HYDROMEDUSAE, SIPHONOPHORES, AND CTENOPHORES. 283 



The bases of the tentacles are laterally flattened, and on the outer 

 surface there is a more or less pronounced spur clasping the exum- 

 brella (pi. 39, fig. 6). This character, however, varies both individ- 

 ually and with contraction. It is not shown in Maas's figure of his 

 Japanese specimens, nor was it noticeable in either the Siboga or the 

 Albatross eastern Pacific series. There is usually but not always 

 an abaxial ocellus on the base of each tentacle, and of each marginal 

 bulb. But this is not an invariable rule, for an ocellus may be lack- 

 ing, or there may be two in conjunction with a single tentacle. Thus 

 in one specimen 7 of the 8 large tentacles and 7 of the 9 knobs have 

 1 ocellus, while 2 knobs have 2 ocelli each ; in another, the 8 tentacles 

 and 3 of the 7 knobs have 1 ocellus, while 4 knobs have 2 ocelli each ; 

 in a third 5 tentacles have 1 ocellus; 1 has a diffused pigment spot 

 and 2 have no ocelli, and of the 8 knobs 4 have one and 4 have 2 

 ocelli each. 



The gonads are of the characteristic type, as a comparison of the 

 photograph (pi. 39, fig. 5) with any of the numerous figures of L. 

 octona (Maas 19045, pi. 1, fig. 9; Hartlaub, 1913) will show. The 

 edges of the radial canals are irregularly fluted, especially in the 

 distal one-third of their course (pi. 39, fig. 6) ; but the lobing varies 

 from canal to canal, or even on opposite sides of any given canal. 

 Some canals hardly show it at all. It is not present in young speci- 

 mens, and first appears as a jagged outline (Bigelow 1909a). It is 

 never as pronounced here as in Catablema or in Neoturris. 



Color. — In the preserved specimens the gonads and lips are violet 

 or purple; the tentacles opaque and yellowish, their bases brownish 

 yellow. The ocelli are dark brown. In the large Japanese examples 

 described by Maas (1909, pi. 1, fig. 3) the colors are similar. L. 

 octona is very generally distributed over the warmer parts of the 

 Indo-Pacific region, but is not yet known from the northwest Pacific 

 north of Japan. 



Genus NEOTURRIS Hartlaub, 1913. 



Neoturris Hartlaub, 1913. 



Tiara Lesson, 1837 (part). 



Turris Lesson, 1837 (part). 



Clavula Bigelow, 1909a. — Mayer, 1910 (part). 



The discovery by Browne (1910) and Hartlaub (1913) that the 

 Turris neglecta of Forbes (of which Clavula is probably the hydroid) 

 is not a Pandiid forbids the use of Clavula as the name for the present 

 genus. 



Neoturris is undoubtedly very closely allied to Leuckartiara; 

 indeed, it has often been united with it, as, for instance, by Van- 

 hoffen (1912). But as previously pointed out (1909a), I agree with 



