ADAPTIVE CHANGES IX COLOR OF FUXDL'LUS. 65 



apparatus by means of which twenty-three lights were obtained 

 each thirty wave-lengths in width but equal in radiant energy and 

 extending by steps of ten from 420^ to 670^. The Hilger 

 deviation spectrometer was used to obtain spectral lights and the 

 radiant energy equated by means of a thermopile and galvano- 

 meter, thus determining the distance of the lamp from the slit of 

 the collimator, in order to give, equal energy for all sets of lights. 

 This apparatus was used to test the effect of spectral lights on 

 the color < lulus. It was necessary, however, to select only 



parts of the spectrum. In fact only two sections would 

 tri.-tly necessary, one of long wave-length the other of short 

 wave -length. Since the spectral lights are of equal energy, the 

 es should show the same effect if intensity is the sole factor, 

 ct varies in different spectral lights, it can be due only 

 to wave-'' Using the data provided by Laurens am! 



1 looker, the following spectral lights with corresponding condi- 

 tions in the apparatus were selected. 



a 1 ^it-source, a looo-watt Mazda lamp was used. The 

 lump was placed in an iron box 12 inches square at the base and 

 :iu-lu-s high. On the front of the box an aperture was made for 

 light. T.I secure ventilation apertures were made below in front 

 and near the top of the rear side; these were screened by sheets 

 of s. An asbestos funnel was also placed between the 



aperture of the lamp box and the slit of the collimator. A metal 

 cylinder lined with black paper and closely fitting to the ocular 

 of the telescope conveyed the light to the jar containing the fishes, 

 placed at a distance of 60 cm. The jar, measuring 10x5.5 cm. and 

 15 cm. high was covered with white paper on all sides except that 

 exposed to the spectral light and screened off from any light 

 Icaka-e from lamp box. The apparatus was set up in a large dark 



