210 CARNEGIE INSTITUTION OF WASHINGTON. 



food tliis has been accomplished with Epinephelus and Lachnolaimus, and 

 permanent records of their striking changes in coloration are available, from 

 which it appears that, under appropriate circumstances, they may be far 

 less conspicuous in the open than their own shadows upon the sandy 

 bottom. 



Reverting to the second of the cardinal points of the present report, we 

 may note that there are many indications that the colors of fishes are corre- 

 lated with their habits; hence the importance of being able to classify the 

 animals with reference to their behavior requires no emphasis. As a step 

 towards the accomplishment of this end it was proposed at first to examine 

 a few individuals of each of many species in order to separate the diurnal 

 from the nocturnal feeders. But the stomach contents of the snappers {Neo- 

 mcenis spp.) threw such light upon the question of the distastefulness or immu- 

 nity of some of the small bright-colored fishes, that, for a time, this phase 

 of the investigation was prosecuted at the expense of the original plan. 



From an examination of stomach contents alone it is at present possible 

 to assign only 9 species to the nocturnal group. These are Anisotremus 

 virginicus and Hcemulon rnacrostomum, parra, plumieri, and sciurus; the three 

 snappers, Neomcenis analis, apodus, and griseus; and Upeneus martinicus. 



The statement above is based upon the following facts: 170 Hsemulidse were 

 examined, of which 81 were taken with dj^namite at about 5 a. m. and 89 

 at 5 p. m. Of the whole, only 3 specimens of H. plumieri, out of a total of 

 52 for that species, failed to conform to the rule that full fishes are taken in the 

 morning and empty ones in the late afternoon. 208 snappers, of which 160 were 

 taken about 5 a. m. and the remainder 12 hours later, showed 27 (12.98 per 

 cent) exceptions to the same rule; but, as the average bulk of food the fishes 

 contain is greater in the morning than in the afternoon, the figures do not 

 accurately measure the proportionate amount of feeding done by night and 

 day. It is also noteworthy that 14.3 per cent of the snappers' stomachs 

 are empty in the morning, while only 8.7 per cent contain food in the 

 late afternoon. This means, imless more extended observation should alter 

 these proportions, that upon the average something less than 8.7 per cent 

 of the feeding of the three species of Neomceiiis is done by day. The dis- 

 crepancy between this figure and that first mentioned above may be accounted 

 for in either or both of two ways: some fishes may feed ravenously in the 

 forepart of the night, hunt less eagerly for food during the later hours, and 

 digest their total catch before morning; or a small proportion may make a 

 complete failure of their night's foraging. There is, indeed, some evidence 

 that the latter alternative explains the difference in question, but in no 

 case can there be doubt that the snappers are justly considered nocturnal 

 feeders. 



Only 15 specimens of U. martinicus were examined— 4 in the morning and 

 11 in the afternoon — but as there were no exceptions to the rule among 

 them, it may be assumed that they also feed by night. 



It was interesting and rather unexpected to find that Ocyurus chryswus 

 and Epinephelus morio feed indifferently at any time throughout the 24 hours. 



The food of the snappers, apart from the fishes it includes, calls for little 

 comment at present. It suffices to say that it consists very largely of 

 Crustacea, about 20 species of which were identified in it. Approximately 

 half of these were brachyura. Cephalopods, chiefly octopi, were found in 

 less than 5 per cent of the fishes examined. Annelids occurred rarely, except 

 upon special occasions, as at the time of the palolo swarm, and other food 

 constitutes so small a fraction of the snappers' diet as to be practically 

 negligible. 



