18 D. B. CASTEEL 



and was able to make distinction with a fair degree of accuracy 

 when the lines were reduced to 2 mm. In the series of six tests 

 (shown in table 8) with lines 4 mm. wide an average of 95 per 

 cent of the trials were correct. When the lines were reduced 

 to a width of 2 mm. on both boxes the first five tests resulted 

 in 80 per cent of the trials being correct. With a continuance 

 of 2 mm. line tests this percentage was lowered in five tests to an 

 average of 68 per cent. It would thus appear that 2 mm. is 

 about the limit of this turtle's discriminative ability for line 

 markings running in different directions. Turtle No. 4 was 

 long in learning the difference between the 8 mm. line boxes. 

 Its first five tests with 4 mm line boxes gave a record of 80 

 per cent correct choices, and this was raised to 90 per cent in 

 the five succeeding tests. With reduction of lines to 2 mm. 

 the average of successful trials dropped to 54 per cent when the 

 sulky disposition of the animal brought the experiment to an end. 



TABLE 8 



Turtle No. 10, Vertical and Hori- 

 f zoNTAL Line Discrimination 



Restdt of preference trials: Vertical 18, 

 horizontal 22. Fed in vertical 



Lines 4mm. wide 



Test Right Wrong 



1 10 



2 9 1 



3 9 1 



4 10 



5 9 1 



6 10 



57 



Lines 2mm. imde 



1 10 



2 6 4 



3 8 2 



4 9 1 



5 7 3 



40 10 



Lines 2mm. ivide (cont.) 



1 6 4 



2 8 2 



3 4 6 



4 8 2 



5 8 2 



34 16 



