20 D. B. CASTEEL 



lines of that width. Since in the previous experiments all of the 

 parallel lines on one box ran vertically and all on the other 

 horizontally, appreciation of difference might simply mean that 

 the turtle discriminated between the general direction taken 

 by the whole series of lines, and would be able to do this until 

 the lines were so reduced in width that only gray surfaces were 

 discerned. The results thus far given are of themselves inter- 

 esting in that they indicate the ability of these turtles to appre- 

 ciate the difference between fairly narrow line markings which 

 run in different directions; but they do not serve as an accurate 

 measure of the turtle's ability to distinguish between lines of 

 different width, nor do they give us an exact idea of the extent 

 of the animal's ability to distinguish relative widths of different 

 lines. 



So it seemed desirable to conduct another series of experi- 

 ments in which the lines on the faces and backs of the boxes 

 all ran in the same direction, either vertically or horizontally, 

 but differed in width on the two boxes. 



In these tests the turtle was first given choice between two 

 boxes, one of which bore lines 8 mm. wide and the other lines 

 much narrower (i mm. in one experiment, 2 mm. in the two 

 others described). The records of turtles Nos. 10, 8 and 11 will 

 be presented. Upon both of the boxes used in tests with No. 

 10 the parallel lines ran vertically; upon those used with Nos. 

 8 and 11 they ran horizontally. Unfortunately, after establish- 

 ing discrimination between 8 mm. and 2 mm. lines with turtles 

 No. 8 and No. 11, it was necessary to discontinue the tests. 

 Their records are found in tables 11 and 12. 



Turtle No. 10 was carried much farther. It is possible that 

 the former experiences with line boxes which this turtle under- 

 went may, to some extent, have so familiarized it with a similar 

 problem that its performance with these boxes would average 

 better than the rule. No. 10 was started with boxes bearing 

 lines 8 mm. and i mm. in width respectively. When this com- 

 bination was learned the 8 mm. box was refaced with boards 

 bearing lines 4 mm. wide. The choice then lay between a box 

 faced with i mm. lines and one faced with 4 mm. lines. The 

 I mm. lines had indicated the electric box in the preceding tests 

 and were so retained, while the 4 mm. lines replaced the 8 mm. 

 width on the food box. In fifty trials with these lines 96 per 



