374 Papers from the Marine Biological Laboratory at Tortugas. 



a longitudinal incision made with scissors through the bodv wall on one side 

 from in front of the shoulder girdle to behind the vent. The body wall, stiff- 

 ened by formalin, sprang back into place and the air inclosed by it made the 

 fish float. The incision was made in a plane extending latero-ventrally, so 

 that its edges were as oblique as possible. The ventral edge was thus over- 

 lapped and held in place by the dorsal and acted as a valve to retain the air. 



Table 4 gives the result of an experiment in which both dark and light 

 red were offered with blue to a colony of about 100 snappers at the east 

 lighthouse dock. The snappers were first fed slowly for about 15 minutes 

 on blue atherinas, both to dull their appetite and to familiarize them with 

 the blue color. About 80 blue atherinas were thus fed before the blue- 

 light-red trials, which began at 10 a. m. The 5 trials of blue and light-red 

 atherinas followed at once on the preliminary feeding. An interval averag- 

 ing about 5 minutes was allowed to elapse between the trials and during 

 each interval 15 or 20 blue atherinas were fed. In the blue-light-red part 

 of the experiment the 100 snappers thus had offered them 80 blue fish in 

 the preliminary feeding, 9 X 15= 135 blues in the interval-feeding, and 25 

 blues included in the 5 trials, or a total of 240; while but 25 reds were 

 offered. x\ll the 265 fish offered were taken. 



The six blue-dark-red trials were begun at 4'' 12" p. m. of the same day, 

 after a preliminary feeding of 30 blue fish. The last trial included but 

 4 fish of each color, so that the total number of blue fish offered in this 

 blue-dark-red part of the experiment was 194 as against 29 reds. 



The method of staining the atherinas was that employed in experiment 

 26. The reds were stained as in experiment i. The light red was obtained 

 by a shorter stay in the stain. In order to determine the relative bright- 

 ness of the colors used the colored fish were matched with cardboard disks 

 stained with the same dyes and these were compared on a color wheel with 

 grays in the usual manner and with the following result : The blue w^as 

 equivalent in brightness to a gray containing 65 per cent black; the light 

 red to a gray containing 50 per cent black ; the dark red to a grav containing 

 75 per cent black. 



The experiment with the snappers was conducted in bright sunlight. 

 A considerable portion of this light does not penetrate into the water, but 

 is reflected from its surface. The snappers, moreover, view the colored fish 

 from beneath. They see the side which is illuminated by light reflected 

 upward from the white sand bottom. The intensity of this illumination 

 was estimated by examining the floating colored atherinas from below bv 

 means of the reflecting water-glass described in the footnote on page 298. 

 The brightness of the colors used was then measured in a light which was 

 judged to be of the same intensity as that which illuminated the colored ather- 

 inas in the actual experiment. 



Table 4 shows the total number of blue and red atherinas taken in each 



