1907.J TITCHENER AND PYLE— JUDGMENT OF DISTANCE. 103 



Strictly comparable; we had made the weak shadows too weak for 

 our direct purpose. At all events, the illusion-effect of 30.6 mm. 

 with the weak shadows is a large effect, and it is hardly possible 

 that the further weakening of the shadows, to the point realised in 

 the previous experiments, should, if the illusion motive is effective 

 at all, reduce this effect to a magnitude smaller than the MV of 

 practised observers. To make assurance doubly sure, we took 

 twenty-four series of experiments with F under the original con- 

 ditions, and obtained the results : 



Average Setting of Variable with 

 I lusion Long. Illusion short. No Illusion. 



249 ± 2 mm. 249.7 — 1-2 mm. 250.5 ± 1 mm. 



That is to say, there is no evidence of any effect at all exerted by 

 the imperceptible shadows. 



F's practice was continued, until the magnitude of the illusion- 

 effect was approximately the same for her as for our more prac- 

 tised observers. We finally obtained the following average values 

 for the illusion : 



It is an irony of chance that the observer G, whose results have 

 so far been (in Dunlap's sense) accordant with the hypothesis of 

 the influence of imperceptible shadows, should here give a larger 

 illusion-effect with the faint than with the strong shadows. 



The table is not altogether satisfactory, because the weak 

 shadows were, throughout, fluctuating; we had made them a little 

 too weak for comparative purposes. But the main point is clear: 

 even on the very edge of imperceptibility, the weak shadows have 

 an effect that is of the same numerical order as the effect of the 

 strong shadows, and this with observers whose judgments show no 

 influence of imperceptible shadows.^ 



^ The figures of the table represent the average value of the illusion- 

 effect as drawn from 16 methodically planned series, in 8 of which the variable 

 gave the " illusion long," and in 8 the " illusion short." The smaller values 

 obtained for the illusion-effect with shadows weak, in the cases of F and P, 

 are accounted for by the fact (attested both by measurement and by introspec- 

 tion) that in a few series the tips of the shadows were so faint that the illu- 

 sion motive was disregarded by the observer. 



