvii.), and in Clavatella each whorl consists regularly of eight (Allman, loc. cit. pi. xviii.). As a rule, however, the number in a whorl is irregular. The tentacles may form a single whorl, or more than one; thus in Corymorpha (fig. 3) and Tubularia (fig. 4) there are two circlets; in Stauridium (fig. 2) several ; in Coryne and Cordylophora the tentacles are scattered irregularly over the elongated hydranth.

As regards form, the tentacles show a number of types, of which the most important arc (i) filiform, i.e. cylindrical or tapering from 1 The numbers in square brackets [ ] refer to the bibliography at the end of this article; but when the number is preceded by the word Hydrozoa, it refers to the bibliography at the end of the article HYDROZOA.

to medusiform persons by budding from the margin of the disk; B, ^ "3fff Fort? (A~ii, nmfl i a , nf ) an, }' ^ *? * /After Alf base to extremity, as in Clava (fig. 5); (2) capitate, i.e. knobbed at the extremity, as in Coryne (see Allman, loc. cit. pi. iv.); (3) branched, a rare form in the polyp, but seen in Cladocoryne (see Allman, loc. cit. p. 380, fig. 82). Sometimes more than one type of form is found in the same polyp; in Pennaria and Stauridium (fig. 2) the upper whorls are capitate, the lower filiform. Finally, as regards structure,^ the tentacles may retain their primitive hollow nature, or become solid by obliteration of the axial cavity.

The hypostome of the hydropolyp may be small, or, on the other hand, as in Eudendrium (Allman, loc. cit. pis. xiii., xiv.), large and trumpet - shaped. In the curious polyp Myriothela the body of the polyp is differFIG. 4. Diagram of Tubularia entiated into nutritive and indivisa. A single hydriform person reproductive portions. a bearing a stalk carrying numerous Histology. The ectoderm degenerate medusiform persons or of the hydropolyp is chiefly sporosacs b. (After Allman.) sensory, contractile and pro- tective in function. It may also be glandular in places. It consists of two regions, an external epithelial layer and a more internal sub-epithelial layer.

The epithelial layer consists of (l) so-called " indifferent " cells secreting the perisarc or cuticle and modified to form glandular cells in places; for example, the adhesive cells in the foot. (2) Sensory cells, which may be fairly numerous in places, especially on the tentacles, but which occur always scattered and isolated, never aggregated to form sense-organs as in the medusa. (3) Contractile From Altaian's Gymnoblastic Hydroids, by permission of the Council of the Ray Society.

FIG. 5. Colonies of Clava. A, Clava squamata, magnified. B, C. multicornis, natural size; p, polyp; gort, gonophores; rh, hydrorhiza.

or myo-epithelial cells, with the cell prolonged at the base into a contractile muscle-fibre (fig. 6, B). In the hydropolyp the ectodermal muscle-fibres are always directed longitudinally. Belonging primarily to the epithelial layer, the muscular cells may become secondarily sub-epithelial. The sub-epithelial layer consists primarily of the so-called inter- stitial cells, lodged between the narrowed basal portions of the epithelial cells. From them are developed two distinct types of histological elements ; the genital cells and the cnidoblasts or mothercells of the nematocysts. The sub-epithelial layer thus primarily constituted may be recruited by immigration from without of other FIG. 6 A. Portion .of the body-wall of Hydra, showing ectoderm cells above, separated by " structureless lamella " from three flagellate .endoderm cells below. The latter are vacuolated, and contain each a nucleus and several dark granules. In the middle ectoderm cell are seen a nucleus and three nematocysts, with trigger hairs projecting beyond the cuticle. A large nematocyst, with everted thread, is seen in the right-hand ectodermal cell. (After F. E. Schulze.)

elements, more especially by nervous (ganglion) cells and musclecells derived from the epithelial layer. In its fullest development, therefore, the sub-epithelial layer consists of four classes of cellelements.

The genital cells are simple wandering cells (archaeocytes), at first minute and without any specially distinctive features, until they begin to develop into germ-cells. According to Wulfert [60] theprimitive germ-cells of Gonothyraea can be distinguished soon after the fixation of the planula, appearing amongst the interstitial cells of the ectoderm. The germ-cells are capable of extensive migrations, not only in the body of the same polyp, but also from parent to bud through many non-sexual generations of polyps in a colony (A. Weismann [58]).

The cnidoblasts are the mother-cells of the nematocysts, each cell producing one nematocyst in its interior. The complete nematocyst (fig. 7) is a spherical or oval capsule containing a hollow thread, usually barbed, coiled in its interior. The capsule has a double wall, an outer one (o.c.), tough and rigid in nature, and an inner one (i.e.) of more flexible consistence. The outer wall of the capsule is incomplete at one pole, leaving an aperture through which the thread is discharged. The inner membrane is continuous with the wall Fie. 6 B. Epidermo-muscular of the hollow thread at a spot im- ce ll s o f Hydra, m, muscular-fibre mediately below the aperture in the processes. (After Kleinenberg, outer wall, so that the thread itself from Gegenbaur.) (/) is simply a hollow prolongation of the wall of the inner capsule inverted and pushed into its cavity. The entire nematocyst is enclosed in the cnidoblast which formed it. When the nematocyst is completely developed, the cnidoblast passes outwards so as to occupy a superficial position in the ectoderm, and a delicate protoplasmic process of sensory nature, termed the cnidocil (en) projects from the cnidoblast like a fine hair or cilium. Many points in the development and mechanism of the nematocyst are disputed, but it is tolerably certain (i) that the cnidocil is of sensory nature, and that stimulation, by contact with prey or in other ways, causes a reflex discharge of the nematocyst ; (2) that the discharge is an explosive change whereby the in-turned thread is suddenly everted and turned inside out, being thus shot through the opening in the outer wall of the capsule, and forced violently into the tissues of the prey, or, it may be, of an enemy; (3) that the thread inflicts not merely a mechanical wound, but instils an irritant poison, numbing and paralysing in its action. The points most in dispute are, first, how the explosive discharge is brought about, whether by pressure exerted external to the capsule (i.e. by contraction of the cnidoblast) or by internal pressure. N. Iwanzov [27] has brought forward strong grounds for the latter view, pointing out that the cnidoblast has no contractile mechanism and that measurements show discharged capsules to be on the average slightly larger than undischarged ones. He believes that the capsule contains a substance which swells very rapidly when brought into contact with water, and that in the undischarged condition the capsule has its opening closed by a plug of protoplasm (x, fig. 7) which prevents xiv. 5 a access of water to the contents; when the cnidocil is stimulated it sets in action a mechanism or perhaps a series of chemical changes by which the plug is dissolved or removed; as a result water penetrates into the capsule and causes its contents to swell, with the result that the thread is everted violently. A second point of dispute concerns the spot at which thepoison is lodged. Iwanzoy believes it to be contained within the thread itself before discharge, and to be introduced into the tissues of the prey by the eversion of the thread. A third point of dispute is whether the nematocysts are formed in situ, or whether the cnidoblasts migrate with them to the region where they are most needed; the fact that in Hydra, for example, there are no interstitial cells in the tentacles, where nematocysts are very abundant, is certainly in favour of the view that the cnidoblasts migrate on to the tentacles from the body, and that like the genital cells the cnidoblasts are wandering cells.

The muscular tissue consists primarily of processes from the bases of the epithelial cells, processes which are contractile in nature and may be distinctly striated. A FIG. 7. Diagrams~to show the struc- further stage in evolution ture of Nematocysts and their mode of } s that the muscle-cells working. (After Iwanzov.)

a, Undischarged nematocyst b. Commencing discharge. Discharge complete. CnidociL Nucleus of cnidoblast. Outer capsule.

c, en, N, o.c x.

lose their connexion with the epithelium and come to lie entirely beneath it, forming a sub-epithelial contractile layer, developed chiefly in the tentacles of the polyp. The t.c.

Plug closing the opening of the evolution of the ganglionouter capsule. cells is probably similar; Inner capsule, continuous with the an epithelial cell develops wall of the filament, /. processes of nervous nature b Barbs. from the base, which come into connexion with the bases of the sensory cells, with the muscular cells, and with the similar processes of other nerve-cells; next the nerve-cell loses its connexion with the outer epithelium and becomes a sub-epithelial ganglion-cell which is closely connected with the muscular layer, conveying stimuli from the sensory cells to the contractile elements. The ganglion-cells of Hydromedusae are generally very small.

In the polyp the nervous tissue is always in the form of a scattered plexus, never concentrated to form a definite nervous system as in the medusa. The endoderm of the polyp is typically a flagellated epithelium of large cells (fig. 6) , from the bases of which arise contractile muscular processes lying in the plane of the transverse section of the body. In different parts of the coelenteron the endoderm may be of From Gegenbaur's Elements oj Comparative Anatomy.

FIG. 8. Vacuolated Endoderm Cells of cartilaginous __ _ consistence from the axis of the three principal types (i) tentacle of a Medusa (Cunina). digestive endoderm, the primitive type, with cells of large size and considerably vacuolated, found in the hydranth; some of these cells may become special glandular cells, without flagella or contractile processes; (2) circulatory endoderm, without vacuoles and without basal contractile processes, found in the hydrorhiza and hydrocaulus; (3) supporting endoderm (fig. 8), seen in solid tentacles as a row of cubical vacuolated cells, occupying the axis of the tentacle, greatly resembling notochordal tissue, particularly that of Amphioxus at a certain stage of development; as a fourth variety of endodermal cells excretory cells should perhaps be reckoned as seen in the pores in the foot of Hydra and elsewhere (cf. C. Chun HYDROZOA [i], pp. 314, 315).

The mesogloea in the hydropolyp is a thin elastic layer, in which may be lodged the muscular fibres and ganglion cells mentioned abovei jut which never contains any connective tissue or skeletogenous cells or any other kind of special mesogloeal corpuscles.

2. The Polyp-colony. All known hydropolyps possess the power of reproduction by budding, and the buds produced may become either polyps or nedusae. The juds may all become detached after a time and _ive rise to separate and independent individuals, as in the :ommon Hydra, n which only jolyp-individuals are produced and sexual elements F rom Allman's Gymnoblaslic Bydroids, by permission of are developed the Council of the Ray Society.

upon the polyps FIG. 9. Colony of Hydractinia echinata, grow- themselves; or, ing on the Shell of a Whelk. Natural size.

on the other hand, the polyp -individuals produced by budding may remain permanently in connexion with the parent polyp, in which case sexual elements are never developed on polyp-individuals but only on medusa-individuals, and a true colony is formed. Thus the typica.1 hydroid colony starts from a " founder " polyp, which n the vast majority of cases is fixed, but which may be floating, as in Nemopsis, Pelagohydra, etc. The founder-polyp usually produces by rmdding polyp-individuals, and these in their turn produce other Duds. The polyps are all non-sexual individuals whose function s purely nutritive. After a time the polyps, or certain of them, aroduce by budding medusa-individuals, which sooner or later develop sexual elements; in some cases, however, the founder_ polyp remains solitary, that is to say, does not produce polypbuds, but only medusa-buds, from the first (Corymorpha, fig. 3, Myriothela, etc.). In primitive forms the medusa-individuals ; set free before reaching sexual maturity and do not contribute anything to the colony. In other cases, however, the medusa-individuals become sexually mature while still attached to the parent polyp, and are then not set free at all, but become appanages of the hydroid colony and undergo degenerative changes leading to reduction and even to complete obliteration of their original medusan structure. In this way the hydroid colony becomes composed of two portions of different function, the nutritive " trophosomc," composed of non-sexual polyps, and the reproductive " gonosome," composed of sexual medusaindividuals, which never exercise a nutritive function while attached to the colony. As a general rule polyp-buds are produced from the hydrorhiza and hydrocaulus, while medusa-buds are formed on the hydranth. In some cases, however, medusabuds are formed on the hydrorhiza, as in Hydrocorallines.

In such a colony of connected individuals, the exact limits of the separate " persons " are not always clearly marked out. Hence it is necessary to distinguish between, first, the " zooids," indicated in the case of the polyps by the hydranths, each with mouth and tentacles; and, secondly, the "coenosarc," or rh, hydrorhiza. common flesh, which cannot be assigned more to one individual than another, but consists of a more or less complicated network of tubes, corresponding to the hydrocaulus and hydrorhiza of the primitive independent polypindividual. The coenosarc constitutes a system by which the digestive cavity of any one polyp is put into communication with that of any other individual either of the trophosome or gonosome. In this manner the food absorbed by one individual contributes to the welfare of the whole colony, and the coenosarc has the From Allman's Gymnoblaslic Bydroidt, by permission of the Council of the Ray Society.

FIG. 10. Polyps from a Colony of Hydractinia, magnified, dz, dactylozoid; gz, gastrozoid; b, blastostyle; gon, gonophores; function of circulating and distributing nutriment through the colony.

The hydroid colony shows many variations in form and architecture which depend simply upon differences in the methods in which polyps are budded.

In the first place, buds may be produced only from the hydrorhiza, which grows out and branches to form a basal stolon, typically net-like, spreading over the substratum to which the founderpolyp attached itself. From the stolon the daughter-polyps grow up vertically. The result is a spreading or creeping colony, with the coenosarc in the form of a root-like horizontal network (fig. 5, B; n, A). Such a colony may undergo two principal modifications. The meshes of the basal network may become very small or virtually obliterated, so that the coenosarc becomes a crust of tubes tendingtofusetogether, and covered over by a common perisarc. Encrusting colonies of this kind are seen in Clava squamata (fig. 5, A) and Hydractinia (figs. 9, 10), the latter having the perisarc calcified. A further very important modifiAfter Hincks, Forbes, and Browne. A and B modified cation is seen when the from Hincks; C modified from Forbes's Brit. Naked- ^ U.L_ Kocol eyed Medusae.

perisarc do not remain FIG. ii. Lar sabellarum and two stages spread out in one plane, of its Medusa, Willia stellata. A, colony of but grow in all planes Lar; B and C, young and adult medusae. forming a felt-work; the result is a massive colony, such as is seen in the so-called Hydrocorallines (fig. 60), where the interspaces between the coenosarcal tubes are filled up with calcareous matter, or coenosteum, replacing the chitinous perisarc. The result is a stony, solid mass, which contributes to the building up of coral reefs. In massive colonies of this kind no sharp distinction can be drawn between hydrorhiza and hydro- ,. caulus in the coenosarc; it ^7- Tis practically all hydrorhiza. ==r=^" Massive colonies may assume various forms and are often branching or tree-like. A further peculiarity of this type of colony is that the entire coenosarcal complex is covered externally by a common layer of ectoderm; it is not clear how this covering layer is developed.

In the second place, the buds may be produced from the hydrocaulus, growing out laterally from it; the result is an arborescent, tree-like colony (figs. 12, 13). Budding from the hydrocaulus may be combined with budding from the hydrorhiza, so that numerous branching colonies arise from a common basal stolon. In the formation of arborescent colonies, two sharply FIG. 12. Colony of Bougainvillea distinct types of budding are fruticosa, natural size, attached to the found, which are best deunderside of a piece of floating tim- scribed in botanical termmober. (After Allman.) logy as the monopodial or racemose, and the sympodial or cymose types respectively; each is characteristic of one of the two sub-orders of the Hydroidea, the Gymnoblastea and Calyptoblastea.

In the monopodial method (figs. 12, 14) the founder-polyp is, theoretically, of unlimited growth in a vertical direction, and as il grows up it throws out buds right and left alternately, so that the first bud produced by it is the lowest down, the second bud is above the first, the third above this again, and so on. Each bud produced FIG. 13. Portion of colony of Bougainmllea fruticosa ( Anthomedusae-Gymnoblasted) more magnified. (From Lubbock, after Allman.)

by the founder proceeds to grow and to bud in the same way as the founder did, producing a side branch of the main stem. Hence, in a colony of gymnoblastic hydroids, the oldest polyp of each system, that is to say, of the main stem or of a' branch, is the topmost polyp ; FIG. 14. Diagrams of the monopodial method of budding, shown in five stages (1-5). F, the founder-polyp; I, 2, 3, 4, the succession of polyps budded from the founder-polyp; a', b', c' , the succession of polyps budded from I ; a 2 , V, polyps budded from 2 ; a 3 , polyp budded from 3.

the youngest polyp of the system is the one nearest to the topmost polyp ; and the axis of the system is a true axis.

In the sympodial method of budding, on the other hand, the founder-polyp is of limited growth, and forms a bud from its side, which is also of limited growth, and forms a bud in its turn, and so on (figs. 15, 16). Hence, in a colony of calyptoblastic hydroids, the oldest polyp f a system is the lowest ; the youngest polyp is the top- most one; and the axis of the system is a false axis composed of portions of each of the consecutive polyps. In this method of budding there are two types. In one, the biserial type (fig. 1 5), the polyps produce buds right i 3 and left alternately, so that the hydranths are arranged in a zigzag fashion, forming a " scorpioid cyme," as in Obelia and Sertularia. In the other, the uniserial type (fig. 1 6), the buds are formed always on the same side, forming a " heliFIG. 15. Diagram of sympodial budding, coid cyme," as in biserial type, shown in five stages (1-5). F, Hydrallmania, founder-polyp; j, 2, 3, 4, 5, 6, succession of according to H. polyps budded from the founder; a, b, c, Driesch, in which, second series of polyps budded from the founder ; h o w ey e r, the a 3 , b 3 , series budded from 3. primitively uni- serial arrangement becomes masked later by secondary torsions of the hydranths. In a colony formed by sympodial budding, a polyp always produces first a bud, which contributes to the system to which it belongs, i.e. continues the stem or branch of which its parent forms a part. The I polyp may then form a second bud, which becomes the starting ^v point of a new system, the j \ beginning, that is, of a new *>. A ) branch; and even a third bud, starting yet another system, may be produced from the same polyp. Hence the colonies of Calyptoblastea may be complexly branched, and the budding may be biserial throughout, uniserial throughout, or partly one, partly the other. Thus in Plumularidae (figs. 17, 7. , 1 8) there is formed a main stem FIG. 16. Diagram of sympodial by bise rial budding; each polyp budding, uniserial type shown on the main stem forms a in four stages (1-4).. F, founder- second bud which usuall E ol IPliV 2 ' 3, succession of polyps forms a side branch or pinnu i e budded from the founder. by uniserial budding. In this way are formed the familiar feathery colonies of Plumularia, in which the pinnules are all in one plane, while in the allied Antennularia the pinnules are arranged in whorls round the main biserial stem. The pinnules never branch again, since in the uniserial mode of budding a polyp never forms a second polyp-bud. On the other hand, a polyp on the main stem may form a second bud which, instead of forming a pinnule by uniserial budding, produces by biserial budding a branch, from which pinnules arise as from the main stem (fig. 18 3, 6). Or a polyp on the main stem, after having budded a second time to form a pinnule, may give rise to a third bud, which starts a new biserial FIG. 17. Diagram of sympodial budding, system, from which simple unbranched Plumularia-type. F, uniserial pinnules arise founder; 1-8, main axis formed by biserial as from the main stem budding from founder; a-e, pinnule formed type of Aglaophenia by uniserial budding from founder; a l -d { , (fig. 19). The laws of branch formed by similar budding from I ; budding in hydroids a 2 -d? from 2, and so forth. have been worked out in an interesting manner by H. Driesch [13], to whose memoirs the reader must be referred for further details.

