PATTERN-DISCRIMINATION IN VERTEBRATES 343 



ness used is a comfortable reading condition for the human subject. 

 We have no reason for assuming that it was or was not optimal 

 for any of the animals used. It will be remembered that each 

 test-object could be made to present a sensibly uniform or a 

 sensibly striate field by changing the angle of rotation of the 

 gratings over each other. This change is made between limits 

 determined by the setting of two stops with a micrometer screw, 

 by moving a lever attached to the mounting. Change of the 

 size of the striae does not alter the mean brightness of the field. 

 Hence the test-objects are not interchanged when it is desired 

 to interchange the fields, and any changes in brightness, range or 

 relative intensities of wave-lengths, odors, etc., which a test- 

 object might suffer, could not aid the animal in forming a dis- 

 crimination habit. They would be presented with the negative 

 field as often as with the positive field. For this reason repeated 

 photometric measurements are unnecessary. It is unlikely 

 that any changes take place save the decline in efficiency of the 

 lamps, which progresses slowly and at a fairly uniform rate between 

 wide limits. The writer observed no sensible changes in the 

 experimental conditions during the work on this problem. 



In these experiments I used the well-known discrimination 

 method, combining the motives of punishment and reward. By 

 referring to Fig. 6 of the earlier article cited, the reader may. see 

 the floor plan of the Yerkes box which I used. Food was placed 

 in both food-compartments, Rfb and Lfb. Entrance to the 

 food-boxes had to be made through alleys A 2 and A 1 respectively, 

 at the ends of which the respective test-fields appeared. Each 

 of the alleys was floored with a punishment grill essentially like 

 those described by Yerkes and Watson (op. cit. P.) and by 

 Yerkes. 4 The grills were hinged to the floor of the box at the 

 ends next the test-objects, and supported at the free end by a 

 light spring. A rod passed through a sleeve fastened to the under 

 side of the true floor of the box and was kept in contact with the 

 under surface of the grill by means of the supporting spring. 

 A mercury cup contact, insulated from the rod, was carried by 

 the latter. The circuit through the signal lamps indicated in 

 the figure was kept closed at this contact until the animal stepped 

 on to the punishment grill and depressed it. This broke the cir- 

 cuit through that signal lamp, and enabled the experimenter 

 4 Yerkes, R. M., " The Dancing Mouse." N. Y., Macmillan, 1937. 



