Its ectoderm developes a circular and a longitudinal layer of muscle cells, its endoderm a circular.

The colonies of Physophoridae are mostly of moderate size, but Agalma elegans attains a length of four feet, Apolemia uvaria of twenty to thirty feet. The Calycophoridae are small, the largest, e. g. Diphyes, is a few inches long, Galeolaria two feet or less. The colony is retractile within a furrow or canal formed at the side of the distal nectocalyx in Diphyes and Abyla; by the two nectocalyces in Praya; or into a pit in the side of the single nectocalyx in Monophyidae. The Siphonophora are found in the Mediterranean and the open seas, sometimes at considerable depths (eight hundred to fifteen hundred fathoms or even more). Physalia is brilliantly coloured: Velella and Porpita are greenish blue. The majority are hyaline with spots of colour on the float, the polypites, hydrocysts, and tentacles.

The ovum appears to be impregnated and to develope while floating freely. It is devoid of a membrane save in Hippopodius, Segmentation is regular, and results in the formation of a solid ciliated planula with large vacuolated central cells. The permanent hypoblast is differentiated from this central mass of cells as a superficial layer of cells beneath the epiblast, the central cells undergoing gradual absorption. In the Calycophorid Galeolaria filiormis=Epibulia aurantiaca, a nectocalyx, a polypite, tentacle, hydrophyllium are formed. The second nectocalyx appears on the pedicle of the first group of zooids, i. e. on the developing coenosarc. The buds of a second polypite, tentacle, hydrophyllium, grow out, and the colony is established by the development of successive groups of zooids in the order named. The cavity of the first nectocalyx in Hippopodius is remarkably small, but its mesoglaea extends as a cap over one extremity of the planula. A provisional hydrophyllium is developed in the first instance in most Physophoridae,-Athorybia, Physophora, Agalmopsis, and Crystallodes, but not in Halistemma. The float appears next in succession; see note, p. 772, ante.

Agalmopsis and Crystallodes have a set of provisional hydro-phyllia (the so-called Athorybia-stage), and the tentacle knobs of the young Physo-phorid are different in character to those of the adult. Halistemma has a single provisional nectocalyx. A portion of the planula persists for some time as the yolk-sac of Crystallodes.

1 It appears that in the Diphyidae the functional nectocalyces are constantly replaced by others. The new nectocalyx is produced distally to the old, not proximally as in Physophoridae. See Chun, SB. Akad. Berlin, 1885, pp. 521 et seqq.; and Korotneff, Mitth. Zool. Stat. Naples, v. 1884, pp. 279 et seqq.

The Hippopodidae differ from other Siphonophora in the fact that the nectocalycine section of the coenosarc is bent parallel to the section bearing the remaining zooids, and the point at which the flexure takes place is also the one at which the rudiments of both nectocalyces and other zooids are formed. Other notable peculiarities in the Order are the presence of hydrocysts among the necto-calyces of Apolemia, and the addition of a nectocalyx to each group of zooids in certain species of Praya, hence erected by Chun into a genus, Lilyopsis (SB. Akad. Berlin (1), 1885).

A well-developed diffuse system of ganglion cells has been detected in the Discoideae. Ganglion cells occur also in the tubular coenosarc of some other Siphonophora. The nervous elements are especially differentiated in Physophora. Remarkable branched cells, 'neuro-muscular cells,' in connection with the longitudinal musculature of the coenosarc, have been described in some forms by Korotneff; cf. Mitth. Zool. Stat. Naples, v. p. 235, PI. 14, Fig. 13.

The Diphyozooid (supra) swims by means of the medusiform sexual zooid. If it is derived from a hermaphrodite colony, it is said nevertheless to be itself unisexual. But one derived from the unisexual Abyla pentagona has been observed by Chun to develope first a male and then a female zooid.

The sexual products of the Discoidean medusa are developed after the medusa is detached, therefore probably from the ectoderm of the manubrium; in all other Siphonophora properly examined from the endoderm: in Epibulia (= Galeolaria) in the rudiment of the gonophore: in Hippopodius in the bud of the future peduncle of the bunch of gonophores: in Forskalia in the rudiment of the peduncle: and in Agalma in the rudiments subsequently differentiated into several gonophores. The young sexual cells are not sub-epithelial at first, as they are in those Hydroidea where they appear in the endoderm, but they project into the cavity of the bud: but whether or no they really originate in the first instance from endoderm cells is not known. The sexual cells in the male wander into the ectoderm of the manubrium: so too in the female of the Calycophores named, but the endoderm subsequently grows round each ovum, furnishing it with an incomplete follicle placed to the outer aspect. But in the Physophores Forskalia and Agalma the single ovum is arrested in the endoderm, and surrounded by the spadix as stated p. 771.

Chun has propounded a theory of cyclical development in Monophyidae; e. g. (1) a form called by him Monophyes primordialis, which has been shown by Claus to be a larva, gives origin to (2) a second form=Muggiaea Kochii, in which the swimming bell of the Monophyes is replaced by a second and differently shaped nectocalyx, and the coenosarc developes groups of zooids: and (3) one of these groups of zooids is detached as a Diphyozooid, the sexual Eudoxia Eschscholtzii. Cf. Chun, SB. Akad. Berlin (2), 1882, and A. N. H. (5), xi; Claus, Arb. Zool. Inst.

Wien, v. 1884. According to Chun, the nectocalyx of the genus Monophyes is a second nectocalyx; cf. op. cit. (I), 1885 j and Claus, Z. A. viii. 1885.

A Siphonophoran has been variously regarded (I) as an assemblage of organs, or (2) as a colony of polymorphic zooids. The former view appears to have been based upon the idea that in any Hydrozoan colony the several factors represent organs of an individual, the individual being the colony itself. Such a view cannot be regarded as any longer tenable. There can be no reasonable doubt that a Hydroid colony is an assemblage of zooids which remain connected instead of separating after their formation by budding. It is no less certain that in a Siphonophoran, the nectocalyx, pneumatophore, sexual zooid, a polypite or hydrocyst, represent so many polymorphic individuals, and that structurally a Siphonophoran is as much a colonial organism as a Sertularia1.