RELATIVE PHYSIOLOGICAL VALUE OF SPECTRAL LIGHTS 513 



mm. high is fastened. The width of each slit is such as to allow 30 

 wave-lengths to pass. As is well kno^^^l. in ihe prismatic spectrum the 

 dispersion increases as the wave-length diminishes so that at the 

 violet end a certain number of wave-lengths occupy' a wider space than 

 they do at the red end. A wide diaphragm is therefore necessary at 

 the violet end and a narrow one at the red end to let through the 

 same number of wave-lengths. The correct width was determined 

 with the aid of the sodium line as follows: Given a desired range of 

 wave-lengths, e.g., 460 mm to 490 mm (see no. 5, table 2); the distance 

 on the wave-length drum between the scale marks 460 and 490 was first 

 measured with a strip of paper and the middle point determined. 

 (This middle point, in this case 474, gives the position at which the 

 drum must be set in order that the desired wave-lengths, 460 /x/x to 

 490 nix, pass). The middle point of the measuring strip of paper was 

 then placed at 589 on the drum and an equal distance measured off 

 on each side of this, thus determining the range of the wave-lengths 

 on either side of the D line, in this case 559 to 625. The diaphragm, 

 with a slit of arbitrary^ width, was then placed in position in the tele- 

 scope and the points noted on the drum at which the D line (sodium 

 flame) disappeared when the drum was rotated. The width of the 

 diaphragm slit was then adjusted so that the D line just disappeared 

 when the indicator on the drum pointed to 559 and 625 respectively. 

 Since equal revolutions of the wave-length drum rotate the prism 

 through equal angles this procedure assures that light ranging from 

 460 MM to 490 MM will pass when the diaphragm is in place and the 

 drum set at 474. The width of the diaphragm slit was thus deter- 

 mined for each of the twenty-three lights. 



Having procured the colored lights the next step was to equate the 

 radiant energy in each. This was done by varying the distance be- 

 tween the lamp and the collimator slit of the spectrometer until the 

 desired deflection of the galvanometer was obtained. A direct pro- 

 portionality was found to exist between the bands, as indicated by their 

 middle points, and this distance (see table 2). 



The intensity of the first four lights was increased by placing a 

 spherical reflector, obtained from the Spencer Lens Company, behind 

 the lamp. The spectrometer was placed so that the collimator slit 

 was at a distance of 9 cm. from the lamp, and the reflector adjusted so 

 that the reflected image of the lamp filament was focussed on the slit. 

 At this distance a radiant energ\' content equal to the amount chosen 

 as our unit value was obtained for light no. 1 (420 nix to 450 nn). 



