LIGHT OF DIFFERENT WAVE-LENGTHS BY FISH 



31 



JiSmm l^/c///? o/j////n mrr? 



^ Afc7/cA,9e/ ^ J>u//er- ^ 



20M F-2 2/ as ac ae-a^' m^Arr 



Md 



H 



YP 



/?«/. 



rr7c?Ji fm cyr?^ 



/r*e<^ ^p<rr'&€^j ^^^ 



Fig. 13 



Graphs made bv plotting the percentages of correct or blue choices for each suc- 

 cessive 20 trials of the four dace, //, Md, YL, YP. 



Abscissae indicate successive slit widths in millimeters. 



Ordinate? percentao;e of correct choices 



Short Dash hnes show rapid changes in slit widths. 



Long Dash lines indicate that no tests were made at intermediate widths. 



Numbers on the graphs indicate successive groups of 20 trials from which 

 percentages were calculated. 



