302 Sir W. cle W. Abney. On the Variation in 



media, and he generalised from his experiments, that the smaller the 

 wave-length, the less steep was the gradation, the ultra-violet rays 

 giving the least steep, and the red the most steep gradation. My 

 experiments, which had at that time been partially completed, did not 

 bear out this generalisation to the full when pure silver salts were 

 used ; and my subsequent measurements with them show that the 

 least steep gradation is that given by the monochromatic light to 

 which the simple silver salt experimented with is most sensitive, and 

 that the gradation becomes steeper as the wave-lengths of light em- 

 ployed depart in either direction in the spectrum from this point, the 

 steepest gradation being given by the extreme red. The case of ortho- 

 chromatic plates in which is a complex mixture of silver salt and dye, 

 is necessarily less simple, involving considerations of the localities in 

 the spectrum to which the dye or dyes, together with that of the silver 

 salt, are most sensitive. For this reason the simple salts have been 

 experimented with in preference to the more complex organic com- 

 pounds. 



Methods of Experimenting. 



As pointed out in the opening paragraph, there are two ways of 

 experimenting, one where the illumination is constant, the times of 

 exposure being altered, and the other in which, the time of exposure 

 is constant, and the illumination is altered. This last is the condition 

 under which an image in the camera is photographed. It might 

 appear that both methods should give identical quantitative results, 

 but it was more than probable that they would not do so, from 

 the experiments that I had previously carried out with these two 

 methods with ordinary white light. 



The first set of experiments were with fixed time of exposure and 

 varying intensity of light. To obtain the varying intensity, a photo- 

 graphic plate was exposed to white light, the parts exposed being 

 limited to an area having the form of a triangle with the top cut off 

 at the apex, the two sides being radial to the centre of the plate. The 

 enclosed angle was about 20°, so that by turning the plate round its 

 centre, twelve different spaces would be exposed. After the plate had 

 been developed with ortol or ferrous oxalate, fixed, washed, and dried, 

 the intervals between the exposed parts were blocked out. The 

 opacities were then ready for measurement. Fig. 1 is a reproduction 

 of the " star " graduated opacities. 



Measurement of Star Opacity with different Colours. 



It became necessary to see whether the deposit obstructed light 

 equally for each ray of the spectrum, and the following arrange- 

 ment was adopted. The colour patch apparatus which I have 



