1905. ] BEARING ACTINIANS IN THEIR CLAWS. 507 
capitulum) alone being naked. In aquaria they are very inactive 
and do not readily expand. 
Polydectus was under observation for only two or three days, 
and during most of that time it remained quiescent, hidden under 
fragments of coral. It allowed itself to be pushed over the floor 
of the vessel, making only a feeble attempt to escape, and showed 
little or no activity with its chelipeds. If irritated, the chele 
were not directed against the source of the stimulus as in the case 
of Melia. When the actinians were gently removed from the 
claws and after a time again presented, the crab made no imme- 
diate attempt to seize them. On the whole Polydectus proved 
itself to be a most unsuitable crab for experimental studies. 
INTERDEPENDENCE OF CRAB AND ACTINIAN. 
Enquiry may now be made as to how far the crab and its 
actinians are interdependent. Can the crabs maintain their 
existence deprived of the actinians, and can the latter exist 
separated from their captors? Although a careful search was 
made during three months’ collecting, no free independent 
examples of either Sagartia or Bunodeopsis, the actinians com- 
mensal with Mela, were met with, and neither Mobius nor 
Borradaile speaks of finding such. There seems no reason, 
however, why the actinians should not be able to live separated 
from the crustaceans. Compared with closely allied species 
elsewhere, they present no modifications whatever which indicate 
a correlation with the commensal habit. So far as the actinians 
are concerned, their presence in the claws of the crab seems of the 
most incidental character, and it can scarcely be doubted that 
ordinarily they are fixed isolated species, and may yet be found 
as such either in the Hawaiian Islands or elsewhere. As regards 
Polydectus and its associate Phellia, the latter certainly exists 
independently of any commensalism, for all round the Hawaiian 
Islands specimens of the sea-anemone are very numerous, attached 
to the under surface of stones and coral blocks. These places 
also constitute the habitat of the crab. In the case of the 
actinians Sagartia and Adamsia, commensal with hermit crabs, 
Faurot has shown experimentally that the polyps do not live 
long when separated from their host; but the relationship on 
the part of the actinian is here much closer than in the polyps 
simply held by Melia and Polydectus. In Sargartia palliata, at 
any rate, the commensalism is correlated with a permanent 
modification of form. 
The genus Lunodeopsis occurs also in the West Indian and 
Mediterranean seas, where it lives in shallow water loosely 
adherent to the leaves of the marine phanerogams Thalassia and 
Ruppia. In these regions, however, it is never found associated 
with crabs; indeed, the genus J/elia is absent from the Atlantic. 
A careful comparison of the external characters and internal 
anatomy of the Hawaiian and West Indian species of Buno- 
