PROFESSOR ALLMAN ON THE RELATIONS OF THE CCELENTERATA. 



463 



walls, so as to constitute an air-fillecl chamber (pneumatocyst) (n), which acts 

 as a float.'" 



Continuing to take the Hydroida as a standard of comparison, the other 

 hydrozoal orders may be now contrasted with them. If the atrium, or that 

 portion of the somatic cavity (fig. 5, b') which lies at the base of the manu- 

 brium in a hydroid medusa, be expanded laterally, and the ectoderm of its 

 floor be projected along four or eight symmetrically disposed radiating lines 

 into as many thick pillars (figs. 8 and 9, o, o), which converge towards the axis, 

 and there meet the manubrial extension of the cavity, while the thin interven- 

 ing portions of the floor between the pillars become developed into generative 

 pouches (d), and the velum or perforated diaphragm, which stretches across 

 the codonostome in the hydroid, disappears, we shall have the hydroid medusa 

 converted, in the more essential points of its structure, into a discophorous 

 medusa (figs. 8, 9). 



Again, a Lucernaria (figs. 10, 11) may be conceived of by imagining a hydra 



Fig. 8. 



Fig. 9. 



Fig. 8. — Diagramatic longitudinal section of a Discophorous Medusa, a, Radiating canal ; b, manubrium ; U, 

 somatic cavity ; d, generative pouches ; o, o, o, pillar-like extensions of the oral side of the umbrella ; z, tentacula-like 

 processes of the inner surface of the somatic cavity. 



Fig. 9. — Diagramatic transverse section of a Discophorous Medusa, a, a, a, Radiating canals ; b, manubrium 

 d, '/, generative pouches ; o, o, umbrello-manubrial pillars. 



with four tentacles to have these tentacles expanded laterally, until their sides 

 meet and coalesce, the hypostome still continuing free, and the proximal 

 portion of the body becoming extended into a solid peduncle of attachment, 

 containing a simple prolongation of the somatic cavity, or traversed longitudi- 

 nally by four narrow prolongations of this cavity, while generative sacs become 

 developed on each side of the partitions formed by the coalescent sides of the 

 tentacles. 



* In the above comparison of the siphonophora "with the hydroida, I have adopted for the sipho- 

 nophora the terminology proposed by Huxley, whose views of the homological relations existing 

 between the two orders I have also generally followed. See his " Oceanic Hydrozoa, - ' page 8, &c. 



VOL. XXVI. PART II. 6 E 



