100 TITCHENER AND PYLE— JUDGMENT OF DISTANCE. [April i8, 



decided, therefore, to keep the ilkimination constant, and to remove 

 and attach the Miiller-Lyer angles as occasion demanded.^ On the 

 other hand, there can be no doubt that the direct pairing of two 

 precisely similar series is methodically indefensible; the observer 

 tends to say ' equal ' or ' doubtful,' in the second series, at about 

 the same point at v^hich he passed this judgment in the first. 



The method that would naturally be employed in an investiga- 

 tion of this sort is the method of constant-i? as applied to the deter- 

 mination of equivalent stimuli.^ We desired, however, to keep as 

 closely as possible to the method chosen by Dunlap, and accord- 

 ingly proceeded as follows. For each observer we made out a set 

 of twenty-four single series. In eight series, the angles were used 

 with the " illusion long " as the variable. In eight series, the angles 

 were used with the " illusion short " as the variable.' In the re- 

 maining eight series, the illusion-angles were not used. On the basis 

 of the previous experiments we selected eight different starting- 

 points for these series, four lying well without and four well within 

 the point of subjective equality. The order of the single series was 

 decided by chance; the variable was shown as often on the right as 

 on the left. The intervals between series were kept constant, so 

 that the observer had no means of knowing whether or not the 

 experimenter changed the apparatus. For the rest, the observations 

 were taken and the calculations made as in the preliminary 

 experiments. 



The table shows the results obtained from five observers : the G, 

 Hg, Hn and P of the former experiments, and I\Ir. R. W. Sailor, 

 a trained observer. 



^ It must be remembered that, though the observers were not informed of 

 the object of these experiments, and (with the exception of P) were unfamiliar 

 with Dunlap's work, they nevertheless received a fairly definite suggestion 

 from the preliminary tests with the circle and skeleton square. No one of 

 them reported any difference in the appearance of the white background from 

 series to series. Two, however, differentiated the series by the glow of light 

 shed upon the table by the back lamp when this was turned on. We arranged 

 black curtains to cut off this diffused light; but there was still a faint glow 

 upon the walls of the room. 



^ Titchener, "Experimental Psychology," II., 1905, i., 104; ii., 258. 



^Dunlap, op. cit., 437. In Dunlap's use of the phrases, the reference is 

 always to the left or standard segment of the line. 