Individualization of Polyp-Colonies. As in other cases where animal colonies are formed by organic union of separate individuals, there is ever a tendency for the polyp-colony as a whole to act as a single individual, and for the members to become subordinated to the needs of the colony and to undergo specialization for particular functions, with the result that they simulate organs and their individuality becomes masked to a greater or less degree. Perhaps the earliest of such specializations is connected with the reproductive function. Whereas primitively any polyp in a colony may produce medusa-buds, in many hydroid colonies medusae are budded only by certain polyps termed blastostyles (fig. 10, b). At first not differing in any way from other polyps (fig. 5), the blastostyles gradually lose their nutritive function and the organs connected with it; the mouth and tentacles disappear, and the blastostyle obtains the nutriment necessary for its activity by way of the coenosarc. In the Calyptoblastea, where the polyps are protected by special capsules of the perisarc, the gonothecae enclosing the blastostyles differ from the hydrothecae protecting the hydranths (fig. 54).

In other colonies the two functions of the nutritive polyp, namely, capture and digestion of food, may be shared between different _ polyps (fig. 10). One class /, IG - 18 Diagram showing method of polyps, the dactylozoids of branching in the Plumulana-type; (dz), lose their mouth and compare with fig. 17. Polyps 3 and 6, stomach, and become elon- instead of producing uniserial pinnules, gated and tentacle-like, have produced biserial branches (3 1 , 3', showing great activity of 3 3 - 3 4 : 6 1 ; 6 ),' which give off uniserial movement. Another class, branches in their turn, the gastrozoids (gz), have the tentacles reduced or absent, but have the mouth and stomach enlarged. The dactylozoids capture food and pass it on to the gastrozoids, which swallow and digest it.

Besides the three types of individual above mentioned, there are other appendages of hydroid colonies, of which the individuality is doubtful. Such are the " guard-polyps " (machopolyps) of Plumularidae, which are often regarded as individuals of the nature of dactylozoids, but from a study of the mode of budding in this hydroid family Driesch concluded that the guard-polyps were not true polyp-individuals, although each is enclosed in a small protecting cup of the perisarc, known as a nematophore. Again, the spines arising from the basal crust of Podocoryne have been interpreted by some authors as reduced polyps.

3. The Medusa. In the Hydromedusae the medusa-individual occurs, as already stated, in one of two conditions, either as an independent organism leading a true life the open seas, or as a subordinate individuality in the hydroid colony, from which it is never set free; it then becomes a mere reproductive appendage or gonophore, losing sue- FIG. 19. Diagram showing method of branchcessively its organs ing in the A glaophenia-type. Polyp 7 has proof sense, loco- duced as its first bud, 8; as its second bud, a 7 , motion and nutri- which starts a uniserial pinnule; and as a third tion, until its bud I 7 , which starts a biserial branch (II'-VI 7 ) medusoid nature that repeats the structure of the main stem and and organization gives off pinnules. The main stem is indicated become scarcely by , the new stem by re cognizab 1 e.

Hence it is convenient to consider the morphology of the medusa from these two aspects.

(o) The Medusa as an Independent Organism. The general structure and characteristics of the medusa are described elsewhere (see articles HYDROZOA and MEDUSA), and it is only necessary here to deal with the peculiarities of the Hydromedusa.

As regards habit of life the vast majority of Hydromedusae arc pelagic organisms, floating on the surface of the open sea, propelling themselves feebly by the pumping movements of the umbrella produced by contraction of the sub-umbral musculature, and capturing their prey with their tentacles. The genera Cladonema (fig. 20) and Clavatella (fig. 21), however, are ambulatory, creeping forms, living in rock-pools and walking, as it were, on the tips of the proximal branches of each of the tentacles, while the remaining branches serve for capture of food. Cladonema still has the typical medusan structure, and is able to swim about, but in Clavatella the umbrella is so much reduced that swimming oc.

From Allman's Gymnoblastic Eydroids, by permission of the Council of the Ray Society.

FIG. 2 1 . Clavatella prolifera, ambulatory medusa. /, tentacles; oc t ocelli.

is no longer possible. The remarkable medusa Mnestra />ar<m/e.sisecto- parasitic throughout life From Allman's Gymwblastic Bydroids, by permission of on the pelagic mollusc the Council of the Ray Society. Phyttirrhoe, attached FIG. 20. Cladonema radiatum, the medusa to jt , b y the subwalking on the basal branches of its tentacles umbral surface, and (0, which are turned up over the body. ' ts tentacles have become rudimentary or absent. It is interesting to note that Mnestra has been shown by J. W. Fewkes [15] and R. T. Gunther [19] to belong to the same family ( Cladonemtdae) as Cladonema and Clavatella, and it is reasonable to suppose that the non-parasitic ancestor of Mnestra was, like the other two genera, an ambulatory medusa which acquired louse-like habits. In some species of the genus Cttnma( Narcomedusae) the youngest individuals ( actinulae) are parasitic on other medusae (see below), but in later life the parasitic habit is abandoned. No other instances are known of sessile habit in Hydromedusae.

The external form of the Hydromedusae varies from that of a deep bell or thimble, characteristic of the Anthomedusae, to the shallow saucer-like form characteristic of the Leptomedusae. It is usual for the umbrella to have an even, circular, uninterrupted margin; but in the order Narcomedusae secondary down-growths between the tentacles produce a lobed, indented margin to the umbrella. The marginal tentacles are rarely absent in non-parasitic forms, and are typically four in number, corresponding to the four perradii marked by the radial canals. Interradial tentacles may be also developed, so that the total number present may be increased to eight or to an indefinitely large number. In Willia, Geryonia, etc., however, the tentacles and radial canals are on the plan of six instead of four (figs, ii and 26). On the other hand, in some cases the tentacles are l ess ' n number than the perradii; in Corymorpha (figs. 3 and 22) there is but a single tentacle, while two are found in Amphinema and Gemmaria (An- FIG. 22. Corymorpha thomedusae), and in Solmiindella nut an s, adult female bitentaculata (fig. 67) and Aeginopsis Medusa. Magnified 10 hensenii (fig. 23) (Narcomedusae). The diameters. tentacles also vary considerably in other ways than in number: first, in form, being usually simple, with a basal bulb, but in Cladonemidae they are branched, often in complicated fashion; secondly, in grouping, being usually given off singly, and at regular intervals from the margin of the umbrella, but in Margelidae and in some Trachomedusae they are given off in tufts or bunches (fig. 24) ; After E. T, Browne, from Proc. Zool. Soc. o/ London.

thirdly, in position and origin, being usually implanted on the extreme edge of the umbrella, but in Narcomedusae they become secondarily shifted and are given off high up on the ex-umbrella (figs. 23 and 25) ; and, fourthly, in structure, being hollow or solid, as in the polyp. In some medusae, for instance, the remarkable deep-sea family Pectyllidae, the tentacles may bear suckers, by which the animal may attach itself temporarily. It should be mentioned finally that the tentacles are very contractile and extensible, and may therefore present themselves, in one and the same individual, as long, drawn-out threads, or in the form of short corkscrew - like ringlets ; they may stream downwards from the sub-umbrella, or be held out horizontally, or be directed upwards over the ex-umbrella (fig. 23). Each species of After O. Maas, Die craspedotm Medusen der Plankton Expedition, by permission of Lipsius and Tischer.

FIG. 23. Aeginopsis hensenii, slightly magnified, showing the manner in which the tentacles are carried in life.j After O. Maas, Craspedoten Medusen der SibogaExpeditian, by permission of E. S. Brill & Co.

FIG. 24. Rathkea octonemalis.

characteristic method of carrying its medusa usually has a tentacles.

The sub-umbrella invariably shows a velum as an inwardly projecting ridge or rim at its margin, within the circle of tentacles; hence the medusae of this sub-class are termed craspedote. The manubrium is absent altogether in the fresh-water medusa Limnocnida, in which the diameter of the mouth exceeds half that of the umbrella; on the other hand, the manubrium may attain a great length, owing to the centre of the sub-umbrella with the stomach being drawn into it, as it were, to form a long proboscis, as in Geryonia. The mouth may be a simple, circular pore at the extremity of the manubrium, or by folding of the edges it may become square or shaped like a Maltese cross, with four corners and four lips. The corners of the mouth may then be drawn out into lobes or lappets, which may have a branched or fringed outline (fig. 27), and in Margelidae the subdivisions of the fringe simulate tentacles (fig. 24).

The internal anatomy of the Hydromedusae shows numerous variations. The stomach may be altogether lodged in the manubrium, from which the radial canals then take origin directly as in Geryonia (Trachomedusae); it may be with or without gastric pouches. The radial canals may be simple or branched, primarily four, rarely six in number. The ring-canal is drawn out in Narcomedusae into festoons corresponding with the lobes of the margin, and may be obliterated altogether (Solmaris) . I n this order the radial canals are represented only by wide gastric pouches, and in the family Solmaridae are suppressed altogether, so that the tentacles and the festoons of the ring-canal arise directly from the stomach. In Geryonia, centripetal canals, ending blindly, arise from the ring-canal and run in a radial direction towards the centre of the umbrella (fig. 26).

Histology of the Hydromedusa. The histology described above for the polyp may be taken as the primitive type, from which that After A. Maas, Medusae, in Prince of Monaco's" FIG. 25. Aeginura grimaldii, natural size.

of the medusa differs only in greater elaboration and differentiation of the cell-elements, which are also more concentrated to form distinct tissues.

The ectoderm furnishes the general epithelial covering of the body, and the muscular tissue, nervous system and sense-organs. The FIG. 26. Carmarina (Geryonia) (After Nerve ring.

Radial nerve.

Tentaculocyst.

Circular canal.

Radiating canal.

Ovary.

Peronia or cartilaginous process ascending from the cartilaginous margin of the disk centripetally in the outer surface of the jellylike disk; six of these are perradial, six interradial, corresponding to the twelve hastata, one of the Trachomedusae. Haeckel.)

solid larval tentacles, resembling those of Cunina.

k, Dilatation (stomach) of the manubrium.

/, Jelly of the disk.

p, Manubrium.

/. Tentacle (hollow and tertiary, i.e. preceded by six perradial and six interradial solid larval tentacles).

, Cartilaginous margin of the disk covered by threadcells.

r. Velum.

external epithelium is flat on the ex-umbral surface, more columnar on the sub-umbral surface, where it forms the muscular tissue of the sub-umbrella and the velum. The nematocysts of the ectoderm may be grouped to form batteries on the tentacles, umbrellar margin and oral lappets. In places the nematocysts may be crowded so thickly as to form a tough, supporting, " chondral " tissue, resembling cartilage, chiefly developed at the margin of the umbrella and forming i streaks or bars supporting the tentacles (" Tentakelspangen, ' peronia) or the tentaculocysts ( ' Gehorspangen," otoporpae). The muscular tissue of the Hydromedusae is entirely ectodermal. The muscle-fibres arise as processes from the bases of the epithelial cells; such cells may individually become sub-epithelial in position, as in the polyp; or, in places where muscular tissue is greatly developed, as in the velum or sub-umbrella, the entire muscular epithelium may be thrown into folds in order to increase its surface, so that a deeper sub-epithelial muscular layer becomes separated completely from a more superficial bodyof^thelium.

After O. Maas fa KMB/JJ o/ j n ; ts arrangement the muscular tissue Museum "of" Com^arS fo . rms . two systems: the one composed Zoology, Cambridge, Mass., of striated fibres arranged circularly, that is to say, concentrically round the central FIG. 27.Stomoloca ax j s O f t he umbrella; the other of nondtftta.oneoftheTjarttZae striated fibres running longitudinally, (Anthomedusae). t h a t is to say, in a radial direction from, or (in the manubrium) parallel to, the same ideal axis. The circular system is developed continuously over the entire subumbral surface, and the velum represents a special local development of this system, at a region where it is able to act at the greatest mechanical advantage in producing the contractions of the umbrella by which the animal progresses. The longitudinal system is discontinuous, and is subdivided into proximal, medial and distal portions. The proximal portion forms the retractor muscles of the manubrium, or proboscis, well developed, for example, in Geryonia. The medial portion forms radiating tracts of fibres, the so-called " bell-muscles " running underneath, and parallel to, the radial canals; when greatly developed, as in Tiandae, they form ridges, so-called, mesenteries, projecting into the sub-umbral cavity. The distal portions form the muscles of the tentacles. In contrast with the polyp, the longitudinal muscle-system is entirely ectodermal, there being no endodermal muscles in craspedote medusae.

The nervous system of the medusa consists of sub-epithelial ganglion-cells, which form, in the first place, a diffuse plexus of nervous tissue, as in the polyp, but developed chieny on the subumbral surface; and which are concentrated, in the second place, to form a definite central nervous system, never found in the polyp. In Hydromedusae the central nervous system forms two Flo. 28. Muscular Cells of Medusae concentric nerve- (Lizzia). The uppermost is a purely muscular rings at the margin ce ll f rom tne sub-umbrella; the two lower are of the umbrella, near epidermo-muscular cells from the base of a thebaseof the velum, tentacle; the upstanding nucleated portion One, the upper forms part of the epidermal mosaic on the or ex-umbral nerve- f ree sur f ace o f t h e body. (After Hertwig.) ring, is derived irom the ectoderm on the ex-umbral side of the velum; it is the larger of the two rings, containing more numerous but smaller ganglioncells, and innervates the tentacles. The other, the " lower or subumbral nerve-ring, is derived from the ectoderm on the sub-umbral side of the velum; it contains fewer but larger ganglion-cells and innervates the muscles of the velum (see diagram in article MEDUSAE). The two nerve-rings are connected by fibres passing from one to the other.

The sensory cells are slender epithelial cells, often with a cilium or stiff protoplasmic process, and should perhaps be regarded as the only ectoderm-cells which retain the primitive dilation of the larval ectoderm, otherwise lost in all Hydrozoa. The sense-cells form, in the first place, a diffuse system of scattered sensory cells, as in the polyp, developed chiefly on the manubrium, the tentacles and the margin of the umbrella, where they form a sensory ciliated epithelium covering the nerve-centres; in the second place, the sense-cells are concentrated to form definite sense-organs, situated always at the margin of the umbrella, hence often termed " marginal bodies." The possession of definite senseorgans at once distinguishes the medusa from the polyp, in which they are never found.

The sense-organs of medusae are of two kinds first, organs sensitive to light, usually termed ocelli (fig. 29) ; secondly, organs commonly termed otocysts, on account of their resemblance to the auditory vesicles of higher ^ o Maas Crasfedalen Medusm der sib animals, but serving Expedition, by permission of E. S. Brill & Co. for the s e n s c o f Fl( , 2f) Tiaropsis rosea (Ag. and Mayer) balance and oncnta- showing y thc eight adradial Statocysts, each close to an Ocellus. Cf. fig. 30.

of statocysts (fig. 30). The sense-organs may be tentaculocysts, i.e. modifications of a tentacle, as in Trachylinae, or developed from the margin of the umbrella, in no connexion with a tentacle (or, if so connected, not producing any modification in the tentacle), as in Leptolinae. In Hydromedusae the sense-organs are always exposed at the umbrellar margin (hence Gymnophthalmata), while in Scyphpmedusae they are covered over by flaps of the umbrellar margin (hence Steganophthalmata).

The statocysts present in general the structure of cither a knob or a closed vesicle, composed of (i) indifferent supporting epithelium ; (2) sensory, so-called auditory epithelium of slender cells, each* bearing at its free upper end a stiff bristle and running out at its base into a nerve-fibre ; (3) concrement-cells, which produce intercellular concretions, so-called otoliths. By vibrations -Sub means of or shocks transmitted through the water, or by displacements in the balance or position of the animal, the otoliths are caused to impinge against the bristles of the sensory cells, now on one side, now on the other, causing shocks or stimuli which are transmitted by the basaJ nerve-fibre to the central nervous system. Two stages in the development of the otocyst Modified after Linko, Tramux Soc. Imp. Nat., St. can fo e recognized the Petersburg, BB. first that of , Qpe ' n pi( .

FIG. 30. Section of a Statocyst and on a freely - projecting Ocellus of Tiaropsis diademata ; cf . fig. 29. knob, in which the otoex, Ex-umbral ectoderm. sub, Sub-umbral ectoderm. c.c, Circular canal. v, Velum.

st.c, Cavity of statocyst. con, Concrement-cell with otolith.

St.e liths are exposed, the second that of a closed vesicle, in which the otoliths are covered over. Further, two distinct types of otocyst can be recognized in the Hydromedusae; that of the Leptolinae, in which the entire organ is ectodermal, concrement-cells and all, and the organ is not a tentaculocyst; and that of the Trachylinae, in which the organ is a tentaculocyst, and the concrement-cells are endodermal, derived from the endoderm of the modified tentacle, while the rest of the organ is ectodermal.

In the Leptolinae the otocysts are seen in their first stage in Mitrocoma annae (fig. 31) and Tiaropsis (figs. 29, 30) as an open pit at the base of the velum, on its subumbral side. The pit has its opening turned towards the sub-umbral cavity, while its , ..,,. base or fundus forms a bulge, more or less pronounced, on Nemn- the ex-umbral side of the by velum. At the fundus are placed the concrement-cells FIG. 31. Section of a Statocyst of with their conspicuous oto- Modified after O. and R. Hertwij system und Sinnesorgane der Medusen permission of F. C. W. Vogel.

Mitrocoma annae. sub, Sub-umbral ectoderm c.c, Circular canal. , Velum.

st.c, Cavity of statocyst. con, Concrement-cell with otolith.

liths (con) and the inconspicuous auditory cells, which are connected with the subumbral nerve -ring. From the open condition arises the closed condition very simply by closing up of the aperture of the pit. We then find the typical otocyst of the Leptomedusae, a vesicle bulging on the ex-umbral side of the velum (figs. 32, 33). The otocysts are placed on the outer wall of the .: ex.

Modified after O. and R. Hertwig, Nenensyslcm und Smnesorgane der Medusen, by permission of F. C. W. Vogel.

FIG. 32. Section of a Statocyst of Phialidium. ex, Ex-umbral ectoderm. .sub, Sub-umbral ectoderm. v, Velum.

st.c, Cavity of statocyst. con, Concrement-cell with otolith.

st.c.

Modified after 0. and R. Hertwig, Nervensyslem und Sinnesorgano der Medusen, by permission of F. C. W. Vogd.

FIG. 33. Optical Section of a Statocyst of Octorchis.

con, Concrement - cell with otolith. st.c, Cavity of statocyst.

vesicle (the fundus of the original pit) or on its sides; their arrangement and number vary greatly and furnish useful characters for distinguishing genera. The sense-cells are innervated, as before, from the sub-umbral nerve-ring. The inner wall of the vesicle con.

(region of closure) is frequently thickened to form a so-called " sensecushion," apparently a ganglionic offshoot from the sub-umbral nerve -ring. In many Leptomedusae the otocysts are very small, inconspicuous and embedded completely in the tissues; hence they may be easily overlooked in badly-preserved material, and perhaps are present in many cases where they fnd.- have been said to have been wanting.

In the Trachylinae the n.C.

simplest condition of the otocyst is a freely projecting club, a so-called statorhabd (figs. 34, 35), representing a tentacle greatly reduced in size, ,, s - ,_. ^wns.! >iiu?mvr '' " ""* covered with sensory ^Uf^m&Sn''"'''*" ectodermal epithelium (eft 1 and rnnt-aini'no- on ^tei O. and R. Hertwig, Nenensystem und Sinnes- A A cmitammg an organe da Medusen, by permission of F. C. W.

endodermal core (end.), Vogel.

which is at first continu- F IG. 34. Tentaculocyst (statorhabd) ous with the endoderm o f Cunina solmaris. n.c, Nerve-cushion; of the ring-canal, but en a, endodermal concrement-cells; con, later becomes separated otolith. from it. In the endoderm large concretions are formed (con.). Other sensory cells with long cilia cover a sort of cushion (n.c.) at the base of the club ; the club may be long and the cushion small, or the cushion large and the club small. The whole structure is innervated, like the tentacles, from the ex-umbral nerve-ring. An advance towards the second stage is seen in such a form as Rhopalonema (fig. 36), where the ectoderm of the cushion rises up in a double fold to enclose the club in a protective covering forming a cup or vesicle, at first open distally ; finally the opening closes and the closed vesicle may sink inwards and be found far removed from After 0. and R. Hertwig, Nenensystem und Sinnesthe surface, as in Geryonia "'Sane der Medusen, by permission of F. C. W.

