STUDIES ON SPECTRAL SENSITIVITY OF BIRDS. 



97 



the steps so that a satisfactory value could be obtained and verified in the 

 course of 12 to 16 trials. In several cases where doubt of the correct setting 

 of the sector was possible, I delayed advancing on the following day to a new 

 wave-length, in order to be able to repeat the experiments. It was not deemed 

 advisable to use an ascending series of steps in this work, i. e., to start below the 

 threshold and gradually increase the intensity of the light until positive 

 response was obtained (for fear of onset of position habits). If more time 

 could have been given over to the experiment such a technique would unques- 

 tionably have been possible. 



After the day's experiments upon the three chicks were over (the experi- 

 ment lasted roughly an hour) my own threshold was taken. My wife adjusted 

 the sector. The procedure was as follows : My eye was kept 1 meter from the 

 plaster surface; the tantalum light was turned up for 5 seconds. While it was 

 on I glanced at a neutral gray paper during the first 4 seconds (in order to 

 avoid strong after-image). On the fifth count (metronome) I again fixated the 

 plaster surface and shut my eyes while the light was turned out for 5 seconds. 

 On the sixth second I opened my eyes and observed the plaster surface. If 

 it were visible the experiment was repeated until the plaster surface was not 

 visible immediately upon opening my eyes. This was repeated five times 

 with ascending series and five times with descending series and the results were 

 averaged. 



Table 32 shows the actual sector openings for chick and man and the re- 

 ciprocals of those values. 



TABLE 32. 



The curve (fig. 10) is plotted from the reciprocals. It is clear that the 

 higher the energy required for any given monochromatic light to stimulate 

 the retina, the less the sensitivity of the retina to that light; that is, the recip- 

 rocals of the energy value (1/e, where e equals the energy necessary for stimula- 

 tion) will, when plotted against the wave-length, give the usual form of a 

 sensitivity curve. 



In examining the curve it is well to remember that the higher the point on 

 the curve is above the base line, the less is the energy required to produce 

 stimulation. In this work we do not know in absolute terms (ergs) the stan- 

 dard energy carried by each beam with which we worked, but we do know that 

 all beams had the same initial energy and consequently that the final sector 



*The opening at 660 w is of course merely proportional. At all other wave-lengths a smoked 

 wedge was used after the energy was calibrated, the transmission curve of which, from 396 n/j. to 

 760 /i/x, was known (calibrated by Bureau of Standards). At 660 /iju the efficiency was so low 

 that the wedge had to be removed. 



