44 HTDROZOA. 



of rougli weather usually disappear, their func- 

 tion having been, in all probability, previously 

 discharged. Yet nothing can be more perfect 

 than the series of transitional forms which estab- 

 lii^h the connection between these highly differen- 

 tiated organisms and the simple reproductive ap- 

 paratus occurring in Hydra, A few gradations 

 may be indicated. Thus, the closed gonophore of 

 Cordylophora sends off from its manubrial cavity 

 a system of prolongations, evidently homologous 

 with true gonocalycine canals {jig, 8, g). In 

 Tubidaria indivisa, fully developed canals are 

 exhibited by a distinct gonocal3rx, but this never 

 becomes detached {fig. 9). Neither does it pre- 

 sent marginal tentacles, though even these sur- 

 round the fixed gonocalyces of Campanularia Ld- 

 veni {fig. 1 o). And so we at once pass to the free 

 swimming generative cups of other Hydrozoa, 

 But it would' be easy to dwell on further modifica- 

 tions. Plumulariu pinnata, for example, has its 

 manubrium irregularly lobed, the lobes being, in 

 all probability, as Professor Allman suggests, in- 

 cipient gonocalj^cine canals, while in Cavipjcinu- 

 laria caliculata canals exist, though the manu- 

 brium itself is suppressed. And in some gonophores, 

 the canal system, at first easily recognisable, be- 

 comes obliterated by age. 



A gonophore, therefore, may exhibit in its de- 

 velopment four distinguishable stages, which cor- 

 respond, respectively, to the permanent forms of 

 the reproductive body in particular members of 

 the group. These conditions are : — 



I . That of a simple expansion of the body-wall, 

 as in Hydra, 