(fig. 37). V sel - The ocelli are seen in FIG. 35. Tentaculocyst of Cunina latitheir simplest form as a ventris.

pigmented patch of ecto- ect, Ectoderm, derm, which consists of n.c, Nerve-cushion, two kinds of cells (l) end, Endodermal concrement-cells. pigment-cells, which are con, Otolith. ordinary indifferent cells of the epithelium containing pigment-granules, and (2) visual cells, slender sensory epithelial cells of the usual type, which may develop visual cones or rods at their free extremity. The ocelli occur usually either on the inner or outer sides of the tentacles; if on the inner side, the tentacle is turned upwards and carried over the ex - umbrella, so as to expose the ocellus to the light; if the -end.

-hit FIG. 36. Simple tentaculocyst of Rhopalo' ocellus be on the nema velatum. The process carrying the otolith outer side of a or concretion hk, formed by endoderm cells, is tentacle, two enclosed by an upgrowth forming the " vesicle," nerves run round which is not yet quite closed in at the top. the base of the (After Hertwig.) tentacle to it. In other cases ocelli may occur between tentacles, as in Tiaropsis (fig. 29).

The simple form of ocellus described in the foregoing paragraph may become folded into a pit or cup, the interior of which becomes filled with a clear gelatinous secretion forming a sort of vitreous ex, sub, c.c, v, body. The distal portion of the vitreous body may project from rhe cavity of the cup, forming a non-cellular lens as in Lizzia (fig. 28). Beyond this simple condition the visual organs of the Hydromedusae do not advance, and are far from reaching the wonderful development of the eyes of Scyphomedusae (Charybdaea) .

Besides the ordinary type of ocellus just described, there is found in one genus(Tiaropsis) a type of ocellus in which the visual elements are inverted, and s '- _ have their cones turned away from the light, as in the human retina (fig. 30). In this case the pigment-cells are endodermal, forming a cup of pigment in which the visual cones are embedded. A similar ocellus is formed in Aurelia among the Scyphomedusae (g.f.).

After O. and R. Hertwig, ffenensystem and Sinnesors/me Other sense ier Hcdusm, by permission of F. C. W. Vogel. organs of Hydro- FIG. 37. Section of statocyst of Geryonia medusae are the (Carmarina hastata). so-called sense- st.c, Statocyst containing the minute tentaculo- clubs or cordyli cyst, found in a few nr jt Ex-umbral nerve-ring. Leptomedusae, nr,, Sub-umbral nerve-ring. especially in those Ex-umbral ectoderm. genera, in which Sub-umbral ectoderm. otocysts are mcon- Circular canal. spicuous or absent Velum. (fig- 39)-. Each cordylus is a tentacle-like structure with an endodermal axis containing an axial cavity which may be continuous with the ring-canal, or may be partially occluded. Externally the cordylus is covered by very flattened ectoderm, and bears no otoliths or sense-cells, but the base of the club rests upon the ex-umbral nerve-ring. Brooks regards these organs as sensory, serving for the sense of balance, and representing a primitive stage of the tentaculocysts of Trachylinae; Linko, on the other hand! finding no nerve-elements connected with them, regards them as digestive (?) in function.

The sense-organs of the two fresh-water medusae Limnocodium and Limnocnida are peculiar and of rather doubtful nature (see E. T. Browne [10]).

The endoderm of the medusa shows the same general types of structure as in the polyp, described above. We can distinguish (i) digestive endoderm, in the stomach, often with special glandular elements; (2) circulatory endoderm, in the radial and ringcanals; (3) supporting endoderm in the axes of the tentacles and in the endodermlamella; the latter is primitively a double layer of cells, produced by concrescence of the ex-umbral and sub-umbral layers of the coelenteron, but it is usually found as a single layer of flattened cells (fig. 40); in Geryonia, however, it remains double, and the centripetal canals arise by parting of the two layers; (4) excretory endoderm, lining pores at the margin of the umbrella, occurring in certain Leptomedusae as socalled " marginal tubercles," opening, on the one hand, into the ring-canal and, on the other hand, to the exterior by " marginal funnels," which debouch into the sub-umbral cavity above the velum. As has been deFIG. 38. Ocellus of scribed above, the endoderm may also conLizzia koellikeri. oc, tribute to the sense-organs, but such Pigmented ectodermal contributions are always of an accessory cefis; /, lens. (After nature, for instance, concrement-cells in Hertwig.) the otocysts, pigment in the ocelli, and never of sensory nature, sense-cells being in all cases ectodermal.

The reproductive cells may be regarded as belonging primarily to neither ectoderm nor endoderm, though lodged in the ectoderm in all Hydromedusae. As described for the polyp, they are wandering cells capable of extensive migrations before reaching the particular spot at which they ripen. In the Hydromedusae they usually, if not invariably, ripen in the ectoderm, but in the neighbourhood of the main sources of nutriment, that is to say, not far from the stomach. Hence the gonads are found on the manubrium in Anthomedusae generally; on the base of the manubrium, or under the gastral pouches, or in both these situations (Octorchidae), or under the radial canals, in Trachomedusae; under the gastral pouches or radial canals, in Narcomedusae. When ripe, the germ-cells are dehisced directly to the exterior.

Hydromedusae are of separate sexes, the only known exception being Amphogona apsteini, one of the Trachomedusae (Browne [9]). Moreover, all the medusae budded from a given hydroid colony are either male or female, so that even the non-sexual polyp must be considered to have a latent sex. (In Hydra, on the other hand, the individual is usually hermaphrodite.) The medusa always reproduces itself sexually, and in some cases non-sexually also. The non-sexual reproduction takes the form of fission, budding or sporogony, the details of which are described below. Buds may be produced from the manubrium, radial canals, ring-canal, or tentaclebases, or from an aboral stolon (Narcomedasae). In all cases only medusabuds are produced, never polyp-buds.

The- mesogloea of the medusa is largely developed and of great thickness in the umbrella. The sub-epithelial tissues, i.e. the nervous and muscular ceHs, are lodged in the mesogloea, but in Hydromedusae it never contains tissue-cells or After w |K Brooks Jmrnai o{ Morphology, x., mesogloeal corpuscles. by permission of Ginn & Co.

FIG. 39. Section of a Cordylus of Laodice.

Circular canal.

Velum.

Tentacle.

Cordylus, composed of flattened ectoderm ec covering a large-celled endodermal axis en.

(6) The Medusae as a Subordinate Individuality. It has been shown above that polyps are budded only from polyps and that the medusae may be budded either from polyps or from medusae. I n any case the daughter-individuals produced from the buds may be imagined as remaining attached to the parent and forming a colony of individuals in organic connexion with one another, and thus three possible cases arise. The first case gives a colony entirely composed of polyps, as in many Hydroidea. The second case gives a colony partly composed of polyp-individuals, partly of medusa-individuals, a possibility also realized in many colonies of Hydroidea. The third case gives a colony entirely composed of medusa-individuals, a possibility perhaps realized in the Siphonophora, which will be discussed in dealing with this group.

The first step towards the formation of a mixed hydroid colony is undoubtedly a hastening of the sexual maturity of the medusaindividual. Normally the medusae are liberated in quite an immature state; they swim away, feed, grow and become adult mature individuals. From the bionomical point of view, the medusa is to be considered as a means of spreading the species, supplementing the deficiencies of the sessile polyp. It may be, however, that increased reproductiveness becomes of greater importance to the species than wide diffusion; such a condition FIG. 40. Portions of Sections through will be brought about if the Disk of Medusae the upper one of the medusae mature Lizzia, the lower of Aurelia. (After quickly and are either Hertwig.) set free in a mature d Endoderm i ame ll a . condition or remain in Muscular processes of the ectoderm-cells the shelter of the polyp- in croi . Brt fc n .

Ectoderm.

Endoderm lining the enteric cavity. Wandering endoderm cells of the gelatinous substance.

colony, protected^ from j risks of a free life in the ' open sea. In this way ' the medusa sinks from ' an independent personality to an organ of the polyp-colony, becoming a so-called medusoid gonophore, or bearer of the reproductive organs, and losing gradually all organs necessary for an independent existence, namely those of sense, locomotion and nutrition.

In some cases both free medusae and gonophpres may be produced from the same hydroid colony. This is the case in Syncoryne mirabilis (Allman [1], p. 378) and in Campanularia volubilis; in the latter, free medusae are produced in summer, gonophores in winter (Duplessis [14]). Again in Pennaria, the male medusae are set free in a state of maturity, and have ocelli ; the female medusae remain attached and have no sense organs.

The gonophores of different hydroids differ greatly in structure from one another, and form a series showing degeneration of the medusa-individual, which is gradually stripped, as it were, of its characteristic features of medusan organization and finally reduced to the simplest structure. A very early stage in the degeneration is well exemplified by the so-called " meconidium " of Gonothyraea (fig. 41, A). Here the medusoid, attached by the centre of its ex-umbral surface, has lost its velum and sub-umbral muscles, its sense organs and mouth, though still retaining rudimentary tentacles. The gonads (g) are produced on the manubrium, which has a hollow endodermal axis, termed the spadix (sp.), in open communication with the coenosarc of the polyp-colony and serving for the nutrition of the generative cells. A very similar condition is seen in Tubularia (fig. 41, B), where, however, the tentacles have quite disappeared, and the circular rim formed by the margin of the umbrella has nearly closed over the manubrium leaving only a small aperture through which the embryos emerge. The next step is illustrated by the female gonophores of Cladocoryne, where the radial and ring-canals Modified from Weismaan, Entstehung der Sexucdzellen bei den Hydromedusen. FIG. 41. Diagrams of the Structure of the Gonophores various Hydromedusae, based on the figures of G. J. Allman and A. Weismann.

A, "Meconidium" of Gonothyraea. H, Withspadix branched (Cordy- B, Type of Tubularia.

C, Type of Garveia, etc. [etc.

D, Type of Plumularia, Agalma, E, Type of Coryne, Forskalia, etc.

F, G, H, Sporosacs.

F, With simple spadix.

G, With spadix prolonged (Eudendrium).

lophora) . s.c, Sub-umbral cavity. /, Tentacles. c.c, Circular canal, g, Gonads. sp, Spadix.

e.l, Endoderm-lamella. ex, Ex-umbral ectoderm. ect, Ectotheca.

have become obliterated by coalescence of their walls, so that the entire endoderm of the umbrella is in the condition of the endodermlamella. Next the opening of the umbrella closes up completely and disappears, so that the sub-umbral cavity forms a closed space surrounding the manubrium, on which the gonads are developed; such a condition is seen in the male gonophore of Cladocoryne and in Garveia (fig. 41, C), where, however, there is a further complication in the form of an adventitious envelope or ectotheca (ect.) split off from the gonophore as a protective covering, and not present in Cladocoryne. The sub-umbral cavity (s.c.) functions as a brood-space for the developing embryos, which are set free by rupture of the wall. It is evident that the outer envelope of the gonophore represents the e-umbral ectoderm (ex.), and that the inner ectoderm lining the cavity represents the sub-umbral ectoderm of the free medusa. The next step is the gradual obliteration of the sub-umbral cavity (s.c.) by disappearance of which the sub-umbral ectoderm comes into contact with the ectoderm of the manubrium. Such a type is found in Plumularia and also in Agalma (fig. 41, D) ; centrally is seen the spadix (sp.), bearing the generative cells (g), and external to these (i) a layer of ectoderm representing the epithelium of the manubrium ; (2) the layer of sub-umbral ectoderm; (3) the endoderm-lamella (e.l.) ; (4) the ex-umbral ectoderm (ex.) ; and (5) there may or may not be present also an ectotheca. Thus the gonads are covered over by at least four layers of epithelium, and since these are unnecessary, presenting merely obstacles to the dehiscence of the gonads, they gradually undergo reduction. The sub-umbral ectoderm and that covering the manubrium undergo concrescence to form a single layer (fig. 41, E), which finally disappears altogether, and the endodermlamella disappears. The gonophore is now reduced to its simplest condition, known as the sporosac (fig. 41, F, G, H), and consists of the spadix bearing the gonads covered by a single layer of ectoderm (ex.), with or without the addition of an ectotheca. It cannot be too strongly emphasized, however, that the sporosac should not be compared simply with the manubrium of the medusa, as is sometimes done. The endodermal spadix (sp ) of the sporosac represents the endoderm of the manubrium; the ectodermal lining of the sporosac (ex.) represents the ex-umbral ectoderm of the medusa; and the intervening layers, together with the sub-umbral cavity, have disappeared. The spadix, as the organ of nutrition for the gonads, may be developed in various ways, being simple (fig. 41, F) or branched (fig. 41, H); in Eudendrium (fig. 41, G) it curls round the single large ovum.

The hydroid Dicoryne is remarkable for the possession of gonophores, which are ciliate and become detached and swim away by means of their cilia. Each such sporosac has two long tentacle-like processes thickly ciliated.

It has been maintained that the gonads of Hydra represent sporosacs or gonophores greatly reduced, with the last traces of medusoid Struc- ^te r Allman, Gymnoblastic Hydroids, by permission of the ture completely CouncU o the ^ Sodety ~ obliterated. There FIG. 42. Gonophores of Dicoryne conferta. is, however, no A, A male gonophore still enclosed in its ectoevidence whatever theca.' [liberation, for this, the gonads B and C, Two views of a female gonophore after of Hydra being t, Tentacles.

purely ectodermal ov, Ova, two carried on each female gonophore. structures, while sp, Testis. all medusoid gonophores have an endodermal portion. Hydra is, moreover, bisexual, in contrast with what is known of hydroid colonies.

In some Leptomedusae the gonads are formed on the radial canals and form protruding masses resembling sporosacs superficially, but not in structure. Allman, however, regarded this type of gonad as equivalent to a sporosac, and considered the medusa bearing them as a non-sexual organism, a " blastocheme " as he termed it, producing by budding medusoid gonophores. As medusae are known to bud medusae from the radial canals there is nothing impossible in Allman's theory, but it cannot be said to have; received satisfactory proof.

Reproduction and Ontogeny of the Hydromedusae.

Nearly every possible method of reproduction occurs amongst the Hydromedusae. In classifying methods of generation it is usual to make use of the sexual or non-sexual nature of the reproduction as a primary difference, but a more scientific classification is afforded by the distinction between tissue-cells (histocytes) and germinal cells, actual or potential ( archaeocytes), amongst the constituent cells of the animal body. In this way we may distinguish, first, vegetative reproduction, the result of discontinuous growth of the tissues and cell-layers of the body as a whole, leading to (i) fission, (2) autotomy, or (3) vegetative budding; secondly, germinal reproduction, the result of the reproductive activity of the archaeocytes or germinal tissue. In germinal reproduction the proliferating cells may be undi/erentiated, so-called primitive germ-cells, or they may be differentiated as sexual cells, male or female, i.e. spermatozoa and ova. If the germ-cells are undifierentiated, the offspring may arise from many cells or from a single cell; the first type is (4) germinal budding, the second is (5) sporogony. If the germcells are differentiated, the offspring arises by syngamy or sexual union of the ordinary type between an ovum and spermatozoon, so-called fertilization of the ovum, or by parthenogenesis, i.e. development of an ovum without fertilization. The only one of these possible modes of reproduction not known to occur in Hydromedusae is parthenogenesis.

(1) True fission or longitudinal division of an individual into two equal and similar daughter-individuals is not common but occurs in Gastroblasta, where it has been described in detail by Arnold Lang [30].

(2) Autotomy, sometimes termed transverse fission, is the name given to a process of unequal fission in which a portion of the body separates off with subsequent regeneration. In Tubularia by a process of decapitation the hydranths may separate off and give rise to a separate individual, while the remainder of the body grows a new hydranth. Similarly in Schizocladium portions of the hydrocaulus are cut off to form so-called " spores," which grow into new individuals (see Allman [1]).

(3) Vegetative budding is almost universal in the Hydromedusae. By budding is understood the formation of a new individual from a fresh growth of undifferentiated material. It is convenient to distinguish buds that give rise to polyps from those that form medusae.

(a) The Polyp. The buds that form polyps are very simple in mode of formation. Four stages may be distinguished; the first is a simple outgrowth of both layers, ectoderm and endoderm containing a prolongation of the coelenteric cavity; in the seconc stage the tentacles "Coelenterata," in grow put as secondary diverticula from the side of the first outgrowth; in the thirc stage the mouth is formed as a perfora tion of the two layers, and, lastly, if the but S.o.

Much modified from C. Chun, Bronn's Tieneich.

FIG. 43. Direct Budding of Cunina. A, B, C, E, F, In ver- t, Tentacle.

tical section.

D, Sketch of external view.

st. Stomach.

m, Manubrium.

s.o, v, s.c, n.s, Sense organ. Velum.

Sub-um bral cavity. Nervous system.

from the parent polyp and begins a free existence.

(6) The Medusae. Two types of budding must be distinguished the direct, so-called palingenetic type, and the indirect, so-callet coenogenotic type.

The direct type of budding is rare, but is seen in Cunina anc Millepora. In Cunina there arises, first, a simple outgrowth of both layers, as in a polyp-bud (fig. 43, A) ; in this the mouth is formec distally as a perforation (B) ; next the sides of the tube so formec >ulge out laterally near the attachment to form the umbrella, while :he distal undilated portion of the tube represents the manubrium ^C) ; the umbrella now grows out nto a number of obes or lappets, and the tentacles and tentaculocysts *row out, the ormer in a notch jet ween two appets, the latter on the apex of each lappet (D, E); finally, the velum arises as a growth of the ectoderm alone, the whole bud shapes itself, so to speak, and the little medusa is separated off by rupture of the thin stalk connecting it with the parent (F). The direct method of medusa-budding only differs from the polyp-bud by its greater complexity of parts and organs.

The indirect mode of budding (figs. 44, 45) is the commonest method by which medusabuds are formed. It is marked by the formation in the bud of a characteristic structure termed the entocodon ( Knospenkern, Glockenkern).

The first stage is a simple hollow outgrowth of both body-layers (fig. 44, A); at the tip of this is formed a thickening of the ectoderm, arising Krimitively as a ollow ingrowth (fig. 44, B), but more usually as a solid mass of ectoderm-cells (fig. 45, A). The ectodermal ingrowth is the entocodon (Gc.) ; it bulges into, and pushes down, the endoderm at the apex of the bud, and if solid it soon acquires a cavity (fig. 44, C, s.c.). The cavity of the entocodon increases continually in size, while the endoderm pushes up at the sides of it to form a cup with hollow walls, enclosing but not quite surrounding the _ entocodon, which - remains in contact at its outer side with the ectoderm covering the bud (fig. 44, D, t). The next changes that take place are chiefly in the endoderm-cup (fig. 44, D, E) ; the cavity between the two walls of the cup becomes reduced by concrescence to form the radial canals (r.c.), ring-canal (c.c.), and endoderm-lamella (e.l., fig. 44, E), and at the same time the base of the cup is thrust upwards to form the manubrium (m), converting the cavity of the entocodon into a FIG. 44. Diagrams of Medusa budding with the formation of an entocodon. The endoderm is shaded, the ectoderm left clear.

A, B, C, D, F, Successive stages in vertical section. E, Transverse section of a stage similar to D. Gc, Entocodon. s.c, Cavity of entocodon, forming the future sub-u m b r a 1 cavity. st, Stomach. r.c. Radial canal. c.c, Circular canal. e.l, Endoderm lamella. m, Manubrium. , Velum. t, Tentacle.

