METHODS OF STUDYING VISION IN ANIMALS 615 



wide enough to allow the whole spectrum to pass through, 

 I cm. high and 8 cm. long, is cut in this wall. A part of this 

 window is shown at O. The window is seen more clearly in 

 the back view^ of the mounted slit, shown at W in fig. 9. The 

 two thin metal strips upon which the scales Sc and Sci are ruled, 

 6 mm. wide, 1.5 mm. thick and 20 cm. long, are beveled at an 

 angle of 45°, and are attached above and below the window in 

 such a way as to form a grooved track for holding the slit jaws. 

 This track is indicated at 11 and 12. Two large slit jaws, 

 J, and J 3 knife-edged along their vertical surfaces, slide in this 

 track. Attached to these jaws are two shouldered nuts, K 

 and Ki. By turning the calibrated screw-heads Cal and Calj, 

 the two jaws may be made to advance or to recede. Without 

 the addition of jaws Ji and J 2 the apparatus may be used as 

 a single optical slit. When these two jaws are inserted, a double 

 slit is provided which allows two . beams of monochromatic 

 light to be admitted. These two jaws are also grooved to slide 

 in tracks 11 and 12. They are held in their position by 

 means of small vertical spring clips (c and d in fig. 9), fastened 

 firmly to the center of each in such a way that the two ends 

 of the spring work constantly against the back face of the slit- 

 wall. The clip, combined with their snug fit in the grooved 

 track, holds the jaws quite firmly in a vertical position. Ths 

 jaws J and J 3 are provided on their front faces with indices, 

 I and 1 1, fig. 10, which give their position with respect to the scales 

 Sc and Scj respectively. These scales are needed for controlling 

 the width of the slit and for setting the openings at desired 

 positions. 



After the apparatus has been permanently installed, it is 

 very desirable to calibrate the double slit in terms of wave- 

 lengths. For this purpose it is best to rule the scale Scj for the 

 whole of the length of the strip. With the sHt in position one 

 then burns in the lamp (using the arc between the two carbons, 

 instead of the crater of the positive carbon as the source) a 

 positive carbon soaked in a strong salt solution. A well-definsd 

 D-line appears, the position of which is noted on the scale. 

 Several other metalHc salts, e. g., lithium and barium, are then 

 successively burned in the same way, and their lines noted 

 on the scale. Some ten or twelve such determinations are 

 suiTficient to enable one to plot the wave-length curve of the 



