478 FRANCIS B. SUMNER 



Experiment alone (see p. 467) can place beyond question the 

 accuracy of this interpretation. What is needed especially is a 

 satisfactory determination of just which elements of the visual 

 field it is to which the animal conforms its own appearance, and 

 which ones it is that serve as a criterion by which the shade of 

 the background is apprehended. So far as I can see, differences of 

 direction from the animaVs eyes are the only ones which can be 

 invoked in differentiating these two sets of stimuli, and Keeble 

 and Gamble (whose treatment of this problem was unknown to 

 me when the foregoing discussion was written "*'- incline to the same 

 opinion. They hold (p. 354) that ''in some way, the eye differen- 

 tiates between the direct and the irregularly scattered light, in 

 other words, it displays a certain dorsi-ventrality." Under ordi- 

 nary conditions, the background is below and the source of light 

 above. But the authors find that, if the conditions of illumination 

 be artificially reversed, the ''background" being above the animal 

 and the light entering from below, the reaction to the former is the 

 same as when it lies beneath them. Thus, they hold, "the dorsi- 

 ventrality is probably not due to a permanent structural difference 

 in the two sides of the eye." It is not clear, however, from their 

 account of this experiment, that the conditions of illumination 

 were not complicated by total reflection from the bottom of the jar. 

 Unless the animals looked directly downward, or at least within 

 a certain angle with the bottom, they would see, not a brightly 

 lighted field below them, but the reflections of objects in the upper 

 portions of the tank (pp. 427,428 of the present paper). ^'^ 



»2 See foot-note 81 



*^ I cannot feel quite sure that the experiment of Bauer (Centralblatt fiir Physi- 

 ologie, 1906), in which he used electric lights, placed above and below the glass 

 container, is not open to the same objection. From this experiment and others, 

 Bauer concluded that the assumption of a dark shade by Idotea was determined 

 by "Simultankontrast," irrespective of the position of the contrasting portions 

 of the visual field. This is certainly not true of fishes, as is shown by my ex- 

 periments. (See particularly p. 451 et seq.) 



The experiments of Mayerhofer, likewise, (op. cit., pp. 553, 554) in which a 

 mirror was placed below the glass container, inclined at an angle of 45°, appear 

 to me to be inconclusive, owing to the same apparent technical defect; and it is 

 significant in this connection to note that fishes lighted from below, in this 

 way, assumed the same shade as those kept in total darkness. 