A B C FIG. 45. Modifications of the method of budding shown in fig. 44, with solid Entocodon (Gc.) and formation of an ectotheca (ect.).

space which is crescentic or horse-shoe-like in section. Next tentacles (t, fig. 44, F) grow out from the ring-canal, and the double plate of ectoderm on the distal side of the entocodon becomes perforated, leaving a circular rim composed of two layers of ectoderm, the velum (v) of the medusa. Finally, a mouth is formed by breaking through at the apex of the manubrium, and the now fully-formed medusa becomes separated by rupture of the stalk of the bud and swims away.

If the bud, how ever, is destined to give rise not to a free medusa, but to a gonophore the developmenl is similar buc becomes arrested at various points, according to the degree to which the gonophore i: degenerate. The entocodon is usually formed proving the medusoid nature of the bud, but in sporosacs the entocodon may be rudimentary or absent altogether. The process of budding as above described may be varied or complicated in various ways; thus a secondary, amnion-Iike, ectodermal covering or ectotheca (fig. 45, C, ect.) may be formed over all, as in Garveia, etc. ; or the entocodon may remain solid and without cavity until after the formation of the manubrium, or may never acquire a cavity at all, as described above for the gonophores.

Phylogenetic Significance of the Entocodon. It is seen from the foregoing account FIG. 46. Diagrams to show the significance of the Entocodon in Medusa-buds. (Modified from a diagram given by A. Weismann.)

Ideally primitive method of budding, in of medusa - budwhich the mouth is formed first (la), ding that the entonext the tentacles (16), and lastly the codon is a umbrella.

very important consti- Method of Cunina; (a) the mouth arises, tuent of the bud, next the umbrella (b), and lastly the ten- furnishing some of tacles (c). the most essential , Hypothetical transition from II to the in- portions of the direct method with an entocodon; the medusa ; its cavity formation of the manubrium is retarded, becomes the subthat of the umbrella hastened (Ilia, b). umbral cavity, IV, a, b, c, budding with an entocodon (cf. and its lining fur- fig- 44)- nishes the ecto- V, Budding with a solid entocodon (cf. fig. 45). dermal epithelium of the manubrium and of the sub-umbral cavity as far as the edge of the velum. Hence the entocodon represents a precocious formation of the sub-umbral surface, equivalent to the peristome of the polyp, differentiated in the bud prior to other portions of the organism which must be regarded as antecedent to it in phylogeny.

If the three principal organ-systems of the medusa, namely mouth, tentacles and umbrella, be considered in the light of phylogeny, it is evident that the manubrium bearing the mouth must be the oldest, as representing a common property of all the Coelentera, even of the gastrula embryo of all Enterozoa. Next in order come the tentacles, common to all Cnidaria. The special property of the medusa is the umbrella, distinguishing the medusa at once from other morphological types among the Coelentera. If, therefore, the formation of these three systems of organs took place according to a strictly phylogenetic sequence, we should expect them to appear in the order set forth above (fig. 46, la, b, c). The nearest approach to the phylogenetic sequence is seen in the budding of Cunina, where the manubrium and mouth appear first, but the umbrella is formed before the tentacles (fig. 46, Ha, b, c). In the indirect or coenogenetic method of budding, the first two members of the sequence exhibited by Cunina change places, and the umbrella is formed first, the manubrium next, and then the tentacles; the actual mouthperforation being delayed to the very last (fig. 46, IVa, b, c). Hence the budding of medusae exemplifies very clearly a common phenomenon in development, a phylogenetic series of events completely dislocated in the ontogenetic time-sequence.

The entocodon is to be regarded, therefore, not as primarily an ingrowth of ectoderm, but rather as an upgrowth of both bodylayers, in the form of a circular rim (IVo), representing the umbrellar margin ; it is comparable to the bulging that forms the umbrella in the direct method of budding, but takes place before a manubrium is formed, and is greatly reduced in size, so as to become a little pit. By a simple modification, the open pit becomes a solid ectodermal ingrowth, just as in Teleostean fishes the hollow medullary tube, or the auditory pit of other vertebrate embryos, is formed at first as a solid cord of cells, which acquires a cavity secondarily. Moreover, the entocodon, however developed, gives rise at first to a closed cavity, representing a closing over of the umbrella, temporary in the bud destined to be a free medusa, but usually permanent in the sessile gonophore. As has been shown above, the closing up of the sub-umbral cavity is one of the earliest degenerative changes in the evolution of the gonophore, and we may regard it as the umbrellar fold taking on a protective function, either temporarily for the bud or permanently for the gonophore.

To sum up, the entocodon is a precocious formation of the umbrella, closing over to protect the organs in the umbrellar cavity. The possession of an entocodon proves the medusa-nature of the bud, and can only be explained on the theory that gonophores are degenerate medusae, and is inexplicable on the opposed view that medusae are derived from gonophores secondarily set free. In the sporosac, however, the medusa-individual has become so degenerate that even the documentary proof, so to speak, of its medusoid nature may have been destroyed, and only circumstantial evidence of its nature can be produced.

4. Germinal Budding. This method of budding is commonly described as budding from a single body-layer, instead of from both layers. The layer that produces the bud is invariably the ectoderm, i.e. the layer in which, in Hydromedusae, the generative cells are lodged; and in some cases the buds are produced in the exact spot in which later the gonads appear. From these facts, and from those of the sporogony, to be described below, we may regard budding to this type as taking place from the germinal epithelium rather than from ordinary ectoderm.

(a) The Polyp. Budding from the ectoderm alone has been described by A. Lang [29] in Hydra and other polyps. The tissues of the bud become differentiated into ectoderm and endoderm, and the endoderm of the bud becomes secondarily continuous with that of the parent, but no part of the parental endoderm contributes to the building up of the daughter-polyp. Lang regarded this method of budding as universal in polyps, a notion disproved by O. Seeliger [52] who went to the opposite extreme and regarded the type of budding described by Lang as non-existent. In view, however, both of the statements and figures of Lang and of the facts to be described presently for medusae (Margellium), it is at least theoretically possible that both germinal and vegetative budding may occur in polyps as well as in medusae.

(b) The Medusa. The clearest instance of germinal budding is furnished by Margellium (Rathkea) octopunctatum, one of the Margelidae. The budding of this medusa has been worked out in detail by Chun (HvDROZOA, [1]), to whom the reader must be referred for the interesting laws of budding regulating the sequence and order of formation of the buds.

The buds of Margellium are produced on the manubrium in each of the four interradii, and they arise from the ectoderm, that is to say, the germinal epithelium, which later gives rise to the gonads. The buds do not appear simultaneously but successively on each of the four sides of the manubrium, thus: 3 4 and secondary buds may be produced on the medusa-buds before the latter are set free as medusae. Each bud arises as a thickening of the epithelium, which first forms two or three layers (fig. 47, A), and becomes separated into a superficial layer, future ectoderm, surrounding a central mass, future endoderm (fig. 47, B). The ectodermal epithelium on the distal side of the bud becomes thickened, grows inwards, and forms a typical entocodon (fig. 37, D, E, F). The remaining development of the bud "s just as described above for the indirect method of medusa-budding 'fig. 47, G, H). When the bud is nearly complete, the body-wall of the parent immediately below it becomes perforated, placing the coelenteric cavity of the parent in secondary communication with that of the bud (H), doubtless for the better nutrition of the latter.

Especially noteworthy in the germinal budding of Margellium is the formation of the entocodon, as in the vegetative budding of the indirect type.

5. Sporogony. This method of reproduction has been described by E. Metchnikoff in Cunina and allied genera. In individuals either of the male or female sex, germ-cells which are quite undifferentiated and neutral in character, become amoeboid, and wander into the endoderm. They divide each into two sistercells, one of which the spore becomes enveloped by the other. The spore-cell multiplies by division, while the enveloping cell is nutrient and protective. The spore cell gives rise to a " sporelarva," which is set free in the coelenteron and grows into a medusa. Whether sporogony occurs also in the polyp or not remains to be proved.

6. Sexual Reproduction and Embryology. The ovum of Hydromedusae is usually one of a large number of oogonia, and grows at the expense of its sister-cells. No regular follicle is formed, but the oocyte absorbs nutriment from the remaining oogonia. In Hydra the oocyte is a large amoeboid cell, which sends out pseudopodia amongst the oogonia and absorbs nutriment from them. When the oocyte is full grown, the residual oogonia die off and disintegrate.

The spermatogenesis and maturation and fertilization of the germ-cells present nothing out of the common and need not be FIG. 47. Budding from the Ectoderm (germinal epithelium) in Margellium. (After C. Chun.)

A, The epithelium becomes two- the bud forms an entocodon layered. (Gc.).

B, The lower layer forms a solid G,H, Formation of the medusae.

mass of cells, which (C) s.c, Sub-umbral cavity, becomes a vesicle, the future r.c. Radial canal, endoderm, containing the st, Stomach, which in H accoelenteric cavity (coel), quires a secondary corn- while the outer layer munication with the diges- furnishes the future ecto- tive cavity of the mother, derm. c.c. Circular canal.

D, E, F, a thickening of the ecto- v. Velum, derm on the distal side of /, Tentacle.

described here. These processes have been studied in detail by A. Brauer [2] for Hydra.

The general course of the development is described in the article HYDROZOA. We may distinguish the following series of stages: (i) ovum; (2) cleavage, leading to formation of a blastula; (3) formation of an inner mass or parenchyma, the future endoderm, by immigration or delamination, leading to the so-called parenchymula-stage; (4) formation of an archenteric cavity, the future coelenteron, by a splitting of the internal parenchyma, and of a blastopore, the future mouth, by perforation at one pole, leading to the gast ru la-stage ; (5) the outgrowth of tentacles round the mouth (blastopore), leading to the actinula-stage; and (6) the actinula becomes the polyp or medusa in the manner described elsewhere (see articles HYDROZOA, POLYP and MEDUSA). This is the full, ideal development, which is always contracted or shortened to a greater or less extent. If the embryo is set free as a free-swimming, so-called planula-larva, in the blastula, parenchymula, or gastrula stage, then a free actinula stage is not found; if, on the other hand, a free actinula occurs, then there is no free planula stage.

The cleavage of the ovum follows two types, both seen in Tubularia (Brauer [3]). In the first, a cleavage foljows each nuclear division; in the second, the nuclei multiply by division a number of times, and then the ovum divides into as many blastomeres as there are nuclei present. The result of cleavage in all cases is a typical blastula, which when set free becomes oval and develops a flagcllum to each cell, but when not set free, it remains spherical in form and has no fiagella.

The germ-layer formation is always by immigration or delamination, never by invagination. When the blastula is oval and freeswimming the inner mass is formed by unipolar immigration from the hinder pole. When the blastula is spherical and not set free, the germ-layer formation is always multipolar, either by immigration or by delamination, i.e. by tangential division of the cells of the blastoderm, as in Geryonia, or by a mixture of immigration and delamination, as in Hydra, Tubularia, etc. The blastopore is formed as a secondary perforation at one spot, in free-swimming forms at the hinder pole. Formation of archenteron and blastopore may, however, be deferred till a later stage (actinula or after).

The actinula stage is usually suppressed or not set free, but it is seen in Tubularia (fig. 48), where it is ambulatory, in Gonionemus (Trachomedusae), and in Cunina ( Narcomedusae), where it is parasitic.

In Leptolinae the embryonic development culminates in a polyp, which is usually formed by fixation of a planula (parenchymula), rarely by fixation of an actinula. The planula may fix itself (i) by one end, and then becomes the hydrocaulus and hydranth, while the hydrorhiza grows out from the base; or (2) partly by one side and then gives rise to Modified from a plate by L. the hydrorhiza as well as to the other Agassiz, Contributions to Nat. parts of the polyp; or (3) entirely by its'" side, and then forms a recumbent hydro- pi G . 48. Free Actinula rhiza from which a polyp appears to be o f Tubularia.

budded as an upgrowth.

In Trachylinae the development produces always a medusa, and there is no polyp-stage. The medusa arises direct from the actinulastage and there is no entocodon formed, as in the budding described above.

Life-cycles of the Hydromedusae. The life-cycle of the Leptolinae consists of an alternation of generations in which non-sexual individuals, polyps, produce by budding sexual individuals, medusae, which give rise by the sexual process to the non-sexual polyps again, so completing the cycle. Hence the alternation is of the type termed metagenesis. The Leptolinae are chiefly forms belonging to the inshore fauna. The Trachylinae, on the other hand, are above all oceanic forms, and have no polyp-stage, and hence there is typically no alternation in their life-cycle. It is commonly assumed that the Trachylinae are forms which have lost the alternation of generations possessed by them ancestrally, through secondary simplification of the life-cycle. Hence the Trachylinae are termed " hypogenetie " medusae to contrast them with the metagenetic Leptolinae. The whole question has, however, been argued at length by W. K. Brooks [4], who adduces strong evidence for a contrary view, that is to say, for regarding the direct type of development seen in Trachylinae as more primitive, and the metagenesis seen in Leptolinae as a secondary complication introduced into the life-cycle by the acquisition of larval budding. The polyp is regarded, on this view, as a form phylogenetically older than the medusa, in short, as nothing more than a sessile actinula. In Trachylinae the polyp-stage is passed over, and is represented only by the actinula as a transitory embryonic stage. In Leptolinae the actinula becomes the sessile polyp which has acquired the power of budding and producing individuals either of its own or of a higher rank; it represents a persistent larval stage and remains in a sexually immature condition as a neutral individual, sex being an attribute only of the final stage in the development, namely the medusa. The polyp of the Leptolinae has reached the limit of its individual development and is incapable of becoming itself a medusa, but only produces medusa-buds; hence a true alternation of generations is produced. In Trachylinae also the beginnings of a similar metagenesis can be found. Thus in Cunina octonaria, the ovum develops into an actinula which buds daughteractinulae; all of them, both parent and offspring, develop into medusae, so that there is no alternation of generations, but only larval multiplication. In Cunina parasitica, however, the ovum develops into an actinula, which buds actinulae as before, but only the daughter-actinulae develop into medusae, whije the original, parent-actinula dies off; here, therefore, larval budding has led to a true alternation of generations. In Gonionemus the actinula becomes fixed and polyp-like, and reproduces by budding, so that here also an alternation of generations may occur. In the Leptolinae we must first substitute polyp for actinula, and then a condition is found which can be compared to the case of Cunina parasitica or Gonionemus, if we suppose that neither the parent-actinula (i.e. founder-polyp) nor its offspring by budding (polyps of the colony) have the power of becoming medusae, but only of producing medusae by budding. For further arguments and illustrations the reader must be referred to Brooks's most interesting memoir. The whole theory is one most intimately connected with the question of the relation between polyp and medusa, to be discussed presently. It will be seen elsewhere, however, that whatever view may be held as to the origin of metagenesis in Hydromedusae, in the case of Scyphomedusae (?.f.) no other view is possible than that the alternation of generations is the direct result of larval proliferation.

To complete our survey of life-cycles in the Hydromedusae it is necessary to add a few words about the position of Hydra and its allies. If we accept the view that Hydra is a true sexual polyp, and that its gonads are not gonophores (i.e. medusa-buds) in the extreme of degeneration, then it follows from Brooks's theory that Hydra must be descended from an archaic form in which the medusan type of organization had not yet been evolved. Hydra must, in short, be a living representative of the ancestor of which the actinula-stage is a transient reminiscence in the development of higher forms. It may be pointed out in this connexion that the fixation of Hydra is only temporary, and that the animal is able at all times to detach itself, to move to a new situation, and to fix itself again. There is no difficulty whatever in regarding Hydra as bearing the same relation to the actinula-stage of other Hydromedusae that a Rotifer bears to a trochophore-larva or a fish to a tadpole.

The Relation of Polyp and M edusa. Many views have been put forward as to the morphological relationship between the two types of person in the Hydromedusae. For the most part, polyp and medusa have been regarded as modifications of a common type, a view supported by the existence, among Scyphomedusae (q.v.), of sessile polyp-like medusae (Lucernaria, etc.). R. Leuckart in 1848 compared medusae in general terms to flattened polyps. G. J. Allman [1] put forward a more detailed view, which was as follows. In some polyps the tentacles are webbed at the base, and it was supposed that a medusa was a polyp of this kind set free, the umbrella being a greatly developed web or membrane extending between the tentacles. A very different theory was enunciated by E. Metchnikoff. In some hydroids the founder-polyp, developed from a planula after fixation, throws out numerous outgrowths from the base to form the hydrorhiza; these outgrowths may be radially arranged so as to form by contact or coalescence a flat plate. Mechnikov considered the plate thus formed at the base of the polyp as equivalent to the umbrella, and the body of the polyp as equivalent to the manubrium, of the medusa; on this view the marginal tentacles almost invariably present in medusae are new formations, and the tentacles of the polyp are represented in the medusa by the oral arms which may occur round the mouth, and which sometimes, e.g. in Margelidae, have the appearance and structure of tentacles. Apart from the weighty arguments which the development furnishes against the theories of Allman and Mechnikov, it may be pointed out that neither hypothesis gives a satisfactory explanation of a structure universally present in medusae of whatever class, namely the endoderm-lamella, discovered by the brothers O. and R. Hertwig. It would be necessary to regard this structure as a secondary extension of the endoderm in the tentacle-web, on Allman's theory, or between the outgrowths of the hydrorhiza, on Mechnikov's hypothesis. The development, on the contrary, shows unequivocally that the endodermlamella arises as a local coalescence of the endodermal linings of a primitively extensive gastral space.

The question is one intimately connected with the view taken as to the nature and individuality of polyp, medusa and gonophore respectively. On this point the following theories have been put forward.

i. The theory that the medusa is simply an organ, which has become detached and has acquired a certain degree of independence, like the well-known instance of the hectocotyle of the cuttle-fish. On this view, put forward by E. van Beneden and T. H. Huxley, the sporosac is the starting-point of an evolution leading up through the various types of gonophores to the free medusa as the culminating point of a phyletic series. The evidence against this view may be classed under two heads: first, comparative evidence; hydroids very different in their structural characters and widely separate in the systematic classification of these organisms may produce medusae very similar, at least so far as the essential features of medusan organization are concerned; on the other hydroids closely allied, perhaps almost indistinguishable, may produce gonophores in the one case, medusae in the other; for example, Hydractinia (gonophores) and Podocoryne (medusae), Tubularia (gonophores) and Ectopleura (medusae), Coryne (gonophores) and Syncoryne (medusae), and so on. Tf it is assumed that all these genera bore gonophores ancestrally, then medusa of similar type must have been evolved quite inde- pendently in a great number of cases. Secondly, there is the evidence from the development, namely, the presence of the entocodon in the medusa-bud, a structure which, as explained above, can only be accounted for satisfactorily by derivation from a medusan type of organization. Hence it may be concluded that the gonophores are degenerate medusae, and not that the medusae are highly elaborated gonophores, as the organ-theory requires.

2. The theory that the medusa is an independent individual, fully equivalent to the polyp in this respect, is now universally accepted as being supported by all the facts of comparative morphology and development. The question still remains open, however, which of the two types of person may be regarded as the most primitive, the most ancient in the race-history of the Hydromedusae. F. M. Balfour put forward the view that the polyp was the more primitive type, and that the medusa is a special modification of the polyp for reproductive purposes, the result of division of labour in a polypcolony, whereby special reproductive persons become detached and acquire organs of locomotion for spreading the species. W. K. Brooks, on the other hand, as stated above, regards the medusa as the older type and looks upon both polyp and medusa, in the Hydromedusae, as derived from a free-swimming or floating actinula, the polyp being thus merely a fixed nutritive stage, possessing secondarily acquired powers of multiplication by budding.

