STUDIES ON SPECTRAL SENSITIVITY OF BIRDS. 



93 



Pfund has found two filters which work well in the determination of limits. 

 The filter for the red end consists of an aqueous solution of cyanin and potas- 

 sium bichloride placed in a convenient glass cell with plane parallel sides. 

 Such a cell was phiced just behind the selecting unit, which was set at \ = 7000. 

 This filter transmits 80 per cent of the light from X = 7000 up to Iju, but 

 excludes all light below X = 7000. When placed behind the slit, the red surface 

 was clearly visible to me after 20 seconds of darkness adaptation.* The slit 

 in the collimater was set at 0.75 mm. and that in the objective at 0.52 mm. 

 The complete records of the chicks are shown in table 26. 



On the following day I set the slit at X = 7150 with other conditions exactly 

 as before. The light was quite faint to me. After 5 minutes' adaptation to 

 darkness the red surged in quite clearly. 



Table 26. 



Chick D chose correctly 3 times, then made 3 consecutive errors, after which 

 he refused to resjmnd even after long adaptation. I then turned the slit back 

 to X = 5480 (green of original training series) and gave him 5 trials; all were 

 correct and the average time of reaction was 2 seconds. 



Chick R chose twice cori'ectly and then made three consecutive errors. 

 When tested on green of training series, 5 correct choices were made in an 

 average time of 2 .seconds. 



Chick G would give only 8 resjionses; during these 8 trials two errors were 

 made. The animal chose with great difficulty. The average time of this chick 



♦When it appeared it came in a burst of red light, 

 whatever. 



There was no photochromatic interval 



