160 CARNEGIE INSTITUTION OF WASHINGTON. 



compartment was probably the interloper. A single burrow in one instance 

 contained three living and uninjured crabs. In this case an accessory pocket 

 at one side below contained a small female, while the large terminal chamber 

 held an adult male. 



0. ceratophthalma is found both on shell or coral beaches of white sand and 

 on the dark beaches formed from disintegrated lava. Its color in the two 

 cases accords very well with that of its surroundings. Upon the light-colored 

 beaches, so far as the dorsal surfaces of the adults are concerned, it is of their 

 light hue, with or without a few comparatively small maroon blotches on the 

 cardiac region. Upon the dark beaches, on the contrary, it is of an almost 

 uniform dark reddish brown. The younger individuals show corresponding 

 differences in shade, but are variously mottled rather than uniformly colored. 

 If tested by placing them in black and in white dishes they show speedily 

 that they possess capacity to change their shade in the direction of that of 

 the substratum on which they stand. The conformity of the species to its 

 environment in nature, therefore, seems capable of simple explanation. It 

 may be added parenthetically that Thalamita Integra (Dana) and Portunus 

 sanguinolentus (Herbst) are also changeable in shade, a statement particu- 

 larly applicable to the former. 



In my last report it was stated that the possibility of successful submarine 

 photography in colors had been established by experiments performed during 

 the summer of 1917. It must apparently be added, however, in view of this 

 summer's results, that an accurate photographic record showing in color the 

 nature of the environment to which tropical reef-fishes possess such marked 

 power of color-adaptation may not be obtained by use of the same screens 

 as are used for color-photography in air. When the first report was made 

 it was supposed that the failure to secm'e accurate reproduction of the differ- 

 ent color values in the field of the lens was due to the use of another screen 

 than was made by the manufacturer of the plates used. In the present 

 instance essentially the same defects appear in the developed plates, although 

 it was demonstrated in advance that in air the adjustment of screen to plates 

 was perfect. The difficulty seems to lie in the fact that differential absorp- 

 tion of light rays of different wave-lengths occurs in their passage through 

 the water. The light which falls upon a particular color-screen in air and in 

 water is not composed, therefore, of light rays of the same sorts in the same 

 proportions in the two cases, and its components can not be reduced to equal- 

 ity of influence upon the sensitive plate in both media. It is necessary that a 

 special screen stopping more of the rays of shorter wave-length than is required 

 in air should be used in submarine color-photography, and even then the 

 best results at a given depth and time could probably be obtained by its use 

 at only one distance from the object photographed. 



About 90 species of fishes were seen during the summer, of which several 

 are either new to science or hitherto unreported from the Hawaiian Islands. 

 Some, however, were encountered rarely, so that those under daily observa- 

 tion scarcely exceeded half the total. 



As is the case with the tropical fishes of every region, the colors of many 

 Hawaiian species are inadequately described in systematic works, since these 

 are concerned almost wholly with the appearance of dead specimens. It may 

 be mentioned in passing that what is probably the best existing record of the 

 form and coloration of a group of tropical fishes of comparable numbers is 

 embodied in the series of painted casts of more than 340 species prepared by 

 Mr. John W. Thompson for the Bishop Museum, Honolulu. Yet even this 

 notable series, in spite of the fact that in a number of instances several color- 

 phases of a single species are represented, fails to do complete justice to its 

 subject. 



