DEPARTMENT OF MARINE BIOLOGY. 183 



protected by stuffing-boxes it allowed one to focus accurately upon desired 

 objects and to make either instantaneous or time exposures. The complete 

 apparatus worked well; some desirable pictures were secured, and the possi- 

 bility of successful submarine photography in colors was established. 



In this connection I desire particularly to express appreciation of the privi- 

 lege I enjoyed in being permitted to work at Tortugas for two weeks after the 

 close of the regular season, which alone made it possible to attempt color- 

 photography this year. For invaluable assistance rendered me throughout 

 the whole period I am deeply indebted to Mr. John Mills, laboratory engineer, 

 who voluntarily remained with me until my work was completed. 



No sustained effort was made to collect additional evidence affecting the 

 conclusion already drawn, that the coloration of tropical fishes is obliterative 

 in its general effect. Such evidence, however, has continued to accumulate. 



The red goatfish {Upeneus maculatus), which is commonly seen over sandy 

 bottom in a gray phase with three lateral brown spots, turns green among 

 turtle-grass, or becomes striped or banded with reddish brown when its sur- 

 roundings are dark-colored. The common grunt {Hcemulon plumieri) has 

 a dark-brown phase consistently exhibited when it is schooUng among such 

 brown corals as Orhicella, yet is exceedingly pale when resting over bare white 

 bottom or straw-colored Palythoa. The schoolmaster {Neomcenis apodus) 

 makes appropriate adjustment to a brown or gray environment. Additional 

 adaptive color-phases or hitherto unnoted ability to change color have also 

 been observed in Calamus arctifrons, Lachnolaimus maximus, Neomcenis analis, 

 and Ocyurus chrysurus. 



Two additional green species have been seined on the grass-flats. Since one, 

 as yet unidentified, was represented by several specimens, and the type speci- 

 men of the other {Doratonotus megalepis) was taken in the same way in similar 

 surroundings, it is probable that both occur normally in such places. This 

 fact emphasizes anew the remarkably high degree of correlation already 

 demonstrated between the occurrence of this color upon fishes and in their 

 environment. 



It has been noted in a preceding report^ that field observations do not sup- 

 port the hypothesis that specific phases of coloration are associated with dif- 

 ferent activities. It is quite true, nevertheless, as has been stated by Dr. 

 Charles H.Townsendj^that in a number of species different color phases may 

 be observed in individuals according as they are in motion or at rest. The 

 differences in such cases seem to be primarily in pattern, although the crea- 

 tures' shade may be incidentally affected and doubt be raised concerning the 

 efficiency of the control exercised over it by that of their surroundings. This 

 is especially true when dark bands appear upon a light-colored fish as it comes 

 to rest in the open on sandy bottom, but it has been demonstrated repeat- 

 edly that amid dark surroundings such resting-phases are darker than under 

 other conditions. 



The tendency for transverse bands to replace longitudinal stripes or self- 

 color when swimming specimens come to rest on or near the bottom is very 

 striking. It appears more or less regularly in Calamus arctifrons, Epinephelus 

 striatus, Iridio hivittatus, Lachnolaimus maximus, Neomcenis analis, Sphyrcena 

 barracuda, and Upeneus maculatus. I am not aware of any instance in which 

 the reverse change occurs under the same conditions during daylight. 

 Whether or not the idea is entertained that banded patterns are more effective 



'Carnegie Institution of Washington, Year Book, No. 14 (1915), p. 208. 



'Townsend, C. H. : Observations on instantaneous changes in color among tropical fishes. 

 Thirteenth Annual Report N. Y. Zool. Soc, 1909. 



