383 Papers from the Marine Biological Laboratory at Tortugas. 



series I and II, but it first intersects the line of dots and dashes at a point 

 farther from the ordinate of percentages than in series II and III. a further 

 evidence of discrimination. Of the lOO fish thrown in the lo trials of series 

 III, 21 remained untaken or were eHminated from the experiment through 

 an error of the record. Of the 79 fish taken. 44 were bkie and 35 were red ; 

 30 bhie fish were taken in the first 5 places as against 20 red, while 14 blue 

 fish were taken in the second 5 places as against 15 red; 16 fish remained 

 untaken at the close of the experiment, and of these 12 were red, while 

 4 were blue fish to which a small amount of red stain had been accidentally 

 transferred. The diagram (table 10) shows in percentages in each order the 



Table 10. 



frequency with which a colony of loo gray snappers took atherinas in each 

 order in three series of trials. The atherinas were thrown in lots of lo. In 

 the first two series 5 white and 5 blue were thrown together ; in the third 

 series 5 blue and 5 red were thrown together. Percentages are calculated 

 as in table 9: 21 atherinas remained untaken in the third trial. 



The plotted data seem to show conclusively that the fish discriminate be- 

 tween the red and blue atherinas. The approximation of the two lines in 

 the second half of their course in series III and their descent toward the 

 zero level is due to the fact tliat the snappers were at this time no longer 

 hungry, took any but normal fish very slowh', and left many untaken. The 

 behavior of the snappers toward the red and blue fish is further proof of 

 discrimination, as shown by the following extract from my notes: 



