REPORT ON THE SIPHONOPHORjE. 315 



or a monogastric larva of the poly gastric Epibidia ritteriana (PI. XXII. figs. 6-8). 

 Indeed it is possible that this latter has been developed immediately from the 

 former, or that Cystalia is a larva of Epibulia, which has reached sexual maturity only 

 exceptionally (Paedogenesis). But comparing the structure of the pneumatophore in both 

 similar forms, we find that the polygastric Epibulia possesses the eight radial clusters of 

 hypocystic vilH characteristic of the Rhizophysidae, whilst these are wanting in the 

 monogastric Cystalia. In any case the fully developed state of the gonodendron in 

 the latter justifies its position as an independent genus. 



In the same month (December 1881) in which I captured Cystalia larvalis off the 

 coast of Ceylon, I took by the tow-net a number of young larvae of Siphonanthae, 

 which I supposed at the first glance to be the larvae of some Physonect, perhaps 

 an Agalmid. The most important stages of them are figured in PL XXII. figs. 1-4. 

 Comparing them with the larvae of the Agalmid Cupulita, which Metschnikoff has 

 described as Stephanomia pictum (85, Taf. xii.), we find a great likeness between these 

 larvae. The youngest larva observed (fig. 1) exhibits a spindle-shaped gastrula, 

 composed of a large-celled entoderm (d) and a small-celled ciliated exoderm (c), with 

 the invagination of the pneumatophore on the apical pole. In the second stage (fig. 2) 

 the medusiform body is divided by a transverse annular constriction into a proximal and 

 a distal portion ; the superior portion is the rudimentary umbrella with the pneumato- 

 phore (already containing a gas-bubble) ; the inferior portion is the primary siphon, from 

 the base of which arises a single tentacle. In the third stage (fig. 3) the distal mouth of 

 the siphon is open, and from its base, opposite to the dorsal tentacle, arises in the 

 ventral side the first bud (i), probably of a palpon. The fourth and last stage observed 

 (fig. 4) exhibits the number of buds augmented (as the beginning of a corona of 

 palpons V), and the single tentacle beset with a series of simple filiform tentilla (ts). 



Since I was not able to recognise the origin of these pelagic larvae, nor to follow 

 their further development, the cpiestion remains open, whether they were produced by a 

 Physonect or a Cystonect. In the latter case they may possibly have been derived 

 either from Cystalia or from the closely allied Epibulia. 



Cormidium. — The central or axial portion of the single cormidium, which represents 

 the entire adult corm of Cystalia, must be regarded as an individual medusome, the 

 modified umbrella of which is the pneumatophore (p) and the manubrium the siphon (s). 

 The base of this latter bears a single tentacle on its dorsal side, a single large gonoden- 

 dron (gd) on its ventral side. The short tubular pedicle of the siphon which connects it 

 with the base of the float represents the axial trunk of the corm, and has produced by 

 budding the corona of palpons, which are expanded between them. The gonodendron 

 itself is the sexual portion of the single cormidium, composed of numerous male and 

 female gonophores, each of which is a modified medusome. 



Comparing the single parts of the corm with the similar parts of related Siphono- 



