146 DR A. MILNES MARSHALL ON THE 



In fig. 33 one of these zooids is represented in longitudinal vertical section, 

 together with the part of the rachis from which it springs. 



The tentacle, which is hollow, overhangs the mouth on the abaxial side. 

 a point of some interest, inasmuch as the calycular processes of the large 

 ventral zooids of P. phosphorea var. aculeata were also found to be abaxial 

 (cf. PL XXII. fig. 8). 



The mouth n leads into the stomodaeum s, the abaxial wall of which is 

 clothed with very long cilia r. At its lower end the stomodseum opens into 

 the body cavity h, which is lined by endoclerm, and is prolonged into the 

 tentacle. The body cavity is, at any rate in some cases, in direct communication 

 with that of adjacent zooids. 



As in zooids generally, there are only two mesenterial filaments present, of 

 which one is shown in the figure. These are borne by the axial septa, and are 

 extremely long and much convoluted. 



In some of the larger zooids I have noticed a slight notching of the margin 

 of the mouth, which may possibly indicate the rudiments of additional ten- 

 tacles. 



Below the polyp-bearing part of the rachis the zooids become much smaller. 

 The tentacles at first increase slightly in length, but become much more slender, 

 and lose their pinnules, with the exception of a single one, which is often 

 retained, giving a bifid appearance to the tentacle. These tentaculiferous 

 zooids are, as shown in fig. 30, almost confined to the lateral margin of the 

 rachis, the zooids of the dorsal and ventral surface becoming very early reduced 

 to the condition of small wart-like knobs. These become 'smaller in size and 

 more irregular in arrangement as we pass downwards, and finally cease about 50 

 mm. from the upper end of the rachis. 



In possessing single tentacles the zooids of U. gracilis resemble those of U. 

 Huxleyi and U. Carpenteri* two of the species obtained by the " Challenger " 

 from the North Pacific and South Polar seas respectively, but differ from 

 all other species, and indeed from all other Pennatulida yet described. 

 In possessing pinnated tentacles the zooids of U. gracilis stand absolutely 

 alone. 



I have pointed out above, when discussing the nature of the zooids of P. 

 phosphorea, that zooids must be considered as abortive polyps arrested in an 

 early stage of development. 



It becomes now an interesting inquiry how this unitentacular condition of 

 the zooids of Umbellula arose. So far as is at present known, the earliest rudi- 

 ments of all eight tentacles arise simultaneously in the Pennatulid polyps. I 

 have described this above in the case of the asexually formed polyps of Pennatula 



, * Kolliker, Zool. Chall. Exp., part ii. 1880, pp. 21-24. 



