As to the special character of the colony two views have found acceptance. One is, that the coenosarc, as it is called, is the homologue of the manubrium of a Medusa, the primitive hydrophyllium, present, e. g. in the larval Physophora, representing the umbrella: the other that it is a floating colony, the individuals of which are connected by a coenosarcal stem or stolon the equivalent of a hydrorhiza. The former theory is supported by the analogy of such a medusa as Sarsia or Willia, in which the manubrium produces medusae by budding. The following objections, however, appear fatal to it. (I) A primitive hydrophyllium is not present in all Physo-phoridae, and is not present at all in Calycophoridae, and to suppose that it has been aborted in these instances is the purest assumption; (2) that the hydrophyllium itself is probably the homologue of a medusa in its entirety; (3) that it is a lateral appendage invariably of the coenosarc, and there is no reason to suppose it is anything else in the planula; (4) taking such a form as Epibulia (Calycophore) or Halistemma (Physophore), it is clear that the coenosarc is attached to the aboral apex of the exumbrella of the first nectocalyx developed, and cannot possibly represent a manubrium.

For these reasons the second view mentioned above is certainly to be preferred. The coenosarc of a Siphonophoran is then comparable to the aboral stolon of the young Narcomedusan Cunina rhododactyla, or of Cunocantha octonarta, which produces buds for a stated period2. The fact that the planula gives rise to more than one zooid is simply an instance of precocious gemmation, to be paralleled by the fission of the embryo of Lumbricus trapezoides, and possibly of the Acrasped Chrysaora. It may be noted also that the Hydroid planula does not always form the hydranth directly; e.g. in Eutima it attaches itself and produces a hydranth by gemmation, itself becoming a hydrorhiza1.

1The hydrophyllium is usually regarded as the homologue of a nectocalyx. Its central endo-derm canal is furnished with accessory canals such as occur in the nectocalyces of Diphyes and Praya. The homology of the tentacle is more difficult. In the Discoideae it probably represents a zooid, i. e. is a dactylozooid, but in Physalia it is possible that the sac represents a hydrocyst. The tentacle itself is in that case attached to this hydrocyst in the same position, i. e. at its base, as is the tentacle of most other Siphonophora. The position is a most unusual one, but Metschnikoff states that he has observed a single tentacle attached similarly to the base of the manubrium of a Medusa (Dipurena). The Siphonophoran tentacle, however, is sometimes attached directly to the coenosarc as in Athorybia rosacea. And it may be noted that in the Physophirid Stephanospira insignis the hydrocyst bearing the female zooids is prolonged apically into a tentacular process with lateral branches bearing nettle-batteries; see Gegenbaur, Nova Acta, xxvii. p. 399. It is quite possible that such a structure as the tentacle of Porpita, beset with capitate processes, represents a zooid form from which the more complicated tentacles of other Siphonophora have been derived.

In determining such a question, much stress cannot be laid on the origin of the tentacles from separate buds. The tentacles of a Hydra originate in the same manner. The place of origin of the buds is however another matter.

2 See Metschnikoff, Z. W. Z. xxiv. 1874, pp. 28 et seqq., PI. V. Figs. 4-8; and Brooks, Mem. Boston Soc. Nat. Hist. iii. (12), 1886, pp. 362 et seqq.