THE QUARTERLY REVIEW OF BIOLOGY 



ally light reflected from colored pieces of 

 paper, painted food containers, colored 

 balls, and the like has been utilized. 

 Control of the intensity factor has never 

 been adequate and in many cases has been 

 quite overlooked. The results reported, 

 and they are about equally divided for and 

 against the existence of color vision in 

 dogs, therefore cannot be taken seriously. 

 Perhaps the least inadequate study making 

 use of reflected rather than transmitted 

 light, is that of Smith. This writer has 

 been generally quoted as having concluded 

 in favor of color vision (see Washburn, 

 Animal Mind, p. 156). A careful reading 

 of his paper shows that this does not 

 fairly represent his position and that the 

 data he presents certainly support no such 

 conclusion. Although Smith is inclined 

 to suppose that certain dogs may be able 

 to discriminate colors to a limited extent, 

 he emphasizes his opinion that such dis- 

 crimination "is highly unstable and 

 cannot be supposed to play any part in the 

 animal's normal existence." For all 

 practical purposes this is equivalent to 

 a denial of color vision. This tentative 

 conclusion is quite in agreement with the 

 work of the Russian school, as is indicated 

 by the following quotation from Pavlov 

 (3 5, page 13Z f.): "Dr. Orbeli in a first 

 series of experiments was unable to detect 

 any differentiation of colours on the part 

 of his dogs. In a second series of experi- 

 ments, however, positive results were 

 obtained in one dog, but only with great 

 difficulty, and even in this case the experi- 

 ments were still open to criticism. The 

 results obtained by other investigators, 

 both Russian and foreign, lead to the 

 conclusion that colour vision in dogs, if 

 present, is only of a very rudimentary 

 form, and that in most dogs it cannot be 

 detected at all." 



It must be concluded, therefore, on the 

 basis of laboratory tests that the dog's 



vision is decidedly inferior. The casual 

 observations made at home and in the 

 field do not conflict with this conclusion. 

 In the every day life of the dog it is no 

 doubt a supplementary sense, rather than 

 the dominant sense as in man. In the 

 laboratory it seems called into play only 

 when the more dominant senses, the 

 olfactory and the kinesthetic, fail to bring 

 the animal success. The discrimination 

 experiments already reported are examples, 

 and there are others which even more 

 clearly indicate the tendency of the animal 

 to overlook visual clues in favor of those 

 of other modalities. This conclusion 

 applies only to still and not necessarily to 

 moving objects. De Jong (14), after 

 training his dogs to open a box by lifting 

 a latch in order to obtain food, turned the 

 box ninety degrees, in the dog's absence. 

 The dogs then went to the point on the 

 box where the latch had previously been 

 and made the movement of the paw that 

 had formerly served to lift the latch. 

 This of course failed to produce the 

 desired results, but the dog nevertheless 

 continued to attack that same point for 

 some time, entirely overlooking the rather 

 prominent latch, which, although at a 

 new place, was still clearly within his 

 visual range. Learning to open the door 

 with the box so turned seemed to be an 

 entirely new problem, accomplished in 

 the usual trial and error manner. Only 

 after the position of the box had been 

 shifted a number of times, and relearned 

 as often, did the dogs finally come to 

 respond to the latch itself, and even then 

 not necessarily on the basis of vision alone. 

 Experiments similar to these have been 

 made by Johnson (zo) with similar results. 

 An even more striking demonstration of 

 the insignificant role of eyesight is found 

 in other of Johnson's experiments. He 

 found that problem boxes were solved as 

 readily by dogs temporarily deprived of 



