PATTERN-DISCRIMINATION IN VERTEBRATES 



173 



invariably to compare the test-fields and choose the coarser 

 system regardless of the absolute width of its members, it would 

 seem feasible to adopt the method of right and wrong cases, 

 presenting in haphazard order large and small differences in 

 both directions from the standard, and to treat the results 

 statistically. The advantages of using such a method in pre- 

 ference to the cruder methods on which we have thus far had 

 to rely are obvious. 



TABLE 1 

 Daily Record of Chick 2 



