Vision 215 



within it do vary. The objects seen proportionately largest are those 

 directly overhead. If an object should swing down a semicircle from the 

 zenith toward the horizon, along a meridian of the aerial hemisphere, it 

 would get shorter and shorter in its meridional length and in its width 

 measured parallel to the surface. Thus, even though its linear distance 

 from the fish were constant, its apparent size would become smaller, the 

 closer it approached the horizon. It would be seen more and more dimly, 

 too, for light rays which make small angles with the water are largely 



Figure 8.1. A. Visual field of a fish looking upward. B. As seen from hori- 

 zontal view. [From Walls, G. L., "The Vertebrate Eye and Its Adaptive 

 Radiation," Cranbrook Inst, of Science, Bull. 19, Bloomfield Hills, Mich., 1942] 



reflected, and but little of such light is refracted down through the sur- 

 face to enter the eye of a fish. 



"The entire circumference of the 'horizon,' which a swimming man 

 could see by treading water and rotating 360° on his axis, is, for the fish, 

 contracted to the few inches or feet of circumference of his surface 

 window. (Figure 8.1A and B. ) It follows that a man standing on the bank 

 of a pool is seen as a tiny doll by a fish which is a few yards away and 

 only a few inches below the surface. Our tendency is to suppose that the 

 fish will see us more poorly still; just as we see him less well, if he drops 

 deeper in the water; but since dropping lower enlarges his window, it 



