DISCRIMINATIVE RESPONSES TO VISUAL STIMULI 



471 



of the work. The absolute brightnesses of the various surfaces 

 can not be perfectly maintained without frequent photometry, 

 as the lamps deteriorate gradually with use. The variation in 

 this case was slight, as the lamps were not actually operated 

 over 100 hours after being seasoned, and those illuminating the 

 interior of C were operated at subnormal voltage. The relative 

 brightnesses were easily kept very constant, as the lamps were 

 connected in multiple on a steady circuit from large accumulator 

 cells, and the variations in potential after the resistances had 

 reached a stable temperature were not detectable with the ordi- 

 nary voltmeter. 



It will be noted that the absolute brightness of the test field 

 (measured with LI and L z in operation) varied between 23 and 



TABLE 1 



27 candles per square meter with the different brightnesses of 

 surroundings. This is due to light from the interior of C being 

 reflected into the eye of the observer from MGi, MG Z and MG$. 

 This could have been avoided by varying the distance between 

 L 3 and MGz to compensate, and it would have been done had 

 the purpose of the experiment been a test of the conditions 

 rather than a test of the method itself. The mounting of L 3 

 which was used did not permit of convenient and accurate 

 adjustment. 



The absolute brightness contributed by LI and L 2 to their 

 respective halves of the photometric field was 1J4 candles per 

 square meter. This, with respect to the total brightness of the 

 field was as follows : For condition BI, 4.68 per cent ; for condition 

 DI, 4.94 per cent; for condition B 2 , 4.16 per cent. This consti- 

 tutes a variability in the intensity of the stimulus to reaction, 

 which in the most precise work might cct be desired, but which 



