328 H. M. JOHNSON 



more stimulus differences might be found to have an interesting 

 effect on the threshold. For example, if two systems of striae 

 were made to differ in width of their respective members, as well 

 in direction, the difference-threshold for direction might be lowered 

 even though the width-difference were subliminal when the two 

 systems lay in the same direction. In the present w T ork I have 

 limited the attack to the four elementary problems, under one 

 set of experimental conditions. The results are presented as 

 holding good only under those conditions. But it impresses the 

 writer that a successful method having been satisfactorily devel- 

 oped, a rich field is open for further extension. 



METHOD OF PREPARING AND PRESENTING THE STIMULI 



For investigating the first three problems a single instrument 

 has been used for preparing the test-fields, with very satisfactory 

 results. Since it has not been used before in animal experimenta- 

 tion, and since familiarity with its principles is necessary to adapt 

 it to such work, it requires detailed description. It consists 

 essentially of two systems of fine opaque lines ruled on glass, the 

 width of the lines and of the clear interspaces being equal. When 

 these two systems are rotated over each other about an axis per- 

 pendicular to the surfaces of the glass, a series of bright and 

 dark bands of equal width appear. The size and number of 

 these striae are reciprocal, dependent on the angle of rotation, 

 and are highly controllable. The direction of the striae is variable 

 at will. The instrument possesses certain peculiar features 

 which are especially advantageous in this work. The luminous 

 intensity does not change under experimental conditions. The 

 mean brightness of the field remains constant provided the area 

 selected include an equal number of dark and bright bands of 

 the same length, or if it be large with respect to the area of a 

 given band. Possible disturbances from constant differences in 

 olfactory stimuli are absolutely neutralized in all stages of experi- 

 mentation. Change from any stimulus- value to any other can 

 be made by insensible gradations. Large differences in pattern 

 can be made without changing any other stimulus-factor In 

 addition to these peculiar features, the apparatus is not unduly 

 expensive; it permits an instantaneous and practically noiseless 

 interchange of stimuli to be made from the operator's chair; 

 and its operation is quite simple. 



