124 THE NATURE AND VARIETIES OF 



special interest in establishing the relations between medu- 

 soids and sporosacs. He describes them as closed cap- 

 sules, each containing in its interior a second sac, with four 

 canals in its walls, radiating from the base, and communi- 

 ating with another circular canal round an opening at its 

 apex. An imperf orate columella or manubrium projects 

 into the interior from the centre of its concavity, so that 

 we may be said to have all the parts represented which are 

 found in the free medusa, except the marginal tentacles, 

 the eye-spots, and some lesser details.* The genital pro- 

 ducts are formed in the wall of the manubrium. In Sertu- 

 laria the medusoid conformation is less discernible, and the 

 sacs generally mature and discharge their contents while 

 still within the outer capsule. In Cordylophora the only 

 medusoid features presented, either by the spermatic or the 

 ovarian cysts, are the manubrium and some irregular longi- 

 tudinal canals in the wall of the sac ; there is no aperture 

 (at first) nor any circular canal. In Hydractinia we have 

 the columella without the canals ; and the sporosacs of some 

 species of Plumularia and Eudendrium are if possible of 

 still simpler structure, the latter containing but a single 

 ovum.*!* The progress of degradation finally reaches its 

 maximum in the common Hydra, in which the large me- 

 dusae of the " Hydra Tuba' are represented only by sper- 

 matic and ovarian cysts of the most rudimentary organiza- 

 tion, attached to the exterior of the polype. 



An equally great variety is presented by the reproductive 

 organs of the allied orders of Calycophorida and Physo- 

 phorida, ranging from the form of mere sacs to that of free 

 moving bodies (as in Velella), precisely resembling medusae, 



* Annals of Nat. Hist., 3d Ser., IV., 48. 



f In E. lacciferwm, Allnaan (Op. Cit.) describes short blind radiating 

 canals. The closely allied species of Atractylis detach well formed me- 

 dnsoids. See many interesting details on this head by Dr. T. S. Wright, 

 in Edinb. Philos. Journal, 1856-57-58. 