The Hertwigs when they discovered the endoderm-lamella showed on morphological grounds that polyp and medusa are independent types, each produced by modification in different directions of a more primitive type represented in development by the actinulastage. If a polyp, such as Hydra, be regarded simply as a sessile actinula, we must certainly consider the polyp to be the older type, and it may be pointed out that in the Anthozoa only polyp- individuals occur. This must not be taken to mean, however, that the medusa is derived from a sessile polyp; it must be regarded as a direct modification of the more ancient free actinula form, without primitively any intervening polyp-stage, such as has been introduced secondarily into the development of the Leptolinae and represents a revival, so to speak, of an ancestral form or larval stage, which has taken on a special role in the economy of the species.

SYSTEMATIC REVIEW OF THE HYDROMEDUSAE ORDER I. Eleutheroblastea. Simple polyps which become sexually mature and which also reproduce non-sexually, but without any medusoid stage in the life-cycle.

The sub-order includes the family Hydridae, containing the common fresh-water polyps of the genus Hydra. Certain other forms of doubtful affinities have also been referred provisionally to this section.

Hydra. This genus comprises fresh-water polyps of simple structure. The body bears tentacles, but shows no division into hydrorhiza, hydrocaulus or hydranth ; it is temporarily fixed and has no perisarc. The polyp is usually hermaphrodite, developing both ovaries and testes in the same individual. There is no free-swimming planula larva, but the stage corresponding to it is passed over in an enveloping cyst, which is secreted round the embryo by its own ectodermal layer, shortly after the germ-layer formation is complete, i.e. in the parenchymula-stage. The envelope is double, consisting of an external chitinous stratified shell, and an internal thin elastic membrane. Protected by the double envelope, the embryo is set free as a so-called " e gg>" a "d in Europe it passes the winter in this condition. In the spring the embryo bursts its shell and is set free as a minute actinula which becomes a Hydra.

Many species are known, of which three are common in European waters. It has been shown by C. F. Jickeli (28) that the species are distinguishable by the characters of their nematocysts. They also show characteristic differences in the egg (Brauer [2]). In Hydra viridis the polyp is of a green colour and produces a spherical egg with a smooth shell which is dropped into the mud. H. grisea is greyish in tint and produces a spherical egg with a spiky shell, which also is dropped into the mud. H. fusca (=H. vulgaris) is brown in colour, and produces a bun-shaped egg, spiky on the convex surface, and attached to a water-weed or some object by its flattened side. Brauer found a fourth species, similar in appearance to H. fusca, but differing from the three other species in being of separate sexes, and in producing a spherical egg with a knobby shell, which is attached like that of H. fusca.

The fact already noted that the species of Hydra can be distinguished by the characters of their nematocysts is a point of great interest. In each species, two or three kinds of nematocysts occur, some large, some small, and for specific identification the nematocysts must be studied collectively in each species. It is very remarkable that this method of characterizing and diagnozing species has never been extended to the marine hydroids. It is quite possible that the characters of the nematocysts might afford data as useful to the systematist in this group as do the spicules of sponges, for instance. It would be particularly interesting to ascertain how the nematocysts of a polyp are related to those possessed by the medusa budded from it, and it is possible that in this manner obscure questions of relationship might be cleared up.

Protohydra is a marine genus characterized by the absence of tentacles, by a great similarity to Hydra in histological structure, and by reproduction by transverse fission. It was found originally in an oyster-farm at Ostend. The sexual reproduction is unknown. For further information see C. Chun (HYDROZOA [1J.P1. I.).

Polypodium hydriforme Ussow is a fresh-water form parasitic on the eggs of the sterlet. A " stolon " of unknown origin produces thirty-two buds, which become as many Polypodia-, each has twenty-four tentacles and divides by fission repeated twice into four individuals, each with six tentacles. The daughter-individuals grow, form the full number of twenty-four tentacles and divide again. The polyps aref ree and walk on their tentacles. See Ussow [54].

Tetraplatia volitans Viguier is a remarkable floating marine form. See C. Viguier [56] and Delage and Herouard (HYDROZOA [2]).

Haleremita Schaudinn. SeeF.Schaudinn [50] and Delage and Herouard (HYDROZOA [2]).

In all the abovementioned genera, with the exception of Hydra, the life-cycle is so imperfectly known that their true position cannot be determined in the present state of our knowledge. They may prove eventually to belong to other orders. Hence only the genus Hydra can be considered as truly representing the order Eleutheroblastea. The phylogenetic position of this genus has been discussed above.

FIG. 49. Diagram showing possible modifications of persons of a gymnoblastic Hydromedusa. (After Allman.)

a, Hydrocaulus (stem).

b, Hydrorhiza (root).

c, Enteric cavity.

d, Endoderm.

e, Ectoderm.

/, Perisarc, (horny case).

g, Hydranth (hydriform person)expanded.

g', Hydranth (hydriform person) contracted.

h, Hypostome, bearing mouth at its extremity.

k, Sporosac springing from the hydrocaulus.

k', Sporosac springing from m, a modified hydriform person (blastostyle) : the genitalia are seen surrounding the spadix or manubrium.

/, Medusiform person or medusa.

m, Blastostyle.

ORDER II. Hydroidea seu Leptolinae. Hydromedusae with alternation of generations (metagenesis)in which a non-sexual polypgeneration ( trophosome) produces by budding a sexual medusa-generation (gonosome). The polyp may be solitary, but more usually produces polyps by budding and forms a polyp-colony. The polyp usually has the body distinctly divisible into hydranth, hydrocaulus and hydrorhiza, and is usually clothed in a perisarc. The medusae may be set free or may remain attached to the polyp-colony and degenerate into a gonophore. When fully developed the medusa is characterized by the sense organs being composed entirely of ectoderm, developed independently of the tentacles, and innervated from the sub-umbral nerve-ring.

The two kinds of persons present in the typical Hydroidea make the classification of the group extremely difficult, for reasons explained above. Hence the systematic arrangement that follows must be considered purely provisional. A natural classification of the Hydroidea has yet to be put forward. Many genera and families are separated by purely artificial characters, mere shelfand-bottle groupings devised for the convenience of the museum curator and the collector. Thus many subdivisions are diagnosed by setting free medusae in one case, or producing gonophores in another; although it is very obvious, as pointed out above, that a genus producing medusae may be far more closely allied to one producing gonophores than to another producing medusae, or vice versa, and that in some cases the production of medusae or gonophores varies with the season or the sex. Moreover, P. Hallez [22]_has recently shown that hydroids hitherto regarded as distinct species are only forms of_the same species grown under different conditions.

SUB-ORDER i. HYDROIDEA GYMNOBLASTEA (ANTHOMEDUSAE). Trophosome without hydrothecae or gonothecae, with monopodial type of budding. Gonosome with free medusae or gonophores; medusae usually with ocelli, never with otocysts. The gymnoblastic polyp usually has a distinct perisarc investing the hydrorhiea and the hydrocaulus, sometimes also the hydranth as far as the bases of the tentacles (Bimeria); but in such cases the perisarc forms a closely-fitting investment or cuticule on the hydranth, never a hydrotheca standing off from it, as in the next sub-order. The polyps may be solitary, or form colonies, which may be of the spreading or encrusting type, or arborescent, and then always of monopodial growth and budding. In some cases, any polyp of the colony may bud medusae; in other cases, only certain polyps, the blastostyles, have this power. When blastostyles are present, however, they are never enclosed FIG. 50. Sarsia (Dipurena) gemnifera. b, The long manubrium, bearing medusiform buds; a, mouth.

FIG. 51. Sarsia prolifera. Ocelli are seen at the base of the tentacles, and also (as an exception) groups of medusiform buds.

in special gonothecae as in the next sub-order. In this sub-order the characters of the hydranth are very variable, probably owing to the fact that it is exposed and not protected by a hydrotheca, as in Calyptoblastea.

Speaking generally, three principal types of hydranth can be distinguished, each with subordinate varieties of form.

1. Club-shaped hydranths with numerous tentacles, generally scattered irregularly, sometimes with a spiral arrangement, or in whorls (" verticillate ").

(a) Tentacles filiform; type of Clava (fig. 5), Cordylophora, etc.

(b) Tentacles capitate, simple; type of Coryne and byncoryne; Myriothela is an aberrant form with some of the tentacles modified as " claspers " to hold the ova. it (c) Tentacles capitate, branched, wholly or in part; type of Cladocoryne.

(d) Tentacles filiform or capitate, tending to be arranged in definite whorls; type of Stauridium (fig. 2), Cladonema and Pennaria.

2. Hydranth more shortened, daisy-like in form, with two whorls of tentacles, oral and aboral.

(a) Tentacles filiform, simple, radially arranged or scattered irregularly; type of Tubularia (fig. 4), Corymorpha (fig. 3), Nemopsis, Pelagohydra, etc.

(b) Tentacles with a bilateral arrangement, branched tentacles in addition to simple filiform ones; type of Branchiocerianthus.

3. Hydranth with a single circlet of tentacles.

(a) With filiform tentacles; the commonest type, seen in Bougainvillea (fig. 13), Eudendrium, etc.

(b) With capitate tentacles; type of Clavatella.

4. Hydranth with tentacles reduced below four; type of Lar(fig. il), Monobrachium, etc.

The Anlhomedusa in form is generally deep, bell-shaped. The sense organs are typically ocelli, never otocysts. The gonads are borne on the manubrium, either forming a continuous ring (Codonid type), or four masses or pairs of masses (Oceanid type). The tentacles may be scattered singly round the margin of the umbrella (" monerenematous ") or arranged in tufts (" lophonematous ") ; in form they may be simple or branched (Cladonemid type); in structure they may be hollow (" coelomerinthous "); or solid (" pycnomerinthous ") When sessile gonophores are produced, they may show all stages- of degeneration.

Classification. Until quite recently the hydroids (Gymnoblastea) and the medusae (Anthomedusae) have been classified separately, since the connexion between them was insufficiently known. Delage and Herouard (HYDROZOA [2]) were the first to make an heroic attempt to unite the two classifications into one, to which Hickson (HYDROZOA [4]) has made some additions and slight modifications. The classification given here is for the most part that of Delage and Herouard. It is certain, however, that no such classification can be considered final at present, but must undergo continual revision in the future. With this reservation we may recognize fifteen wellcharacterized families and others of more doubtful nature. Certain discrepancies must also be noted.

1. Margelidae ( = medusa-family Margelidae+hydrold families Bougainvtilidae, Dicorynidae, Bimeridae and Eudendridae). Trophosome arborescent, with hydranths of Bougainvillea-type; gonosome free medusae or gonophores, the medusae with solid tentacles in tufts (lophonematous). Common genera are the hydroid Bougainvillea (figs. 12, 13), and the medusae Hippocrene (budded from Bougainvillea), Margelis, Rathkea (fig. 24), and MargeUium. Other hydroids are Garveia, Bimeria, Eudendrium and Heterocordyle, with gonophores, and Dicoryne with peculiar sporosacs.

2. Podocorynidae( = medusa-families Thamnostomidae and Cytaeidae -fhydroid families Podocorynidae and Hydractiniidae) . Trophosome encrusting with hydranths of BougainviUea-type, polyps differentiated into blastostyles, gastrozoids and dactylozoids ; gonosome free medusae or gonophores. The typical genus is the well-known hydroid Podocoryne, budding the medusa known as Dysmorphosa ; Thamnostylus, Cytaeis, etc., are other medusae with unknown hydroids. Hydractinia. (figs. 9, 10) is a familiar hydroid genus, bearing gonophores.

3. Cladonemidae. T rophosome, polyps with two whorls of tentacles, the lower filiform, the upper capitate; gonosome, free medusae, with tentacles solid and branched. The type-genus Cladonema (fig. 20) is a common British form.

4. Clavatellidae. Trop hosome, polyps with a single whorl of capitate tentacles; gonosome, free medusae, with tentacles branched, solid. Clavatella (fig. 21), with a peculiar ambulatory medusa is a British form.

5. Pennariidae. T rophosome, polyps with an upper circlet of numerous capitate tentacles, and a lower circlet of filiform tentacles. Pennaria, with a free medusa known as Globiceps, is a common Mediterranean form. Stauridium (fig. 2) is a British hydroid. 6. Tubulariidae. Trophosome, polyps with two whorls of tentacles, both filiform. Tubularia (fig. 4), a well-known British hydroid, bears gonophores.

After Haeckel, System der Medusen, by permission of Gustav Tischer.

FIG. 52. Tiara pileata, L. Agassiz.

7. Corymorphidae (including the medusa-family Hybocodonidae). Trophosome solitary polyps, with two whorls of tentacles; gonosome, free medusae or gonophores. Corymorpha (fig. 3), a wellknown British genus, sets free a medusa known as Steenstrupia (fig. 22). Here belong the deep-sea genera Monocaidus and Branchiocerianthus, including the largest hydroid polyps known, both genera producing sessile gonophores.

8. Dendrodamdae. T r o p h o- some, polyp with filiform tentacles in three or four whorls. Dendroclava, a hydroid, produces the medusa known as Turritopsis.

9. Clavidae (including the medusa-family Tiaridae (figs. 27 and 5 1 )- Trophosome, polyps with scattered filiform tentacles; gonosome, medusae or gonophores, the medusae with hollow tentacles. Clava (fig. 5), a common British hydroid, produces gonophores; so also does Cordylophora, a form inhabiting fresh or brackish water. Turns produces free medusae. Amphinema is a medusan genus of unknown hydroid.

10. Bythotiaridae. Trophosome unknown; gonosome, free medusae, with deep, bell-shaped umbrella, with interradial gonads ' on the base of the stomach, with branched radial canals, and correspondingly numerous hollow tentacles. Bythotiara, Sibogita.

11. Corynidae( = hydroid families Corynidae, Syncorynidae and Cladocorynidae+medusan family Sarsiidae). Trophosome polyps After Haeckel, System der Medusen, by with capitate tentacles, simple or pe.ss,on of Gustav F,scher. branched, scattered or verticillate ; ^, FIG - 53- Pteronema darunnn. gonosome, free medusae or gono- The apex of the stomach is prophores. Coryne, a common British longed into a brood pouch conhydroid, produces gonophores ; taming embryos. Syncoryne, indistinguishable from it, produces medusae known as Sarsia (fig. 51). Cladocoryne is another hydroid genus ; Codonium and Dipurena (fig. 50) are medusan genera.

12. Myriothelidae. The genus Myriothela is a solitary polyp with scattered capitate tentacles, producing sporosacs.

13. Hydrolaridae, Trophosome (only known in one genus), polyps with two tentacles forming a creeping colony; gonosome, free medusae with four, six or more radial canals, giving off one or more lateral branches which run to the margin of the umbrella, with the stomach produced into four, six or more lobes, upon which the gonads are developed; the mouth with four lips or with a folded margin; the tentacles simple, arranged evenly round the margin of the umbrella. The remarkable hydroid Lar (fig. n) grows upon the tubes of the worm Sabella and produces a medusa known as Willia. Another medusan genus is Proboscidactyla.

14. Monobrachiidae. The genus Monobrachium^ is a colonyforming hydroid which grows upon the shells of bivalve molluscs, each polyp having but a single tentacle. It buds medusae, which, however, are as yet only known in an immature condition (C. Mereschkowsky [41]).

15. Ceratellidae. Trophosome polyps forming branching colonies of which the stem and main branches are thick and composed of a network of anastomosing coenosarcal tubes covered by a common ectoderm and supported by a thick chitinous perisarc; hydranths similar to those of Coryne; gonosome, sessile gonophores. Ceratella, an exotic genus from the coast of East Africa, New South Wales and Japan. The genera Dehitella Gray and Dendrocoryne Inaba should perhaps be referred to this family; the last-named is regarded by S. Goto [16] as the type of a distinct family, Dendrocorymdae.

Doubtful families, or forms difficult to classify, are : Pteronemidae, Medusae of Cladonemid type, with hydroids for the most part unknown. The British genus Gemmariq, however, is budded from a hydroid referable to the family Corynidae. Pteronema (fig. 53).

Nemopsidae, for the floating polyp Nemopsis, very similar to Tubularia in character; the medusa, on the other hand, is very similar to Hippocrene (Margelidae). See C. Chun (HYDROZOA[!]).

Pelagohydridae, for the floating polyp Pelagohydra, Dendy, from New Zealand. The animal is a solitary polyp bearing a great number of medusa-buds. The body, representing the hydranth of an ordinary hydroid, has the aboral portion modified into a float, from which hangs down a proboscis bearing the mouth. The float is covered with long tentacles and bears the medusa-buds. The proboscis bears at its extremity a circlet of smaller oral tentacles. Thus the affinities of the hydranth are clearly, as Dendy points out, with a form such as Corymorpha, which also is not fixed but only rooted in_the mud. The medusae, on the other hand, have thi tentacles in four tufts of (in the buds) five each, and thus resemble the medusae of the i family Mareelidae.

See A. Dendy [12].

Perigonimus. This common British hydroid belongs by it characters to the family Bougainvillidae; it produces, however, a medusa of the genus Tiara (fig. 52), referable to the family Clavidae ; a fact sufficient to indicate the tentative character of even the most modern classifications of this order.

SUB-ORDER II. HYDROIDEA CALYP-TOBLASTEA (LEPTO-MEDUSAE). Trophosome with polyps always differentiated into nutritive and reproductive individuals(blastostyles) enclosed in hydrothecae and gonothecae respectively; with sympodial type FIG. 54. Diagram showing possible modi- of budding. Gonofications of the persons of a Calyptoblastic some with free medHydromedusa. Letters a to h same as in <nnnr>tin>c.

fig. 49- i, The horny cup or hydrotheca of u * ae Or 8 on P nores .> the hydriform persons; /, medusiform person tne medusae typlspringing from m, a modified hydriform cally with otocysts, person (blastostyle); n, the horny case or sometimes with corgonangium enclosing the blastostyle and j i; __ n/ ._ii; if. a . its buds. This and the hydrotheca i give dyh . r oceUl origin to the name Calyptoblastea. (After 54, 55)- Allman.) The calyptoblastic polyp of the nutritive type is very uniform in character, its tendency to variation being limited, as it were, by the enclosing hydrotheca. The hydranth almost always has a single circlet of tentacles, like the Bougainvillea-type in the preceding sub-order; an exception is the curious genus Clathrozoon, in which the hydranth has a single tentacle. The characteristic hydrotheca is formed by the bud at an early stage (fig. 56); when complete it is an open cup, in which the hydranth develops and can be protruded from the opening for the capture of food, or is withdrawn into it for protection. Solitary polyps are unknown in this sub-order; the colony may be creeping or arborescent in form ; FIG. 55. View of the Oral Surface of if tne latter, the budding one of the Leptomedusae (Irene pellu- of the polyps, as already g*. Genital glands, re, The four mdi M, Manubrium. ating canals.

ot t Otocysts. Ve, The velum.

T forming stems capable of further branching, or uniserial, forming pinnules not capable of further branching. In the biserial type the polyps on the two sides of the stem have primitively an alternating, zigzag arrangement; but, by a process of differential growth, quickened in the ist, 3rd, sth, etc., members of the stem, and retarded in the 2nd, 4th, 6th, etc., members, the polyps may assume secondarily positions opposite to one another on the two sides of the stem. Other variations in the mode of growth or budding bring about further differences in the building up of the colony, which are not in all cases properly understood and cannot be described in detail here. The stem may contain a single coenosarcal tube (" monosiphonic ") or several united in a common perisarc (" polysiphonic "). An important variation is seen, in the form of the hydrotheca itself, which may come off from the main stem by a stalk, as in Obelia, or may be sessile, without a stalk, as in Sertularia.

In many Calyptoblastea there occur also reduced defensive polyps or dactylozoids, which in this sub-order have received the special name of sarcostyles. Such are the " snake-like zoids " of Ophiodes and other genera, and as such are generally interpreted the " machopolyps" of the P lumular idae. These organs are supported by cuplike structures of the perisarc, termed nematophores, regarded as modified hydrothecae supporting the specialized polyp- individuals. They are specially characteristic of the family Plumularidae.

