Chapter V: V V (3)

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 polyp-colony, 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 actinula-stage. 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 "egg," and 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 [1]. Pl. I.).

a, Hydrocaulus (stem).
b, Hydrorhiza (root).
c, Enteric cavity.
d, Endoderm.
e, Ectoderm.
f, 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.
l, Medusiform person or medusa.
m, Blastostyle.]

_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 are free and walk
on their tentacles. See Ussow [54].

_Tetraplatia volitans_ Viguier is a remarkable floating marine form.
See C. Viguier [56] and Delage and Hérouard (Hydrozoa [2]).

_Haleremita_ Schaudinn. See F. Schaudinn [50] and Delage and Hérouard
(HYDROZOA [2]).

In all the above-mentioned 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.

ORDER II. Hydroidea seu Leptolinae.--Hydromedusae with alternation of generations (metagenesis) in which a non-sexual polyp-generation (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 shelf-and-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 1. 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 hydrorhiza 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 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_,
&c.

(b) Tentacles capitate, simple; type of _Coryne_ and _Syncoryne_;
_Myriothela_ is an aberrant form with some of the tentacles modified
as "claspers" to hold the ova.

(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_, &c.

(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_, &c.

(b) With capitate tentacles; type of _Clavatella_.

4. Hydranth with tentacles reduced below four; type of _Lar_ (fig.
11), _Monobrachium_, &c.

The _Anthomedusa_ 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
Hérouard (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
Hérouard. 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
well-characterized families and others of more doubtful nature.
Certain discrepancies must also be noted.

1. _Margelidae_ (= medusa-family _Margelidae_ + hydroid families
_Bougainvillidae_, _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
_Margellium_. Other hydroids are _Garveia_, _Bimeria_, _Eudendrium_
and _Heterocordyle_, with gonophores, and _Dicoryne_ with peculiar
sporosacs.

FIG. 52.--_Tiara pileata_, L. Agassiz.]

2. _Podocorynidae_ (= medusa-families _Thamnostomidae_ and _Cytaeidae_
+ hydroid families _Podocorynidae_ and _Hydractiniidae_). Trophosome
encrusting with hydranths of _Bougainvillea_-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_, &c., are other medusae with
unknown hydroids. _Hydractinia_ (figs. 9, 10) is a familiar hydroid
genus, bearing gonophores.

3. _Cladonemidae_.--Trophosome, 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._--Trophosome, 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_.--Trophosome, 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.

7. _Corymorphidae_ (including the medusa-family
_Hybocodonidae_).--Trophosome solitary polyps, with two whorls of
tentacles; gonosome, free medusae or gonophores. _Corymorpha_ (fig.
3), a well-known British genus, sets free a medusa known as
_Steenstrupia_ (fig. 22). Here belong the deep-sea genera _Monocaulus_
and _Branchiocerianthus_, including the largest hydroid polyps known,
both genera producing sessile gonophores.

FIG. 53.--_Pteronema darwinii_. The apex of the stomach is prolonged
into a brood pouch containing embryos.]

8. _Dendroclavidae._--Trophosome, 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
51). 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. _Turris_
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
with capitate tentacles, simple or branched, scattered or
verticillate; gonosome, free medusae or gonophores. _Coryne_, a common
British hydroid, produces gonophores; _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. 11) 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 colony-forming
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, _Dendrocorynidae_.

Doubtful families, or forms difficult to classify, are: Pteronemidae,
Medusae of Cladonemid type, with hydroids for the most part unknown.
The British genus _Gemmaria_, 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 [1]).

_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 the tentacles in four tufts of (in
the buds) five each, and thus resemble the medusae of the family
_Margelidae_. See A. Dendy [12].

_Perigonimus._--This common British hydroid belongs by its 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 CALYPTOBLASTEA (LEPTOMEDUSAE).--Trophosome with polyps always differentiated into nutritive and reproductive individuals (blastostyles) enclosed in hydrothecae and gonothecae respectively; with sympodial type of budding. Gonosome with free medusae or gonophores; the medusae typically with otocysts, sometimes with cordyli or ocelli (figs. 54, 55).

ge, Genital glands.
M, Manubrium.
ot, Otocysts.
rc, The four radiating canals.
Ve, The velum.]

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; if the latter, the budding of the polyps, as already stated, is of the sympodial type, and either biserial, 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 1st, 3rd, 5th, &c., members of the stem, and retarded in the 2nd, 4th, 6th, &c., 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_.

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.]

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 _Plumularidea_. 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 halecinum_, 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 gonotheca is formed in its early stage in the same way as the hydrotheca, 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).

The blastostyles, gonophores and gonothecae furnish a series of
variations which can best be considered as so many stages of
evolution.

Stage 1, 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, 19).

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 the
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), enclosing
and protecting the embryos. Then the spadix withers, leaving the
embryos in the acrocyst, which may be further protected by a so-called
marsupium, a structure formed by tentacle-like processes 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, pl. ix.,
figs. 7, 8.)

