6 K. S. LASHLEY 



of the chick's time of reaction and the time required by the 



experimenter to adapt to the stimulus lights, 1/45 of the standard 



intensity. The experimenter's adaptation was always lighter 



than that produced by looking directly at the 40-watt lamp 



for 10 seconds. 



Chick Experimenter 



22 seconds 20 seconds 



14 " 16 

 4 6 



18 " 10 

 6 4 



14 " 20 



2 3 " 

 6 " 2 " 



3 * 3 



When darkness adaptation was desired the chick was exposed 

 to light for only the moment when he pecked at the food. Be- 

 fore the experiments he was kept in the dark for one hour, this 

 being the time required for complete darkness adaptation 



(Hess '07). 



THE PURKINJE PHENOMENON 



Experiments were begun with two Bantam cocks, A and C. 

 They were first tested with white light and were found to be 

 markedly photopositive, invariably going to an illuminated 

 plaster surface. Two. white lights of different intensity were 

 then introduced. The chicks showed a tendency to choose the 

 brighter of the lights but their discrimination was very inaccurate 

 and when the dimmer of the lights was well above their threshold 

 they no longer selected the brighter: the reaction was evidently 

 to the light as an object. An attempt was then made to increase 

 the accuracy of their discrimination. Food was given at the 

 brighter light and no food at the other. Training under these 

 conditions was continued from October 23 until January 31 

 with ten trials per day. 



During this time it was found impossible to get the chicks 

 to distinguish accurately between white lights of relative in- 

 tensities less than ten to one (absolute intensities of 18 and 

 1.8 ca. m.). They showed a slight preference for the brighter 

 of two lights of intensities three to one (18.0 and 6.0 ca. m.) 

 but never made ten successive reactions without error. I 

 believe that the ratio of three to one represents very nearly 

 the difference limen of the fowl under the conditions of these 

 experiments. It is possible, however, that with a stronger 



