626 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



and fairly wide stripes of dark pigment, while the other was uniformly covered 

 by red pigment. 



Potts states that there are two factors in the color variations: One is the 

 character of the pigment, which ranges from a clear red to a purple so deep as to 

 appear black ; the second depends on the area of the body covered by the pigment. 

 The first seems to depend very largely on the species of the host. Synalpheus 

 hrucei is found, though more rarely, on a second host, Comatula purpurea, in 

 which reddish pigment predominates, and here the crustacean was sometimes 

 observed to show a red pigment covering the whole body, thus matching the host; 

 but sometimes purple and white striped individuals are found on Comatula 

 purpurea. 



Among the commensals of Comanthus annulatus also the pigment may be red 

 or red brown : but this is only so in the cases wliere the colored stripes are narrow ; 

 where they are broader the pigment is invariably darker. 



Potts gives the chief variations in color as follows: Uniformly red brown in 

 Comatula purpurea; with narrow red stripes in light green varieties of Covianthus 

 annulatus; with purple stripes of uniform dark color in dark green varieties of 

 Comanthus annulatus. 



He states that the purple pigment would seem to be either something addi- 

 tional or a more complex product of the red pigment. In the alpheids which were 

 preserved in formalin and glycerin to keep the original color as far as possible the 

 purple was speedily lost, leaving such a red as occurs in the lighter colored forms. 



Synalpheus hmcei thus exhibits a marked protective resemblance to its host. 

 Against the very light-colored crinoids the almost colorless alpheid is quite incon- 

 spicuous. The striped specimens are found upon comatulids in which bars of 

 light and dark pigments alternate, and so they, too, are not easily seen. And 

 lastly, when the dark pigment is uniformly spread over the body of the crinoid, 

 this is also the case with the alpheid. 



But though this is often true, there are many cases where the resemblance is 

 by no means close, and the alpheid is rather conspicuous. Potts is inclined to 

 explain these cases by supposing the alpheids to have migrated from some other 

 crinoid at a comparatively recent period. So thickly do the crinoids lie in the 

 crevices of the reef, all conceivable color varieties being herded together without 

 distinction, that it is more than likely that an interchange of commensals should 

 occasionally take place; for Synalj)heus, though tending to become a truly seden- 

 tary animal, is at times very active, and an excellent swimmer. 



Potts suggests that the conditions of commensalism do not allow natural 

 selection to come into play to any marked extent. The commensals are for the 

 most part inclosed by the arms of the crinoid as in a living cage. When the 

 water is calm these arms are relaxed and extended, but on the approach of an 

 intruder they curl up and so protect the soft central disk. It is scarcely to be 

 supposed that even a rapacious fish would take a mouthful of these hard and 

 unsatisfying arms for the sake of the shrimp which lies amongst them. In no case, 

 at least, which Potts saw were the arms of crinoids mutilated. 



