Oliver. — Mollusca of the Kermadec Islands. 



50 1 



islands, and outlined their probable history. Briefly, my conclusions are 

 these : All the islands are of recent volcanic formation, two of them being 

 still in the solfatara stage. Sunday Island is built on a base of which part 

 is known to be composed of syenite, a rock usually associated with con- 

 tinental areas. The first eruptions were submarine, the later ones sub- 

 aerial. Sundav Island has thus arisen from beneath the sea in recent 

 geological times, and has not exceeded its present limits more than can be 

 accounted for by marine and subaerial denudation. Whatever portion 

 of the continental base was ever formerly above sea-level was entirely 

 submerged before the present group of islands came into existence, and any 

 terrestrial life it may have contained completely destroyed. Assuming 

 that I have interpreted the geological evidence correctly, then the whole 

 of the terrestrial life now found at the Kermadecs has arrived by accidental 

 transoceanic migration ; but among its marine forms one might expect to 

 find an element suggestive of a continental connection. 



Owing to the vast amount of work still to be done on the Mollusca it 

 is not possible to speculate on their origin and relationships except by a 

 statistical method. The relationships of the 261 species enumerated in 

 the present paper may be expressed in tabular form thus : — 



Total. 



Per Cent. 



Gastropoda. 



££££ Amphlnenra. 



Cephalopoda. 



Endemic 

 Polynesian 

 New Zealand 

 Pelagic 



12 

 16 



Endemic Element. 



A third of the Kermadec molluscs are not known elsewhere. This 

 proportion will probably be decreased when the molluscan fauna of the 

 south-west Pacific is more closely studied. Taking first the Gastropoda, 

 the bulk are probably of Polynesian affinities. Half the species are small 

 shells, mostly under 5 mm. in length, most of them falling within the 

 families Rissoidae, Cerithiidae, Eulimidae, and Turritidae. About one- 

 fifth are fairly large shells, and these form a very remarkable collection 

 of species which one might hardly expect on an isolated volcanic island. 

 Possibly they indicate a former continental connection. Some of them 

 appear to have no closely related species in the adjacent regions, though 

 their affinities are mainly with the north (Scutellastra kermadecensis, Tectus 

 royanus, Conus kermadecensis, Cassidea royana, Spondylus raoulensis). The 

 highly variable members of the genera Cellana and Siphonaria were appa- 

 rently derived from the south. The land shells are, according to Iredale, 

 almost entirely of Polynesian origin. 



Members of the class Amphineura appear to be particularly useful for 

 indicating the routes of migration of marine faunas, especially as they are 

 peculiarly sedentary in their habits, and, no doubt as a consequence, 

 restricted in their distribution. The affinities of the Kermadec species are, 

 as noted by Iredale, entirely with those of New Zealand. The presence 

 of a variable member of the characteristic New Zealand genus Eudoxochiton 

 is worthy of special notice. 