The medusa-buds, as already stated, are always produced from blastostyles, reduced non- nutritive polyps without mouth or tentacles. An apparent, but not real, exception is Halecium kalecinum, in which the blastostyle is produced from the side of a nutritive polyp, and both are enclosed in a common theca without a partition between them (Allman [1] p. 50, fig. 24). The jonotheca is formed in its early stage in the same way as the iydrotheca, but the remains of the hydranth persists as an operculum closing the capsule, to be withdrawn when the medusae or genital products are set free (fig. 56) .

Theblastostyles, gonophoresand gonothecaefurnishaseriesof variations which can best be considered as so many stages of evolution. Stage i , seen in Obelia. Numerous medusae are budded successively within the gonotheca and set free; they swim off and mature in the open sea (Allman [1], p. 48, figs. 18, IQ).

Stage 2, seen in Gonothyraea. Medusae, so-called " meconidia," are budded but not liberated; each in turn, when it reaches sexual maturity, is protruded from the gonotheca by elongation of the stalk, and sets free the embryos, after which it withers and is replaced by another (Allman [1], p. 57, fig. 28).

Stage 3, seen in Sertularia. The gonophores are reduced in varying degree, it may be to sporosacs; they are budded successively from he blastostyle, and each in turn, when ripe, protrudes the spadix through the gonotheca (fig. 57, A, B). The spadix forms a gelatinous cyst, the so-called acrocyst (ac), external to the gonotheca (gth), inclosing and protecting the embryos. Then the spadix withers, caving the embryos in the acrocyst, which may be further protected >y a so-called_ marsupium, a structure formed by tentacle-like >rocesses growing out from the blastostyle to enclose the acrocyst, each such process being covered by perisarc like a glove-finger secreted by it (fig. 57, C). (Allman [1], pp. 50, 51, figs., 21-24; Weismann [58], p. 170, pi. ix., figs. 7, 8.)

After Allman, Gymnoblaslic Bydnids, by permission of the council of the Ray Society .

FIG. 56. Diagrams to show the mode of formation of the Hydrotheca and Gonotheca in Calyptoblastic Hydroids. A-D are stages common to both ; from D arises the hydrotheca (E) or the gonotheca (F); th, theca; st, stomach; t, tentacles; m, mouth; mb, medusa-buds.

Stage 4, seen in Plumularidae. The generative elements are produced in structures termed corbulae, formed by reduction and modification of branches of the colony. Each corbula contains a central row of blastostyles enclosed and protected by lateral rows of branches representing stunted buds (Allman [1], p. 60, fig. 30).

The Leptomedusa in form is generally shallow, more or less saucer-like, with velum less developed than in Anthomedusae (fig. 55). The characteristic sense-organs are ectodermal otocysts, absent, however, in some genera, in which case cordyli may replace them. When otocysts are present, they are at least eight in number, situated adradially, but are often very numerous. The cordyli are scattered on the ring-canal. Ocelli, if present, are borne on the tentacle-bulbs. The tentacles are usually hollow, rarely solid (Obelia). In number they are rarely less than four, but in Dissonema there are only two. Primitively there are four perradial tentacles, to which may be added four interradial, or they may become very numerous and are then scattered evenly round the margin, never arranged in 'tufts or clusters.

In addition to tentacles, there may be marginal cirri (Laodice) with a solid endodermal axis, spirally coiled, very contractile, and bearing a terminal battery of nematocysts. The gonads are developed typically beneath the radial canals or below the stomach or its pouches, often stretching as long bands on to the base of the manubrium. In Octorchidae (fig. 58) each such band is interrupted, forming one mass at the base of the manubrium and After Allman, Gymnoblastic Hydroids, by permission of the council of the Ray Society.

FIG. 57. Diagrams to show the mode of formation of an Acrocyst and a Marsupium. In A two medusa-buds are seen within thegonotheca (gth), the upper more advanced than the lower one. In B thespadixof the upper bud has protruded itself through the top of the gonotheca and the acrocyst (ac) is secreted round it. In C the marsupium (m) is formed as finger-like process from the summit of the blastostyle, en- another below the closing the acrocyst; b, medusa-buds on the blastostyle.

radial canal in each radius, in all eight separate gonad-masses, as the name implies. In some Leptomedusae excretory " marginal tubercles " are developed on the ring-canal.

Classification. As in the Gymnoblastea, the difficulty of uniting the hydroid and medusan systems into one scheme of classification is very great in the present state of our knowledge. In a great many Leptomedusae the hydroid stage is as yet unknown, and it is by no means certain even that they possess one. It is quite possible that some of these medusae will be found to be truly hypogenetic, that is to say, with a life-cycle secondarily simplified by suppression of metagenesis. At present, ten recent and one extinct family of Calyptoblastea (Leptomedusae) may be recognized provisionally:

1. Eucopidae (figs. 55, 59). Trophosome with stalked hydrothecae; gonosome, free medusae with otocysts and four, rarely six or eight, unbranched radial canals. Two of the commonest British hydroids belong to this family, Obelia and Clytia. Obelia forms numerous polyserial stems of the characteristic zigzag pattern growing up from a creeping basal stolon, and buds the medusa of the same name. In Clytia the polyps arise singly from the stolon, and the medusa is known as Phialidium (fig. 59).

2. Aequoridae. Trophosome only known in one genus ( Polycanna), and similar to the preceding; gonosome, free medusae with otocysts and with at least eight radial canals, often a hundred or more, simple or branched. Aequorea is a common medusa.

3. Thaumantidae. Trophosome only known in one genus ( Thaumantias), similar to that of the Eucopidae; gonosome, free medusae with otocysts inconspicuous or absent, with usually four, sometimes eight, rarely more than eight, radial canals, simple and unbranched, along which the gonads are developed, with numerous tentacles bearing ocelli and with marginal sense-clubs. Laodice and Thau* mantias are representative genera.

4. Berenicidae. Trophosome unknown; gonosome, free medusae, with four or six radial canals, bearing the gonads, with numerous tentacles, between which occur sense-clubs, without otocysts. Berenice, Staurodiscus, etc.

5. Polyorchidae. Trophosome unknown; gonosome, free medusae of deep form, with radial canals branched in a feathery manner, and After Haeckel, System dcr Mcdusen, by permission of Gustav Fischer.

FIG. 58. Octorchandra canariensis, from life, magnified 4 diameters.

bearing gonads on the main canal, but not on the branches, with numerous hollow tentacles bearing ocelli, and without otocysts. Polyorchis, Spirocodon.

6. Campanularidae. Trophosome as in Eucopidae; gonosome, sessile gonophores. Many common or well-known genera belong here, such as Halecium, Campanularia, Gonothyraea, etc.

7. Lafoeidae. Trophosome as in the preceding; gonosome, free medusae or gonophores, the medusae with large open otocysts. The hydroid genus Lafoea is remarkable for producing gonothecae on the hydrorhiza, each containing a blastostyle which bears a single gonophore; this portion of the colony was formerly regarded as an independent parasitic hydroid, and was named Coppinia. Medusan genera are Mitrocoma, Halopsis, Tiaropsis (fig. 29, etc.).

(So far as the characters of the trophosome are concerned, the seven preceding families are scarcely distinguishable, and they form After E. T. Browne, Proc. tool. Sac. of London, 1896.

FIG. 59. Three stages in the development of Phialidium temporarium. a, The youngest stage, is magnified 22 diameters; 6, older, is magnified 8 diameters; c, the adult medusa, is magnified 6 diameters.

a section apart, contrasting sharply with the families next to be mentioned, in none of which are free medusae liberated from the colony, so that only the characters of the trophosome need be considered.)

8. Sertularidae. Hydrothecae sessile, biserial, alternating or opposite on the stem. Sertularia and Sertularella are two very common genera of this family.

9. Plumularidae. Hydrothecae sessile, biserial on the main stem, uniserial on the lateral branches or pinnules, which give the colony its characteristic feathery form; with nematophores. A very abundant and prolific family; well-known British genera are Plumularia, Antennularia and Aglaophenia.

10. Hydroceratinidae. This family contains the single Australian species Clathrozoon wilsoni Spencer, in which a massive hydrorhiza bears sessile hydrothecae, ('detaining hy^lranths each with a single tentacle, and numerous nematophores. See W. B. Spencer [53].

II. Dendrograptidae, containing fossil (Silurian) genera, such as Dendrograptus and Thamnograptus, of doubtful affinities.

Order III. Hydrocorallinae. Metagenetic colony-forming Hydromedusae, in which the polyp-colony forms a massive, calcareous corallum into which the polyps can be retracted; polyp-individuals always of two kinds, gastrozoids and dactylozoids; gonosome either free medusae or sessile gonophores.

The trophosome consists of a mass of coenosarcal tubes anastomosing in all planes. The interspaces between the tubes are filled up by a solid mass of lime, consisting chiefly of calcium carbonate, which replaces the chitinous perisarc of ordinary hydroids and forms a stony corallum or coenosteum (fig. 60). The surface of the coenosteum is covered by a layer of common ectoderm, containing large nematocysts, and is per- FIG. 60. Portion of the forated b Y P ores of two kinds > calcareous corallum of Mille- gastropores and dactylopores, pora nodosa, showing the giving exit to gastrozoids and cyclical arrangement of the dactylozoids respectively, which pores occupied by the per- , . , . .. , sons " or hydranths. Twice are lod ed m vertical pore-canals the natural size. (From f wider calibre than the coenoMoseley.) sarcal canals of the general net- work. The coenosteum increases in size by new growth at the surface; and in the deeper, older portions of massive forms the tissues die off after a certain time, only the superficial region retaining its vitality down to a certain depth. The living tissues at the surface are cut off from the underlying dead portions by horizontal partitions termed tabulae, which are formed successively as the coenosteum increases in age and size. If the coenosteum of Millepora be broken across, each pore-canal (perhaps better termed a polypcanal) is seen to be interrupted by a series of transverse partitions, representing successive periods of growth with separation from the underlying dead portions.

Besides the wider vertical pore-canals and the narrower, FIG. 61. Enlarged view of the surface of a living Millepora, showing five dactylozooids surrounding a central gastrozooid (From Moseley.)

irregular coenosarcal canals, the coenosteum may contain, in its superficial portion, chambers or ampullae, in which the reproductive zoids (medusae or gonophores) are budded from the coenosarc.

The gastropores and dactylopores are arranged in various ways at the surface, a common pattern being the formation of a cyclosystem_(fig. 60), in which a central gastrozoid is surrounded by a ring of dactylozoids (fig. 61). In such a system the dactylopores may be confluent with the gastropore, so that the entire cyclosystem presents itself as a single aperture subdivided by radiating partitions, thus having a superficial resemblance to a madreporarian coral with its radiating septa (figs. 62 and 63).

The gastrozoids usually bear short capitate tentacles, four, six or twelve in number; but in Astylus (fig. 63) they have no tentacles. The dactylozoids have no mouth; in MUleporidae they have short capitate tentacles, but lack tentacles in Stylasleridae.

Thegonosomecon- F , G 62 ._ Diagrams in ustrating the suc . sists of free medusae cessive stages in the development of the in MUleporidae, cyclosystems of the Stylasteridae. (After Moseley.)

i, Sporadopora dichotoma.

s, Style.

dp, Dactylopore.

gp, Gastropore.

o, In fig. 6, inner horseshoe- shaped mouth ' of gastropore.

which are budded from the apex of a dactylozoid in Mille- 2> ^AUopora nobilis.

pora murrayi, but in 4, Allopora profunda.

other species from 5, Allopora miniacea.

the r o e n o <s a r r a 1 6 ' Ast V lus subviridis.

1 7, Distichopora coccinea. canals. The medusae are produced by direct budding, without an entocodon in the bud. They are liberated in a mature condition, and probably live but a short time, merely sufficient to spread the species. The manubrium bearing the gonads is mouthless, and the umbrella is without tentacles, sense-organs, velum or radial canals. In the Stylasleridae sessile gonophores are formed, always by budding from the coenosarc. In Dislichopora the gonophores have radial canals, but in other genera they are sporosacs with no trace of medusoid structure.

Classification. Two families are known :

1. Milleporidae. Coenosteum massive, irregular in form; pores scattered irregularly or in cyclosystems, without styles, with transverse tabulae; free medusae. A single genus, Millepora (figs. 60, 61).

2. Stylasteridae. Coenosteum arborescent, sometimes fanlike, with pores only on one face, or on the lateral margins of the branches; gastropores with tabulae only in two genera, but with (except in Astylus) a style, i.e. a conical, thorn-like projection from the base of the pore, sometimes found also in dactylopores; sessile gonophores. Sporadopora has the pores scattered irregularly. Distichopora has the pores arranged in rows. Stylaster has cyclosystems. In Allopora the cyclostems resemble the calyces of Anthozoan corals. In Cryptohelia the cyclosystem is covered by a cap or oper- F IG - 63. Portion of culum. In Astylus (fig. 63) styles are the corallum of A stylus absent. subviridis (one of the Affinities of the Hydrocorallinae. There Stylasteridae), showing can be no doubt that the forms comprised cyclosystems placed at in this order bear a close relationship to the intervals on the Hydroidea, especially the sub-order Gymno- branches, each with a blastca, with which they should perhaps be central gastropore and classed in a natural classification. A hydro- zone of slit-like daccoralline may be regarded as a form of tylopores. (After hydroid colony in which the coenosarc Moseley.) forms a felt-work ramifying in all planes, and in which the chitinous perisarc is replaced by a massive calcareous skeleton. So far as the trophosome is concerned, the step from an encrusting hydroid such as Hydractinia to the hydrocoralline Millepora is not great.

Hickson considers that the families Milleporidae and Stylasteridae should stand quite apart from one another and should not be united in one order. The nearest approach to the Stylasteridae is perhaps to be found in Ceratella, with its arborescent trophosome formed of anastomosing coenosarcal tubes supported by a thick perisarc and covered by a common ectoderm. Ceratella stands in much the same relation to the Stylasteridae that Hydractinia does to the Milleporidae, in both cases the chitinous perisarc being replaced by the solid coenosteum to which the hydrocorallines owe the second half of their name.

ORDER IV. Qraptolitoidea (Rhabdophora, Allman). This order has been constituted for a peculiar group of palaeozoic fossils, which have been interpreted as the remains of the skeletons of Hydrozoa of an extinct type.

A typical graptolite consists of an axis bearing a series of tooth-like projections, like a saw. Each such projection is regarded as representing a cup or hydrotheca, similar to those borne by a calyptoblastic hydroid, such as Sertularia. The supposed hydrothecae may be present on one side of the axis only ( monoprionid) or on both sides (diprionid); the first case may be conjectured to be the result of uniserial (helicoid) budding, the second to be produced by biserial (scorpioid) budding. In one division (Retiolitidae) the axis is reticulate. In addition to the stems bearing cups, there are found vesicles associated with them, which have been interpreted as gonothecae or as floats, that is to say, aif-bladders, acting as hydrostatic organs for a floating polyp-colony.

Since no graptolites are known living, or, indeed, since palaeozoic times, the interpretation of their structure and affinities must of necessity be extremely conjectural, and it is by no means certain that they are Hydrozoa at all. It can only be said that their organization, so far as the state of their preservation permits it to be ascertained, offers closer analogies with the Hydrozoa, especially the Calyptoblastea, than with any other existing group of the animal kingdom.

See the treatise of Delage and Herouard (HYDROZOA, [4]), and the article GRAPTOLITES.

ORDER V. Trachylinea. Hydromedusae without alternation of generations, i.e. without a hydroid phase; the medusa develops directly from the actinula larva, which may, however, multiply by budding. Medusae with sense-organs represented by otocysts derived from modified tentacles (tentaculocysts), containing otoliths of endodermal origin, and innervated from the ex-umbral nerve-ring.

This order, containing the typical oceanic medusae, is divided into two sub-orders.

SUB-ORDER i. TRACHOMEDUSAE. Tentacles given off from the margin of the umbrella, which is entire, i.e. not lobed or indented; tentaculocysts usually enclosed in vesicles; gonads on the radial canals. The medusae of this order are characterized by the tough, rigid consistence of the umbrella, due partly to the dense nature of the mesogloea, partly to the presence of a marginal rim of chondral tissue, consisting of thickened ectoderm containing great numbers of nematocysts, and forming, as it were, a cushion-tyre supporting the edge of the umbrella. Prolongations from the rim of chondral tissue may form clasps or peronia supporting the tentacles. The tentacles are primarily four in number, perradial, alternating with four interradial tentaculocysts, but both tentacles and sense-organs may be multiplied and the primary perradii may be six instead of four {fig. 26). The tentacles are always solid, containing an axis of endoderm-cells resembling notochordal tissue or plantparenchyma, and are but moderately flexible. The sense-organs are tentaculocysts which are usually enclosed in vesicles and may be sunk far below the surface. The gonads are on the radial canals or on the stomach (Ptychogastridae), and each gonad may be divided into two by a longitudinal sub-umbral muscle-tract. The radial canals are four, six, eight or more, and in some genera blindly-ending centripetal canals are present (fig. 26). The stomach may be drawn out into the manubrium, forming a proboscis (" Magenstiel ") of considerable length.

The development of the Trachomedusae, so far as it is known, shows an actinula-stage which is either free (larval) or passed over in the egg (foetal) as in Geryonia; in no case does there appear to be a free planula-stage. The actinula, when free, may multiply by larval budding, but in all cases both the original actinula and all its descendants become converted into medusae, so that there is no alternation of generations. In Gonionemus the actinula becomes attached and polyp-like and reproduces by budding.

The Trachomedusae are divided into the following families :

1. Petasidae ( Petachnidae). Four radial canals, four gonads; stomach not prolonged into the manubrium, which is relatively short ; tentaculocysts free. Pelasus and other genera make up this family, founded by Haeckel, but no other naturalist has ever seen them, and it is probable that they are simply immature forms of other genera.

2. Olindiadae, with four radial canals and four gonads; manubrium short; ring-canals giving off blind centripetal canals; tentaculocysts enclosed. Olindias miilleri (fig. 64) is a common Mediterranean species. Other genera are Aglauropsis, Gossea and Gonionemus; the last named bears adhesive suckers on the tentacles. Some doubt attaches to the position of this family. It has been asserted that the tentaculocysts are entirely ectodermal and that either the family should be placed amongst the Leptomedusae, or should form, together with certain Leptomedusae, an entirely distinct order. In Gonionemus, however, the concrement-cells are endodermal.

3. Trachynemidae. Eight radial canals, eight gonads, stomach not prolonged into manubrium; tenta- culocysts enclosed. Rhopalonema, Trachynema, etc.

4. Ptychogastridae (Pectyllidae). As in the preceding, but with suckers on the tentacles. Ptychogastria Allman ( = Pectyllis), a deep-sea form.

5. Aglauridae. Eight radial canals, two, four or eight gonads; tentacles numerous; tentaculocysts free; stomach prolonged into manubrium. Aglaura, Aglantha (fig. 65), etc., with eight gonads; Stauraglaura with four ; Persa with two. Amphogona, hermaphrodite, with male and female gonads on alternating radial canals.

6. Geryonidae. Four or six radial canals; gonads band-like; stomach prolonged into a manubrium of great length ; tentaculocysts enclosed. Liriope, etc., with four radial canals; Geryonia, Carmarina (fig. 26), etc., with six.

broad After Haeckel, System drr Medusen, by permission of Gustav Fischer.

FIG. 64. Olindias mfdleri, twice natural size.

After E. T. Browne, Proc. Zool.

7. Halicreidae. Eight very broad Soc. o radial canals; ex-umbrella often pro- p IG _ gj_ Aglantha vided with lateral outgrowths; tentacles rosea (Forbes), a British differing in size, but in a single row. medusa, X5- Halicreas.

