1907] TITCHENER AND PVLE— JUDGMENT OF DISTANCE. 



99 



Expcriiiicnt I. — What, now, of the results? Dunlap quotes no 

 figures. He merely says that '' a set was counted ' for ' or ' against ' 

 the illusion according as the difference between the mean equality 

 points was or was not in the direction which would correspond to 

 the possible effect of the illusion-figure."^ And he tabulates his 

 results as follows : 



Nothing is here said of the magnitude of the illusion-effect. If, 

 however, this point is disregarded, and our own results are treated 

 as Dunlap prescribes, we obtain the following table : 



It is plain that the two sets of results are not in agreement. Since 

 apparatus and method were practically the same, and care and skill 

 in the conduct of the experiments may be assumed to be equal, we 

 can only conjecture that the positive testimony of Dunlap's series is 

 due to chance, operating on a small group of observations. 



Experiment II. — It would, naturally, have been simply a matter 

 of time to extend these series to a point at which the effects of 

 chance should be ruled out. It seemed advisable, however, to 

 modify the procedure. Dunlap himself refers to the experiments 

 above reported as " preliminary to the investigation proper," and 

 asserts that the apparatus was "' not very satisfactory in its opera- 

 tion " :- though he gives no details. We found two principal sources 

 of error. On the one hand, we feared that the presence or absence 

 of light behind the screen, as the change was made from shadow- 

 series to shadowless, might influence the observer's judgment. We 



'Op. cit, 439- 



2 Op. cit, 439, 436. 



