232 E. B. TITCHENER— ETHNOLOGICAL TESTS OF SENSATION. 



paper shown for lo to 20 sec. on a gray ground, or of gray paper on 

 a colored ground ; and in what iUumination ? The method favors 

 R; I can reproduce the results with observers of normal vision. 

 There is nothing to show that the IMurray men were not the victims 

 of circumstances ; and Rivers does not report corresponding tests of 

 Europeans. 



4. Rivers lays no great stress on his observations of color-pref- 

 erence. " Among saturated colors/' however, " R easily had the first 

 place, followed by B, while Y and G were distinctly less favored. 

 . . . Complementary colors were commonly worn together, Y with 

 B and G with R."^- The first of these facts is a little surprising, in 

 view of the practical importance of yellow and green ;^^ taken to- 

 gether with the second, it may perhaps indicate that the natives, once 

 their attention has been called to color as such, have a normal appre- 

 ciation of blue. 



Here, however, we are dealing with fairly large surfaces of 

 color, surfaces that would be viewed in indirect as well as in direct 

 vision. One of Rivers' strongest points against the native is that, 

 peripherally, " the color blue was recognized readily, even more read- 

 ily than other colors. The color of the patch used was saturated,"^* 

 but even so the results do not accord with those obtained in direct 

 vision, so that the two sets of observations need to be reconciled. 

 I have tried to show that the tests of direct vision are not convincing. 

 It seems, also, that Rivers' argument from the sensitivity of the 



52 R, 84. 



52 1 have said nothing of the supposed relative insensitiveness of the 

 islanders to G, partly for reasons of space, partly because the position of B 

 is the more important. H. E. Houston and W. W. Washburn ("On the 

 Naming of Colors," Amcr. Joiirn. Psych., XVIII. , 1907, 523) found no over- 

 lap of B and Y or R and G, but a marked confusion of B with G and B 

 vi^ith P. It is notevi^orthy that H. K. Wolfe ("On the Color-Vocabulary of 

 Children," University of Nebraska Studies, I., 1890, 23) finds no such con- 

 fusion of B and G. Many points of Wolfe's investigation are of interest in 

 connection with the Murray Island results. Thus " the pupils seem loth 

 to confess their ignorance ; four fifths of them attempted to name orange, 

 and only one fifth knew what it was" (24); and the expressions "dark 

 white" and "light dark" were used in good faith and with meaning (28). 



54 R, 79. If the tests were made in the order in which they are reported 

 (of. R, 49, 53, 70) the name bulubulu may have become standardized simply 

 by the progress of the tests themselves. 