SUB-ORDER 2. NARCOMEDUSAE. Margin of the umbrellalobed, tentacles arising from the ex-umbrella at some distance from the margin; tentaculocysts exposed, not enclosed in vesicles; gonads on the sub-umbral floor of the stomach or of the gastric pouches.

The Narcomedusae exhibit peculiarities of form and structure which distinguish them at once from all other Hydromedusae. The umbrella is shallow and has the margin supported by a rim of thickened ectoderm, as in the Trachomedusae, but not so strongly developed. The tentacles are not inserted on the margin of the umbrella, but arise high up on the ex-umbral surface, and the umbrella is prolonged into lobes corresponding to the interspaces between the tentacles. The condition of things can be imagined by supposing that in a medusa primitively of normal build, with tentacles at the margin, the umbrella has grown down past the insertion of the tentacles. As a result of this extension of the umbrellar margin, all structures belonging to this region, namely, the ring-canal, the nerve-rings, and the rim of thickened ectoderm, do not run an even course, but are thrown into festoons, caught up under the insertion of each tentacle in such a way that the ring-canal and its accompaniments form in each notch of the umbrellar margin an inverted V, the apex of which corresponds FIG. 66. Cunina rhododactyla, one of the Narcomedusae. (After Haeckel.) c, Circular canal.

' ' Otoporpae ' ' or centripetal process of the marginal cartilaginous ring connected with tentaculocyst.

of the k, Stomach.

/, Jelly of the disk.

r, Radiating canal (pouch of stomach).

tt, Tentacles.

tw, Tentacle root.

to the insertion tentacle; in some cases the limbs of the V may run for some distance parallel to one another, and may be fused into one, giving a figure better compared to an inverted Y. Thus the ectodermal rim runs round the edge of each lobe of the umbrella and then passes upwards towards the base of the tentacle from the re-entering angle between two adjacent lobes, to form with its fellow of the next lobe a tentacle-clasp or peronium, i.e. a streak of thickened ectoderm supporting the tentacle. Similarly the ring-canal runs round the edge of the lobe as the so-called festoon-canal, and then runs upwards under the peronium to the base of the tentacle as one of a pair of peronial canals, the limbs of the V-like figure already mentioned. The nerve-rings have a similar course. The tentaculocysts are implanted round the margins of the lobes of the umbrella and may be supported by prolongations of the ectodermal rim termed otoporpae (Gehorspangen). The radial canals are represented by wide gastric pouches, and may be absent, so that the tentacles arise directly from the stomach (Solmaridae). The tentacles are always solid, as in Trachomedusae.

The development of the Narcomedusae is in the main similar to that of the Trachomedusae, but shows some remarkable features. In Aeginopsis a planula is formed by multipolar immigration. The two ends of the planula become greatly lengthened and give rise to the two primary tentacles of the actinula, of which the mouth arises from one side of the planula. Hence the principal axis of the future medusa corresponds, not to the longitudinal axis of the planula, but to a transverse axis. This is in some degree parallel to the cases described above, in which a planula gives rise to the hydrorhiza, and buds a polyp laterally.

In Cunina and allied genera the actinula, formed in the manner described, has a hypostome of great length, quite disproportionate to the size of the body, and is further endowed with the power of producing buds from a stolon arising from the aboral side of the body. In these species the actinula is parasitic upon another medusa; for instance, Cunoctantha oclonaria upon Turritopsis, C. proboscidea upon Liriope or Geryonia. The parasite effects a lodgment in the host either by invading it as a free-swimming planula, or, apparently, in other cases, as a spore-embryo which is captured and swallowed as food by the host. The parasitic actinula is found attached to the proboscis of the medusa; it thrusts its greatly elongated hypostome into the mouth of the medusa and nourishes itself upon the food in the digestive cavity of its host. At the same time it produces buds from an aboral stolon. The buds become medusae by the direct method of budding described above. In some cases the buds do not become detached at once, but the stolon continues to grow and to produce more buds, forming a "bud-spike" (Knospenahre) , which consists of the axial stolon bearing medusa-buds in all stages of development. In such cases the original parent-actinula does not itself become a medusa, but remains arrested in development and ultimately dies off, so that a true alternation of generations is brought about. It is in these parasitic forms that we meet with the method of reproduction by sporogony described above.

In other Narcomedusae, e.g. Cunoctantha fowleri Browne, buds are formed from the sub-umbrella on the under side of the stomach pouches, where later the gonads are developed.

Classification. Three families of Narcomedusae are recognized (see O. Maas [40]) :

1. Cunanthidae. With broad gastric pouches which are simple, i.e. undivided, and " pernemal," i.e. correspond in position with the tentacles. Cunina (fig. 66)

with more than eight tentacles; Cunoctantha with eight tentacles, four perradial, four interradial.

2. Aeginidae. Radii a multiple of four, with radial gastric pouches bifurcated or subdivided; the tentacles are implanted in the notch between the two subdivisions of each (primary) gastric pouch, hence the (secondary) gastric pouches appear to be " internemal " in position, i.e. to alternate in position with the tentacles. Aegina, with four tentacles and eight pouches; Aeginura (fig. 25), with eight tentacles and sixteen pouches; Solmundella (fig. 67) , with two tentacles and eight pouches ; A eginopsis (fig. 23), with two or four tentacles and sixteen pouches.

3. Solmaridae. No gastric pouches; the numerous tentacles arise direct from the stomach, into which also the peronial canals open, so that the ring-canal is cut up into separate festoons. Solmaris, Pegantha, Polyxenia, etc. To this family should be referred, probably, the genus Hydroctena, described by C. Dawydov [lla] and regarded by him as intermediate between Hydromedusae and Ctenophora. See O. Maas [35].

Appendix to the Trachylinae.

Of doubtful position, but commonly referred to the Trachylinae, are the two genera of fresh-water medusae, Limnocodium and Limnocnida.

Limnocodium sowerbyi was first discovered in the Victoria regia tank in the Botanic Gardens, Regent's Park, London. Since then it has been discovered in other botanic gardens in various parts of Europe, its two most recent appearances being at Lyons (1901) and Munich (1905), occurring always in tanks in which the Victoria regia is cultivated, a fact which indicates that tropical South America is its original habitat. In the same tanks a small hydroid, very similar to Mtcrohydra, has been found, which bears medusa-buds and is probably the stock from which the medusa is budded. It is a remarkable fact that all specimens of Limnocodium hitherto seen have been males ; it may be inferred from this either that only one polypstock has been introduced into Europe, from which all the medusae seen hitherto have been budded, or perhaps that the female medusa is a sessile gonophore, as in Pennaria. The male gonads are carried on the radial canals.

Limnocnida tanganyicae was discovered first in Lake Tanganyika, but has since been discovered also in Lake Victoria and in the river Niger. It differs from Limnocodium in having practically no manubrium but a wide mouth two-thirds the diameter of the umbrella across. It buds medusae from the margin of the mouth in May and June, and in August and September the gonads are formed in the place where the buds arose. The hydroid phase, if any, is not known.

Both these medusae have sense-organs of a peculiar type, which are said to contain an endodermal axis like the sense-organs of Trachylinae, but the fact has recently been called in question for After O. Maas, Craspedolen Medusen dtr Sitoga Expedition, by permission of E. S. Brill & Co.

FIG. 67. Solmundella bitentaculata (Quoy and Gaimard).

Limnocodium by S. Goto, who considers the genus to be allied to Olindias. Allman, on the other hand, referred Limnocodium to the Leptomedusae.

In this connexion must be mentioned, finally, the medusae budded from the fresh-water polyp Microhydra. The polyp-stages of Limnocodium and Microhydra are extremely similar in character. In both cases the hydranth is extremely reduced and has no tentacles, and the polyp forms a colony by budding from the base. In Limnocodium the body secretes a gelatinous mucus to which adhere particles of mud, etc., forming a protective covering. In Microhydra no such protecting case is formed. In view of the great resemblance between Microhydra and the polyp of Limnocodium, it might be expected that the medusae to which they give origin would also be similar. As yet, however, the medusa of Microhydra has only been seen in an immature condition, but it shows some well-marked differences from Limnocodium, especially in the 'structure of the tentacles, which furnish useful characters for distinguishing species amongst medusae. The possession of a polyp-stage by Limnocodium and Microhydra furnishes an argument against placing them in the Trachylinae. Their sense-organs require renewed investigations. (Browne [10] and [100].)

ORDER VI. Siphonophora. Pelagic floating Hydrozoa with great differentiation of parts, each performing a special function; generally regarded as colonies showing differentiation of individuals in correspondence with a physiological division of labour.

A typical Siphonophore is a stock or cormus consisting of a number of appendages placed in organic connexion with one another by means of a coenosarc. The coenosarc does not differ in structure from that already described in colonial Hydrozoa. It consists of a hollow tube, or tubes, of which the wall is made up of the two body-layers, ectoderm and endoderm, and the cavity is a continuation of the digestive cavities of the nutritive and other appendages, i.e. of the coelenteron. The coenosarc may consist of a single elongated tube or stolon, forming the stem or axis of the cormus on which, usually, the appendages are arranged in groups termed cormidia; or it may take the form of a compact mass of ramifying, anastomosing tubes, in which case the cormus as a whole has a compact form and cormidia are not distinguishable. In the Disconectae the coenosarc forms a spongy mass, the " cenlradenia," which is partly hepatic in function, forming the socalled liver, and partly excretory.

The appendages show various types of form and structure corresponding to different functions. The cormus is always differentiated into two parts; an upper portion termed the nectosome, in which the appendages are FIG. 68. Diagram showing possible modifications of medusiform line represents endoderm, the thinner line ectoderm. (After Allman.) n, Pneumatocyst.

s^sjsssk ; comotor , r hydrostatk in Generative medusiform person, function, that is to say, serve Palpon with attached palpacle, h. for swimming or floating; Siphon with branched grappling anc i a i ower portion termed g^ the siphosome, bearing ap- pendages which are nutritive, reproductive or simply protective in function.

Divergent views have been held by different authors both as regards the nature of the cormus as a whole, and as regards the homologies of the different types of appendages borne by it.

The general theories of Siphonophoran morphology are discussed i, < e, m, below, but in enumerating the various types of appendages it is convenient to discuss their morphological interpretation at the same time.

In the nectosome one or more of the following types of appendage occur:

1. Swimming-bells, termed nectocalyces or nectophores (fig. 68, k), absent in Chondrophorida and Cystophorida ; they are contractile and resemble, both in appearance, structure and function, the umbrella of a medusa, with radial canals, ring-canal and velum; but they are without manubrium, tentacles or sense-organs, and are always bilaterally symmetrical, a peculiarity of form related with the fact that they are attached on one side to the stem. A given cormus may bear one or several nectocalyces, and by their contractions they propel the colony slowly along, like so many medusae harnessed together. In cases where the cormus has no pneumatophore the topmost swimming bell may contain an oil-reservoir or oleocyst.

2. The pneumatophore or air-bladder (fig. 68, ), for passive forming a locomotion, Zoology. float which After A. Agassiz, from Lankester's Treatise on FIG. 69. Porpita, seen from above, keeps the cormus at or showing the pneumatophore and exnear the surface of the panded palpons. water. The pneumatophore arises from the ectoderm as a pit or invagination, part of which forms a gas-secreting gland, while the rest gives rise to an air-sack lined by ft chitinous cuticle. The orifice of invagination forms a pore which may be closed up or may form a protruding duct or funnel. As in the analogous swim-bladder of fishes, the gas in the pneumatophore can be secreted or absorbed, whereby the specific gravity of the body can be diminished or increased, so as to cause it to float nearer the surface or at a deeper level. Never more than one pneumatophore is found in a cormus, and when present it is always situated at the highest point above the swimming bells, if these are present also. In Velella the pneumatophore becomes of complex structure and sends air-tubes, lined by a chitin and resembling tracheae, down into the compact coenosarc, thus evidently serving a respiratory as well as a hydrostatic function.

Divergent views have been held as to the morphological significance of the pneumatophore. E. Haeckel regarded the whole structure as a glandular ectodermal pit formed on the exumbral surface of a medusa-person. C. Chun and, more recently, R. Woltereck [59], on the other hand, have shown that the ectodermal pit which gives rise to the pneumatophore represents an entocodon. Hence the cavity of the air-sack is equivalent to a sub-umbral cavity in which no manubrium is formed, and the pore or orifice of invagination would represent the margin of the umbrella. In the wall of the sack is a double layer of endoderm, the space between which is a continuation of the coelenteron. By coalescence of the endoderm-layers, the coelenteron may be reduced to vessels, usually eight in number, opening into a ring-sinus surrounding the pore. Thus the disposition of the endoderm-cavities is roughly comparable to the gastrovascular system of a medusa.

The difference between the theories of Haeckel and Chun is connected with a further divergence in the interpretation of the stem or axis of the cormus. Haeckel regards it as the equivalent of the manubrium, and as it is implanted on the blind end of the pneumatophore, such a view leads necessarily to the air-sack and gland being a development on the ex-umbral surface of the medusa-person. Chun and Woltereck, on the other hand, regard the stem as a stole prolifer arising from the aboral pole, that is to say, from the ex-umbrella, similar to that which grows out from the ex-umbral surface of the embryo of the Narcomedusae and produces buds, a view which is certainly supported by the embryological evidence to be adduced shortly.

In the siphosome the following types of appendages occur:

1. Siphons or nutritive appendages, from which the order takes its name; never absent and usually present in great numbers (fig. 68, e). Each is a tube dilated at or towards the base and containing a mouth at its extremity, leading into a stomach placed in the dilatation already mentioned. The siphons have been compared to the manubrium of a medusa-individual, or to polyps, and hence are sometimes, termed gastrozoids.

2. Palpons (fig. 68, g), present in some genera, especially in Physonectae ; similar to the siphons but without a mouth, and purely tactile in function, hence sometimes termed dactylozoids. If a distal pore or aperture is present, it is excretory in function; such, varieties have been termed " cystons " by Haeckel.

3. Tentacles (" Fangfdden "), always present, and implanted one at the base of each siphon (fig. 68,/). The tentacles of siphonophores may reach a great length and have a complex structure. They may bear accessory filaments or tentilla (/'), covered thickly with batteries of nematocysts, to which these organisms owe their great powers of offence and defence.

4. Palpacles (" Tastfaden "), occurring together with palpons, one implanted at the base of each palpon (fig. 68, h). Each palpacle is a tactile filament, very extensile, without accessory filaments or nematocysts.

5. Bracts (" hydrophyllia "), occur in Calycophorida and some Physophorida as scale-like appendages protecting other parts (fig. 68, /). The mesogloea is greatly developed in them and they are often of very tough consistency. By Haeckel they are considered homologous with the umbrella of a medusa.

6. Gonostyles, appendages which produce by budding medusae or gonophores, like the blastostyles of a hydroid colony. In their most primitive form they are seen in Velella as " gonosiphons," which possess mouths like the ordinary sterile siphons and bud free medusae. In other forms they have no mouths. They may be branched, so-called " gonodendra," and amongst them may occur special forms of palpons, " gonopalpons." The gonostyles have been compared to the blastostyles of a hydroid colony, or to the manubrium of a medusa which produces free or sessile medusa-buds.

7. Gonophores, produced either on the gonostyles already mentioned or budded, as in hydrocorallines, from the coenosarc, i.e. the stem (fig. 68, *'.). They show every transition between free medusae and sporosacs, as already described for hydroid colonies.

Thus in Velella free medusae are produced, which have been described as an independent genus of medusae, Chrysomitra. In other types the medusae may be set the appendages are arranged as regularly recurrent cormidia along it, and the cormidia are then said to be " ordinate." In such cases the oldest cormidia, that is to say, those furthest from the nectosome, may become detached (like the segments or proglottides of a tape-worm) and swim off, each such detached cormidium then becoming a small free cormus which, in many cases, has been given an independent generic name. A cormidium may contain a single nutritive siphon (" monogastric ") or several siphons (" polygastric "). From G - H - Fowler, after G.

The following are some of the forms uvier, Lankester's Treatise on of cormidia that occur :

1. The eudoxome (Calycophorida), con- FIG. 71. Upper sursisting of a bract, siphon, tentacle and face of Velella, showing gonophore; when free it is known as pneumatophore and sail. Eudoxia.

2. The ersaeome (Calycophorida), made up of the same appendages as the preceding type but with the addition of a nectocalyx; when free termed Ersaea.

3. The rhodalome of some Rhodalidae, consisting of siphon, tentacle and one or more gonophores.

4. The athorome of Physophora, etc., consisting of siphon, tentacle, one or more palpons with palpacles, and one or more gonophores.

5. The crystallpme of Anthemodes, etc., similar to the athorome but with the addition of a group of bracts.

Embryology of the Siphonophora. The fertilized ovum gives rise to a parenchymula, with solid endoderm, which is set free as a free- swimming planula larva, in the manner already described (see HYDROZOA). The planula has its two extremities dissimilar (Bipolaria-larva). The subsequent development is slightly different according as the future cormus is headed by a pneumatophore ( Physophorida, Cystophorida) or by a nectocalyx ( Calycophorida).

(i.) Physophorida, for example Halistemma (C. Chun, HYDROZOA i]). The planula becomes elongated and broader towards one pole, at which a pit or invagination of the ectoderm arises. Next the pit closes up to form a vesicle with a pore, and so gives rise to the pneumatophore. From the broader portion of the planula an From G. H. Fowler, after A. Agassiz, Lankester's Treatise on Zoology.

FIG. 70. Diagram of the structure of Velella, showing the central and outgrowth arises peripheral thirds of a half-section of the colony, the middle third being ^ lch becom es the omitted. The ectoderm is indicated by close hatching, the endoderm by h L St tentacle light hatching, the mesogloea by thick black lines, the horny skeleton of j cormus. the pneumatophore and sail by dotting. endoderm of BL, Blastostyle. M, Medusoid gonophores. planula now ac- C, Centradenia.

D, Palpon.

EC, Edge of colony prolonged be- yond the pneumatophore. G, Cavity of the large central siphon.

free in a mature condition as the so-called " genital swimming bells," comparable to the Globiceps of Pennaria. The most usual condition, however, is that in which sessile medusoid gonophores or sporosacs are produced.

The various types of appendages described' in the foregoing may be arranged in groups termed cormidia. In forms with a compact coenosarc such as Velella, Physalia, etc., the separate cormidia cannot be sharply distinguished, and such a condition is described technically as one with r< scattered " cormidia. In forms in which, on the other hand, the coenosarc forms an elongated, tubular axis or stem, of FIG. 72. A, Diphyes campanulala; e B, a group of appendages (cormidium) of the the same Diphyes. (After C. Gegenbaur.)

PN, Primary central chamber, and qulr , es PN', concentric chamber of a the pneumatophore, showing narr wer an opening to the exterior r and a " trachea."

S, Sail.

a, Axis of the colony. Nectocalyx. Sub-umbral cavity of nectocalyx. Radial canals of nectocalyx.

Orifice of nectocalyx. Bract. Siphon. Gonophore. Tentacle.

a cavity, at the? pole a ' mouth is formed, giving rise to the ' primary siphon. Thus from the original planula three appendages are, as it were, budded off, while the planula itself mostly gives rise to coenosarc, just as in some hydroids the planula is converted chiefly into hydrorhiza.

(ii.) Calycophorida, for ; example, Muggiaea. The planula develops, on the whole, in a similar manner, but the ectodermal invagination arises, not at the pole of the planula, but on the side of its broader portion, and gives rise, not to a pneumatophore, but to a nectocalyx, the primary swimming bell or protocodon (" Fallschirm ") which is later thrown off and replaced by secondary swimming bells, metacodons, budded from the coenosarc.

