74 



FREDERICK S. BREED 



certainly does not apply to the chicks tested. Of three animals 

 that required loo trials to perfect the black-blue modification, 

 one made a perfect persistence test; of two that required 70 

 trials, one made a perfect record; of three that required 60 trials, 

 two made perfect tests; and the chick that completed its work 

 in the shortest time also made a perfect record. The number 

 of trials given in the table is exclusive of preference and final 

 perfect series. In the error column is given for each chick the 

 total number of errors in this number of trials. Of five chicks 

 that made twenty or less errors, four had perfect persistence 



iSt/ccesses 



\Ser/es A 



8 9 10 11 12 13 14 15 16 



Figure 1 1 — The success curve for nine chicks of the same age in black-blue training, 

 plotted from averages based on data given in tables 13 and 20. Distances 

 along axis of abscissae represent series ; along axis of ordinates, right or suc- 

 cessful trials. 



tests; of four that made more than twenty errors, one had a 

 perfect persistence test. 



One may well question the appropriateness of the expression 

 " trial and error " intended to be descriptive of this method of 

 learning. If the number of errors is included in the number of' 

 trials, as for example in this paper, then the original term is 

 awkward if not illogical. These animals learn through successes 

 as well as failures, and the sum of the number of both equals 

 the whole number of trials. In this sense the method is more 

 accurately described as a method of success and failure. Again, 



