REPORT ON THE SIPHONOPHOR^E. 197 



of the French observers, the apical pneumatophore was surrounded by a corona of five 

 nectophores, and the single large basal siphon by a corona of numerous palpons ; two 

 tentacles are figured, but probably these were only parts of a loop of a single one. The 

 tentilla, however, arranged in a single long series along the tentacle, were not simple 

 lateral branches as in our Circalia stephanoma, but provided with an ovate cnidosac at 

 the distal end. It is possible, therefore, that this Circalia haplorhiza represents another 

 genus of Circalidse ; it may be called provisionally Circonalia. 



Another monogastric Physonect, which is very similar to Circalia, and inhabits the 

 western part of the Gulf Stream (near the Tortugas Reefs, Florida), has recently been 

 described by Fewkes under the name Agalma jxipiUosum. Its nectophores are 

 papillate, and the tentilla, which form a long series on the single tentacle, are tricornuate 

 with a median terminal ampulla and two lateral horns at the distal end of the spiral 

 cnidoband (as in Agalma, Agalmopsis, &c). But since no gonophores were observed on 

 this remarkable form, it is perhaps only the larva of another Physonect. 



The single cormidium, which represents the mature corm of Circalia, is of great 

 interest on account of its typical simplicity and its morphological relations to other 

 Siphonanths, especially to the Discolabidge {Physophora, Step>hanospira) on one hand, and 

 to the Rhodalidse (Stephalia, Rhodalia) on the other hand. The entire cormidium may 

 be compared with a Medusa, which has preserved the original simple manubrium (the 

 central siphon), but whose umbrella has been transformed into a pneumatophore, and 

 produced by budding (from the base of the manubrium) a corona of radial medusomes 

 (eight in Circalia stephanoma), each medusome being composed of a proximal nectophore 

 and a distal palpon (or a pair of palpons), and beyond this a gonodendron. 



It is perhaps a fact of great morphological value, that the octoradial type of 

 Medusa in Circalia is expressed not only in the structure of the pneumatophore (with 

 eight radial pouches of the cavity, and eight pigment-rays at the apex), and of the 

 single central siphon (with eight liver-ridges and eight mouth-lobes), but also in the 

 composition of the nectosome (with eight radial nectophores) and of the siphosome 

 (with sixteen palpons and eight gonodendra) ; these latter numbers, however, may be 

 accidental. 



Comparing the entire corm of Circalia with a single Medusa, which has produced 

 a cormidium by budding from the base of the manubrium, we get a further support for 

 our medusome theory (p. 3). The transformed umbrella of the original Medusa person 

 is the apical pneumatophore ; its manubrium is the prolonged central siphon. The 

 distal and lower part of the siphon only has preserved the function of a feeding and 

 digesting stomach, whilst the proximal and upper part represents the axial trunk of the 

 corm, from which the buds arise. These buds may have been originally simple Medusae, 

 but afterwards transformed into loose medusomes ; the umbrella of the secondary 

 Medusa has been developed into a nectophore, the dislocated manubrium into a palpon, 



