HYDROMEDTJSAE, SIPHONOPHORES, AND CTENOPHORES. 345 



In the eastern Pacific specimen the canal system of both bract and 

 nectophore was damaged, consequently the figure (19116, pi. 20, fig. 

 6) was incomplete in this respect. In the present representative the 

 canals are well preserved and follow the same general plan. The 

 ascending branch of the bracteal somatocyst has a transverse branch 

 running to the dorsal surface (represented in the earlier figure by 

 dotted lines). The two descending (hydroecial) trunks are separate 

 from the beginning, all three arising at the same point, descending 

 over the two faces of the hydroecium, on the right and left, respec- 

 tively, to unite near the tip of the bract. In the present specimen 

 their point of union is slightly nearer the extremity of the bract than 

 it was in the eastern Pacific one. 



The canal system of the special nectophore is especially interesting- 

 because it gives us our only clue to the identity of this eudoxid. 

 There is a single main trunk running along the dorsal face of the- 

 hydroecium, forming its pedicular canal at its upper end, and reach- 

 ing the extremity of the basal prolongation of the bell. It gives off 

 four chief branches as follows: one opposite the upper, one opposite 

 the lower face, of the nectosac : the other two together, opposite the 

 center of the nectosac. These four join the nectosac at its widest 

 level — that is, some distance below its apex — to form the subumbral 

 canals. The basal ends of these four trunks could be seen in the east- 

 ern Pacific specimen, but their courses coidd not be traced. 



Now, this same type of subumbral canals is to be seen in the mono- 

 phyid Nectopyramis thetis Bigelow ( 1911a) . and. so far as I am aware, 

 in no other siphonophore. The implication, of course, is that Archi- 

 soma is the free eudoxid of N. thetis. This, of course, is proposed 

 only tentatively ; the gaps in the chain of evidence are much more ex- 

 tensive than its links. One objection is that the only record of 

 N. thetis is from the Atlantic. But, judging from what we know 

 of the distribution of other siphonophores. this is not serious. If my 

 suggestion should prove correct it would be of great importance sys- 

 tematically, because the only other species yet referred to Necto- 

 pyramis , N. albatrossi, lacks the special nectophore, and likewise has 

 subumbral canals of the ordinary type. 



Suborder PHYSOPHORAE. 

 Family FORSKALIIDAE Haeckel, 1888. 



Genus FORSKALIA Kolliker, 1853. 



FORSKALIA, species ?. 



Jolo. One very fragmentary specimen (Cat. No. 29344, U.S.N. M.) 

 minus all the older nectophores and all the bracts, and much con- 

 tracted. 