From a comparison of the two embryological types there can be no doubt on two points; first, that the pneumatophore and the protocodon are strictly homologous, and, therefore if the nectocalyx is comparable to the umbrella of a medusa, as seems obvious, the pneumatophore must be so too; secondly, that the coenosarcal axis arises from the ex-umbrella of the medusa and cannot be compared to a manubrium, but is strictly comparable to the " bud-spike ' of a Narcomedusan.

Theories of Siphonophore Morphology. The many theories that have been put forward as to the interpretation of the cormus and the various parts are set forth and discussed in the treatise of Y. Delage and E. H6rouard (HYDROZOA [4]) and more recently by R. Woltereck [59], and only a brief analysis can be given here.

In the first place the cormus has been regarded as a single individual and its appendages as organs. This is the so-called " polyorgan " theory, especially connected with the name of Huxley; but it must be borne in mind that Huxley regarded all the forms produced, in any animal, between one egg-generation and the next, as constituting in the lump one single individual. Huxley, therefore, considered a hydroid colony, for example, as a single individual, and each separate polyp or medusa budded from it as having the value of an organ and not of an individual. Hence Huxley's view is not so different from those held by other authors as it seems to be at first sight.

In more recent years Woltereck [59] has supported Huxley's view of individuality, at the same time drawing a fine distinction between " individual "and" person." The individual is the product of sexual reproduction ; a person is an individual of lower rank, which may be produced asexually. A Siphonophore is regarded as a single individual composed of numerous zoids, budded from the primary zoid (siphon) produced from the planula. Any given zoid is a person-zoid if equivalent to the primary zoid, an organ-zoid if equivalent only to a part of it. Woltereck considers the siphonophores most nearly allied to the Narcomedusae, producing like the buds from an aboral stolon, the first bud being represented by the pneumatophore or protocodon, in different cases.

Contrasting, in the second place, with the polyorgan theory are the various " poly person " theories which interpret the Siphonophore cormus as a colony composed of FiG.73Physophorahydrostatica. more O r fewer individuals in or- After C. Gegenbaur.

a', Pneumatocyst. t, Palpons. a, Axis of the colony. m, Nectocalyx.

0, Orifice of nectocalyx. n, Siphon.

g, Gonophore.

1. Tentacle.

ganic union with one another. On this interpretation there is still room for considerable divergence of opinion as regards detail. To begin with, it is not necessary on the polyperson theory to regard each appendage as a distinct individual; it is still possible to compare appendages with parts of an individual which have become separated from one another by a process of " dislocation of organs." Thus a bract may be regarded, with Haeckel, as a modified umbrella of a medusa, a siphon as its manubrium, and a tentacle as representing a medusan tentacle shifted in attachment from the margin to the sub-umbrella ; or a siphon may be compared with a polyp, of which the single tentacle has become shifted so as to be attached to the coenosarc and so on. Some authors prefer, on the other hand, to regard every appendage as a separate individual, or at least as a portion of an individual, of which other portions have been lost or obliterated.

A further divergence of opinion arises from differences in the interpretation of the persons composing the colony. It is possible to regard the cormus (i) as a colony of medusa-persons, (2) as a colony of polyp-persons, (3) as composed partly of one, partly of the other. It is sufficient here to mention briefly the views put forward on this point by C. Chun and E. Haeckel.

Chun (HYDROZOA [1]) maintains the older views of Leuckart and Claus, according to which the cormus is to be compared to a floating hydroid colony. It may be regarded as derived from floating polyps similar to Nemopsis or Pelagohydra, which by budding produce a colony of polyps and also form medusa-buds. The polyp- individuals form the nutritive siphosome or trophosome. The medusabuds are either fertile or sterile. If fertile they become free medusae or sessile gonophores. If sterile they remain attached and locomotor in function, forming the nectosome, the pneumatophore and swimming-bells.

Haeckel, on the other hand, is in accordance with Balfcur in regarding a Siphonophore as a medusome, that is to say, as a colony composed of medusoid persons or organs entirely. Haeckel considers that the Siphonophores have two distinct ancestral lines of evolution :

1. In the Disconanthae, i.e. in such forms as Velella, Porpita, etc., the ancestor was an eight-rayed medusa (Disconuld) which acquired a pneumatophore as an ectodermal pit on the ex-umbrella, and in which the organs (manubrium, tentacles, etc.) became secondarily multiplied, just as they do in Gastroblasta as the result of incomplete fission. The nearest living allies of the ancestral Disconula are to be sought in the Pectyllidae.

2. In the Siphonanthae , i.e. in all other Siphonophores, the ancestral form was a Siphonula, a bilaterally symmetrical Anthomedusa Alter Haeckel, from Lankester's Treatise on Zoology.

FIG. 74. Stephalia corona, a young colony. p, Pneumatophore. /, Aurophore. s, Siphon.

n, Nectocalyx. lo, Orifice of the aurophore. t, Tentacle.

with a single long tentacle (cf. Corymorpha), which became displaced from the margin to the sub-umbrella. The Siphonula produced buds on the manubrium, as many Anthomedusae are known to do, and these by reduction or dislocation of parts gave rise to the various appendages of the colony. Thus the umbrella of the Siphonula became the protocodon, and its manubrium, the axis or stolon, which, by a process of dislocation of organs, escaped, as it were, from the sub-umbrella through a cleft and became secondarily attached to the ex-umbrella. It must be pointed out that, however probable Haeckel's theory may be in other respects, there is not the slightest evidence for any such cleft in the umbrella having been present at any time, and that the embryological evidence, as already pointed out, is all against any hpmology between the stem and a manubrium, since the primary siphon does not become the stem, which arises from the ex-umbral side of the protocodon and is strictly comparable to a stolon.

Classification. The Siphonophora may be divided, following Delage and Herouard, into four sub-orders:

I. CHONDROPHORIDA (Disconeclae Haeckel, Tracheophysae Chun). With an apical chambered pneumatophore, from which tracheal tubes may take origin (fig. 70); no nectocalyces or bracts; appendages all on the lower side of the pneumatophore arising from a compact coenosarc, and consisting of a central principal siphon, surrounded by gonosiphons, and these again by tentacles.

Three families: (i) Discalidae, for Discalia and allied genera, deepsea forms not well known; (2) Porpitidae for the familiar genus Porpita (fig. 69) and its allies ; and (3) Velellidae, represented by the well-known genus Velella (figs. 70, 71), common in the Mediterranean and other seas.

II. CALYCOPHORIDA (Cdyconectae, Haeckel). Without pneumatophore, with one, two, rarely more nectocalyces.

Three families: (i) Monophyidae, with a single nectocalyx; examples Muggiaea, sometimes found in British seas, Sphaeronectes, etc. ; (2) Diphyidae, with two nectocalyces; examples Diphyes (fig. 72), Praya, Abyla, etc. ; and (3) Polyphyidae, with numerous nectocalyces ; example Hippupodius, Stephanophyes and other genera.

From O. H. Fowler, modified after G. CUVUT and E. Haeckel, Lankestcr's Treatise on Zoology.

FIG. 75. A. Physalia, general view, diagrammatic; B, cormidium of Physalia; D, palpon; T, palpacle; G, siphon; GP, gonopalpon ; M <f , male gonophore ; M 9 , female gonophore, ultimately set free.

III. PHYSOPHORIDA (Physonectae + Auronectae, Haeckel). With an apical pneumatophore, not divided into chambers, followed by a series of nectocalyces or bracts.

A great number of families and genera are referred to this group, amongst which may be mentioned specially (i) Agalmidae, containing the genera Stephanomia, Agalma, Anthemodes, Halistemma, etc. ; (2) Apolemidae, with the genus Apolemia and its allies; (3) Forskalitdae, with Forskalia and allied forms; (4) Physophoridae, for Physophora (fig. 73) and other genera, (5) Anthophysidae, for Anthophysa, Athorybia, etc. ; and lastly the two families (6) Rhodalidae and (7) Stephalidae (fig. 74), constituting the group Auronectae of Haeckel. The Auronectae are peculiar deep-sea forms, little known except from Haeckel's descriptions, in which the large pneumatophore has a peculiar duct, termed the aurophore, placed on its lower side in the midst of a circle of swimming-bells.

IV. CYSTOPHORIDA (Cyslonectae, Haeckel). With a very large pneumatophore not divided into chambers, but without nectocalyces or bracts. Two sections can be distinguished, the Rhizophysina, with long tubular coenosarc-bearing ordinate cormidia, and Physalina, with compact coenosarc-bearing scattered cormidia.

A type of the Rhizophysina is the genus Rhizophysa. The Physalina comprise the families Physalidae and Epibulidae, of which the types are Physalia (figs. 74, 75) and Epibulia, respectively. Physalia, known commonly as the Portuguese man-of-war, is remarkable for its great size, its brilliant colours, and its terrible stinging powers.

BIBLIOGRAPHY. In addition to the works cited below, see the general works cited in the article HYDROZOA, in some of which very full bibliographies will be found.

1. G. J.Allman," A Monograph of the Gymnoblastic or Tubularian Hydroids," Ray Society (1871-1872); 2. A. Brauer, "Uber die Entwickelung von Hydra," Zeitschr.f. wiss. Zool. Iii. (1891), pp. 169- 216, pis. ix.-xii. ; 3. tJber die Entstehung der Geschlechtsprodukte und die Entwickelung von Tubularia mesembryanthemum Allm.," i.e. pp. 551-579, pis. xxxiii.-xxxv.; 4. W. K. Brooks, "The LifeHistory of the Hydromedusae: a discussion of the Origin of the Medusae, and of the significance of Metagenesis," Mem. Boston Soc. Nat. Hist. iii. (1886), pp. 259-430, pis. xxxvii.-xliv. ; 5. " The Sensory Clubs of Cordyli of Laodice," Journ. Morphology, x. (1895), pp. 287-304, pi. xvii.; 6. E. T. Browne, " On British Hydroids and Medusae," Proc. Zool. Soc. (1896), pp. 459-500, pis. xvi., xvii., (1897), pp. 816-835, pis. xlviii. xlix. 12 text-figs.; 7. " Biscayan Medusae," Trans. Linn. Soc. x. (1906), pp. 163-187, pi. xiii. ; 8. " Medusae " in Herdman, Rep. Pearl Oyster Fisheries, Gulf of Manaar, iv. (1905), pp. 131-166, 4 pis.; 9. "Hydromedusae with a Revision of the Williadae and Petasidae," Fauna and Geogr. Maldive and Laccadive Archipelagos, ii. (1904), pp. 722-749, pis. liv.-lvii.; 10. "On the Freshwater Medusa liberated by Microhydra ryderi, Potts, and a Comparison with Limnocodium," Quart. Journ. Micr. Sci. i (1906), pp. 635, 645, pi. xxxvii.; lOa. " On the Freshwater Medusa Limnocnida tanganicae " Budgett Memorial Volume (Cambridge, 1908, pp. 471-482, pi. xxviii.; 11. C. Claus, " Uber die Struktur der Muskelzellen und uber den Korperbau von Mnestra parasites Krohn," Verhandl. zoo/, hot. Ges. Wien, xxy. (1876), pp. 9-12, pi. i. ; lla. C. Dawydov, " Hydroctena salenskii," Mem. Acad. Imp. St. P6tersbourg (viii.) xiv. No. 9 (1903), 17 pp., I pi.; 12. A. Dendy, " On a Free-swimming Hydroid, Pelagohydra mirabilis," n. gen. et sp.. Quart. Journ. Micr. Sci. xlvi. (1903), pp. 1-24, pis. i. ii.; 13. H. Driesch, " Tektonische Studien an Hydroidpolypen," (i) Jen. Zeitschr., xxiv. (1890), pp. 189-226, 12 figs.; (2) I.e. pp. 657-688, 6 figs.; (3) ibid. xxv. (1891), pp. 467-479, 3 figs.; 14. G. Duplessis, "On Campanularia_volubilis,' Soc. Vaud. Bull. 13 (Lausanne, : ~ 1875); IS. J. W. Fewkes, "On Mnestra," Amer. Natural., xviii. (1884), pp. 197-198, 3 figs.; 16. S. Goto, " Dendrocoryne Inaba, Vertreterin einer neuen Familie der Hydromedusen," Annot. Zool. Tokyo, i. (1897), pp. 93-104, pi. vi., figs. 106-113; 17. " The Craspe dote Medusa Olindias and some of its Natural Allies," Mark Anniversary Volume (New York, 1903), pp. 1-22,3 pis.; 18. H. Grenadier, " Uber die Nesselkapseln von Hydra," Zool. Anz. xviii. (1895), PP- 310-321, 7 figs.; 19. R. T. Gilnther, "On the Structure and Affinities of Mnestra parasites Krohn ; with a revision of the Classification of the Cladonemidae," Mitt. Stat. Neapel, xvi. (1903), pp. 35-62, pis. ii. iii.; 20. E. Haeckel, "Das System der Medusen," Denkschr. med.-nat.-wiss. Ges. (Jena, 1879-1881); 21. "Deep Sea Medusae," in Reports of the Challenger Expedition, Zool. iv. pt. 2 (London, 1882); 22. P. Hallez, " Bougainvillia fruticosa Allm. est le fades d'eau agitee du Bougainvillia ramosa Van Ben." C.-R. Acad. Sci. Paris, cxl. (1905), pp. 457-459; 23. O. & R. Hertwig, Der Organismus der Medusen (Jena, 1878), 70 pp., 3 pis.; 24. Das Nervensystem und die Sinnesorgane der Medusen (Leipzig, 1878), 186 pp., 10 pis.; 25. S. J. Hickson, " The Medusae of Millepora," Proc. Roy. Soc. Ixvi. (1899), pp. 6-10, 10 figs.; 26. T. Hincks, A History of British Hydroid Zoophytes (2 vols., London, 1868); 27. N. Iwanzov, " Uber den Bau, die Wirkungsweise und die Entwickelung der Nesselkapseln von Coelenteraten," Bull. Soc. Imp. Natural. Moscou (1896), pp. 323-355, 4 pis. ; 28. C. F. Jickeli, " Der Bau der Hydroidpolypen, (i) Morph. Jahrbuch, viii. (1883), pp. 373-416, pis. xvi.- Lang, " Gastroblasta Raffaelei. " "Eine" durch eine Art unvollstandiger Theilung entstehende Medusen-Kolonie," Jena Zeitschr. xix. (1886), pp. 735-762, pis. xx., xxi.; 31. A. Linko, " Observations sur les mdduses de la mer Blanche," Trav. Soc. Imp. Nat. St Pelersbourg, xxix. (1899); 32. " Uber den Bau der Augen bei den Hydromedusen," Zapiski Imp. Akad. Nauk (Mem. Acad. Imp. Set.) St Petersburg (8) x. 3 (1900), 23 pp., 2 pis. ; 33. O. Maas, " Die craspedoten Medusen," in Ergebn. Plankton Expedition, ii. (Kiel and Leipzig, 1893), 107 pp., 8 pis., 3 figs.; 34. " Die Medusen," Mem. Mus. Cpmp. Zool. Harvard, xxiii. (1897), i.; 35. "On Hydroctena," Zool.

. , ... 240-243; 36. "Revision des meduses appartenant aux families des Cunanthidae et des Aeginidae, et groupement nouveau des genres," Bull. Mus. Monaco, v. (1904), 8 pp.; 37. " Revision der Cannotiden Haeckels," SB. K. Bayer. Akad. xxxiv. (1904), pp. 421-445; 38. "Meduses," Result. Camp.

Sci. Monaco, xxviii. (1904), 71 pp., 6 pis.; 39. " Die craspedoten Medusen der Siboga-Expedition, Uitkomst. Siboga-Exped. x. (1905), 84 pp., 14 pis.; 40. " Die arktischen Medusen (ausschliesslich der Polypomedusen)," Fauna arctica, iv. (1906), pp. 479-526; 41. C. Mereschkowsky, " On a new Genus of Hydroids (Monobrachium) from the White Sea, with a short description of other new Hydroids," Ann. Mag. Nat. Hist. (4) xx. (1877), pp. 220-229, pis. v. vi.; 42.

E. Metchmkoft, " Studien uber die Entwickelung der Medusen und Siphonophoren," Zeitschr. f. wiss. Zool. xxiv. (1874), pp. 15-83, pis. i.-xii.; 43. " Vergleichend-embryologische Studien" (Geryoniden, Cunina), ibid, xxxvi. (1882), pp. 433-458, pi. xxviii.; 44. Embryologische Studien an Medusen (Vienna, 1886), 150 pp., 12 pis., 10 figs.; 45. " Medusologische Mittheilungen," Arb. zool. Inst. Wien, vi. (1886), pp. 237-266, pis. xxii. xxiii. ; 46. L. Murbach, " Beitrage zur Kenntnis der Anatomie und Entwickelung der Nesselorgane der Hydroiden," Arch. f. Naturgesch. Ix. i. (1894), pp. 217-254, pi. xii. ; 47. " Preliminary Note on the Life-History of Uonionemus ,' Journ. Morph. xi. (1895), pp. 493-496; 48. L. Murbach and C. Shearer, " On Medusae from the Coast of British Columbia and Alaska," Proc. Zool. Soc. (1903), ii. pp. 164-191, pis. xvii.-xxii. ; 49. H. F. Perkins, "The Development of Gonionema murbachii," Proc.Acad. Nat. Sci. Philadelphia (1902), pp. 750-790, pis. xxxi-xxxiv.; 50.

F. Schaudinn, " Uber Haleremita cumulans, n. g. n. sp., einen marinen Hydroidpolypen," SB. Ges. nalforsch. Freunde Berlin (1894), pp. 226-234, 8 figs. ; 51. F. E. Schulze, " On the Structure and Arrangement of the Soft Parts in Euplectella aspergtilum " ( Amphibrachium), Tr. R. Soc. Edinburgh, xxix. (1880), pp. 661-673, pi. xvii. ; 52. O. Seeliger, " Uber das Verhalten der Keimblatter bei der Knospung der Colenteraten," Zeitschr. f. wiss. Zool. Iviii. (1894), pp. 152-188, pis. vii.-ix.; 53. W. B. Spencer, "A new Family of Hydroidea (Clalhrozoon) , together with a description of the Structure of a new Species of Plumularia," Trans. Roy. Soc. Victoria (1890), pp. 121-140, 7 pis.; 54. M. Ussow, " A new Form of Fresh-water Coelenterate " (P ply podium) , Ann. Mag. Nat. Hist. (5) xviii. (1886), pp. 110-124, pi. iv. ; 55. E. Vanhoffen, " Versuch einer natiirlichen Gruppierung der Anthomedusen," Zool. Anzeiger, xiv. (1891), pp. 439-446; 56. C. Viguier, " Etudes sur les animaux inferieurs de la baie d'AIger " (Tetraplatia) , Arch. Zool. Exp. Gen. viii. (1890), pp. 101-142, pis. vii.-ix.; 57. J. Wagner, " Recherches sur I' orgamsation de Monobrachium parasiticum Merejk," Arch. biol. x. (1890), pp. 273-309, pis. viii. ix.; 58. A. Weismann, Die Entstehung der Sexualzellen bei den Hydromedusen (Jena, 1883); 59. R. Woltereck, " Beitrage zur Ontogenie und Ableitungdes Siphonophorenstocks," Zeitschr. f. wiss. Zool. Ixxxii. (1905), pp. 611-637, 21 text-figs.; 60. J. Wulfert, " Die EmbryonalentwickelunK von Gonothyraea loveni Allm.," Zeitschr. f. wiss. Zool. Ixxi. (1902), pp. 296-326, pis. xvi.-xviii. (E. A. M.)

Note - this article incorporates content from Encyclopaedia Britannica, Eleventh Edition, (1910-1911)

Privacy Policy | Cookie Policy | GDPR