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. 66, fig. 30).

Fig. 57.--Diagrams to show the mode of formation of an Acrocyst and a Marsupium. In A two medusa-buds are seen within the gonotheca (gth), the upper more advanced than the lower one. In B the spadix of 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, enclosing the acrocyst; b, medusa-buds on the blastostyle.]

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 another below the 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
_Thaumantias_ 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_, &c.

FIG. 58.--_Octorchandra canariensis_, from life.]

5. _Polyorchidae._--Trophosome unknown; gonosome, free medusae of deep
form, with radial canals branched in a feathery manner, and 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_, &c.

7. _Lafoëidae._--Trophosome as in the preceding; gonosome, free
medusae or gonophores, the medusae with large open otocysts. The
hydroid genus _Lafoëa_ 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, &c.).

(So far as the characters of the trophosome are concerned, the seven
preceding families are scarcely distinguishable, and they form 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.)

FIG. 59.--Three stages in the development of _Phialidium temporarium_.
a, The youngest stage, is magnified about 22 diam.; b, older, is
magnified about 8 diam.; c, the adult medusa, is magnified.]

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, containing hydranths each with a single
tentacle, and numerous nematophores. See W. B. Spencer [53].

11. _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 perforated by pores of two kinds, gastropores and dactylopores, giving exit to gastrozoids and dactylozoids respectively, which are lodged in vertical pore-canals of wider calibre than the coenosarcal canals of the general network. 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 polyp-canal) 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, 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).

1, _Sporadopora dichotoma_.
2, 3, _Allopora nobilis_.
4, _Allopora profunda_.
5, _Allopora miniacea_.
6, _Astylus subviridis_.
7, _Distichopora coccinea_.
s, Style.
dp, Dactylopore.
gp, Gastropore.
b, In fig. 6, inner horseshoe-shaped mouth of gastropore.]

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 _Milleporidae_ they have short capitate tentacles, but lack tentacles in _Stylasteridae_.

The gonosome consists of free medusae in _Milleporidae_, which are budded from the apex of a dactylozoid in _Millepora murrayi_, but in other species from the coenosarcal 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 _Stylasteridae_ sessile gonophores are formed, always by budding from the coenosarc. In _Distichopora_ 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 operculum. In _Astylus_ (fig. 63) styles are absent.

_Affinities of the Hydrocorallinae._--There can be no doubt that the
forms comprised in this order bear a close relationship to the
Hydroidea, especially the sub-order Gymnoblastea, with which they
should perhaps be classed in a natural classification. A
hydrocoralline may be regarded as a form of hydroid colony in which
the coenosarc 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. Graptolitoidea (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, air-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 Hérouard (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 1. 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 plant-parenchyma, 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.

FIG. 64. _Olindias mülleri._]

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. _Petasus_ 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 mülleri_ (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; tentaculocysts enclosed. _Rhopalonema_,
_Trachynema_, &c.

FIG. 65.--_Aglantha rosea_ (Forbes), a British medusa.]

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), &c., 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_, &c., with four radial canals; _Geryonia_, _Carmarina_ (fig.
26), &c., with six.

7. _Halicreidae._--Eight very broad radial canals; ex-umbrella often
provided with lateral outgrowths; tentacles differing in size, but in
a single row. _Halicreas_.

SUB-ORDER 2. NARCOMEDUSAE.--Margin of the umbrella-lobed, 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.

c, Circular canal.
h, "Otoporpae" or centripetal process of the marginal cartilaginous
ring connected with tentaculocyst.
k, Stomach.
l, Jelly of the disk.
r, Radiating canal (pouch of stomach).
tt, Tentacles.
tw, Tentacle root.]

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 to the insertion of the 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_ (_Gehörspangen_). 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 octonaria_ 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" (_Knospenähre_), 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]):

FIG. 67.--_Solmundella bitentaculata_ (Quoy and Gaimard).]

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; _Aeginopsis_ (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_, &c. To this family should be referred,
probably, the genus _Hydroctena_, described by C. Dawydov [11a] 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
_Microhydra_, 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 polyp-stock 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
_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, &c.,
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 [10a].)

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.

n, Pneumatocyst.
k, Nectocalyces (swimming bells).
l, Hydrophyllium (covering-piece).
i, Generative medusiform person.
g, Palpon with attached palpacle, h.
e, Siphon with branched grappling tentacle, f.
m, Stem.]

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 "_centradenia_," which is partly hepatic in function, forming the so-called 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 locomotor or hydrostatic in function, that is to say, serve for swimming or floating; and a lower portion termed the _siphosome_, bearing appendages 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 below,
but in enumerating the various types of appendages it is convenient to
discuss their morphological interpretation at the same time.

FIG. 69.--_Porpita_, seen from above, showing the pneumatophore and
expanded palpons.]

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, n), for passive
locomotion, forming a float which keeps the cormus at or near the
surface of the 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 a 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 ex-umbral 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 _stolo 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:--