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 endoderm-lamella 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. [Illustration: After Allman, _Gymnoblastic Hydroids_, by permission of the Council of the Ray Society. FIG. 42.--Gonophores of _Dicoryne conferta_. A, A male gonophore still enclosed in its ectotheca. B and C, Two views of a female gonophore after liberation. t, Tentacles. ov, Ova, two carried on each female gonophore. sp, Testis.] 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 structure completely obliterated. There is, however, no evidence whatever for this, the gonads of _Hydra_ being purely ectodermal structures, while 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._