170 COLOR-BLINDNESS IN ITS RELATION TO 



must be red, orange, yellow, or green, and the other blue, indigo, or violet. 

 It should very naturally be our object to give the i^reference in the 

 selection to the colors which niost strongly affect the eye at the time of 

 the comparison. Now, the most intense colors of the spectrum, that is 

 to say, the most vivid colors which enter into the white light of the sun, 

 are yellow and blue, one of each of the two groups. We then select them, 

 the more willingly, that the light of the lantern is, without any prepara- 

 tion, and to a very high degree, yellow, though it is not homogeneous. 

 But we are far from being so fortunate with regard to blue. We here 

 encounter a difficulty, on the contrary, which induces us to doubt 

 whether a change of colors will accomplish the desired end. 



On all colored surfaces — flags, paintings, semaphores, etc. — employed 

 by railways to reflect during the day the sunlight or daylight, the pro- 

 posed colors answer perfectly without any doubt, and, in all probability, no 

 color-blind individual of the kinds specified would nominally make mis- 

 takes of judgment. But the night-signals are quite another matter, and 

 are by far more important for many reasons. This is therefore why we 

 prefer to attach here so much importance to them, as during the day a 

 multitude of different circumstances might give warning of danger, 

 while during the night the colored light is the only signal which indi- 

 cates it. 



The colored lights used for night-signals are made, as all know, by plac- 

 ing colored glass before the flame of a lantern. The use of Bengal lights 

 as regular signals could scarcely be introduced into practice. Now, a 

 colored glass produces a colored light, because, of all the kinds of light 

 radiated from the flame, but one kind (or, at least, mainly one kind) is 

 allowed to escape, while all the others, or a greater part of the others, 

 are absorbed by the glass. Thus, blue glass, according to its thickness 

 or degree of coloring, absorbs all the other kinds of light emitted by the 

 flame of the lantern, allowing only the blue rays to escape. But, un- 

 fortunately, as is well known by direct experience, the flame of the lan- 

 tern emits comparatively but a small amount of blue light when rape- 

 seed oil and photogene, or generally any of our ordinary sources of 

 artificial light, are employed ; and this is why all appear yellow or red 

 when compared with the light of day. Under such circumstances, blue 

 glass can naturally transmit only a small amount of blue light; and the 

 light of a' blue lantern must consequently always be very feebly lumi- 

 nous. 



What we have just said of blue applies equally to indigo and violet. 

 The proposed changes of the colors of signals furnishes, therefore, 

 but two colors in place of three; and then one is a very feeble light, 

 so that it is difficult to see it far off" so long as it is sufficiently colored. 

 This state of things scarcely holds out much inducement to introduce a 

 reform of this nature. And it seems the more dangerous that this change 

 of colors in the signals would cause those with normal sight amongst the 

 personnel to run the same risk that the really color-blind do; I mean that 



