38 



FREDERICK S. BREED 



tative relations involved. E.g., on the second day, out of fifty 

 trials the object was struck, on an average, 33.48 times on the 

 first attempt. Of this number only 16.96 included seizing; and 

 of the latter only 8.67 were reactions in which the object was 

 struck, seized, and swallowed. As the chicks advanced in age, 

 each of these sets of figures approximated the same limit, but 

 naturally they preserved the original order as to size, for with 

 each additional reaction in the series there was an added source 

 of error. 



Reverting to the earlier consideration that Curves I, II, and 

 III may be regarded as error curves, an examination of these 



ai/e 12 3 



10 11 12 13 14 15 



Figure 7 — Curves of development of the peeking instinct, based on the averages 

 obtained from the records of twenty-one chicks. Data in table 7. Distances 

 along axis of abscissae represent days of age ; distances along axis of ordi- 

 nates, the number of occurrences of a given type of reaction in a daily series 

 of fifty pecking reactions. Curves I, II, and III trace the course of reactions 

 1, 2, and 3, respectively. Curve IV shows the average improvement in 

 accuracy of reaction 4. 



curves shows that the improvement of pecking^accuracy is 

 retarded more by failures in seizing than by failures in swallow- 

 ing ; and more by failures in swallowing than by failures in strik- 

 ing. After the third day, the difiiculty of seizure remains greater 

 than that of the other two reactions combined. The striking 

 reaction improves in accuracy rapidly and without relapse, 

 closely approximating perfection by the fifth day — a degree of 

 accuracy that might easily inspire belief in the perfection of 

 instinct, for, indeed, this is the reaction which has no doubt 

 been central in the earlier discussions when accuracy has come 

 into question. 



