1905, } BEARING ACTINIANS IN THEIR CLAWS. 499 
actinians at once makes it certain that the commensalism is not 
restricted to a single species of actinian, as might have been 
supposed ; moreover, the experiments given below prove that the 
individual crab will seize whichever of the two forms is presented 
to it. It may be, however, that the commensalism is limited 
to these two forms of actinians, Bunodeopsis and Sagartia. 
The figure of the actinian which Borradaile gives, reproduced on 
p. 496 (text-fig. 73), closely recalls the sagartiids found on the 
Hawaiian crabs, though he writes me that it is to be regarded 
as only a conventional representation. Richter’s figure (p. 495, 
text-fig. 72) is clearly intended for a Bunodeopsis, and there is 
no reason to suppose that the form is different from the Hawaiian 
species. 
Experiments were first conducted to determine the responses 
of Melia toward different actinians. The first crab was deprived 
of both its bunodeopsids and then placed in a dish in which were 
the two sagartiids removed from the claws of the second crab, the 
polyps lying free upon the bottom of the dish, not fixed by their 
base. The crab walked about for some time, showing no response 
whatever which would indicate that it was aware of the presence 
of the actinians; many a time it would pass in close proximity to 
them without any recognition signs. Happening in its wan- 
derings to touch one of the sagartiids, it stopped immediately, 
moved its claws around as if examining the polyp, and then 
grasped it at an oblique angle and carried it away. After the 
crab had moved about for some time longer, with one claw 
occupied and the other vacant, the second Sagartiid was inten- 
tionally placed so as to touch the unoccupied chela, when it was 
likewise examined, seized upon, and carried off. 
Thus the crab which originally held two bunodeopsids had now 
provided itself with two sagartiids; hence the species of actinian 
as regards Bunodeopsis and Sagartia are interchangeable. 
One of the sagartiids was now released and placed in the dish 
along with its original crab having both chelipeds vacant, the 
bunodeopsid fragment being also introduced. After a short time 
the Sagartia was seized, and later the small fragment of Buno- 
deopsis. The first Melia was also placed in a dish along with its 
own bunodeopsis and a Sagartia, and after a time these were 
likewise appropriated. Thus each crab was again provided with 
two actinians but of different species, the one a bunodeopsid and 
the other a sagartiid. In all the experiments the crabs appeared 
to seize either one or the other species with equal readiness. 
Experiments as to the behaviour of the crabs towards other species 
of actinians were very desirable, but at the time no other forms 
were available. 
The crabs exercised what must unquestionably be considered 
an intelligent selection, as far as regards the desirability or other- 
wise of an actinian already held by them. <A Melia carrying a 
small fragment of a Bunodeopsis in one claw and a perfect 
Sagartia in the other was placed in a dish containing the full- 
