ACCIDENTS BY RAIL AND SEA. 169 



any reform that can be iutroduced at present in this matter sliould be 

 established just as if it were already decisively proved that color-blind- 

 ness is incurable. It is therefore this hypothesis that we intend to apply 

 in what follows. 



2.— MODIFICATION OF THE SIGNAL SYSTEM. 



(a) Other colors for signals. — If congenital color-blindness is incurable, 

 or at least if we know no actual remedy for it, it is necessary to devise 

 some other method (while retaining the color-blind in the employment 

 of railways) of guaranteeing the communications against any mistakes 

 they might commit with regard to signals. It is seen by what precedes 

 that these errors can and must occur in the use of the signal colors 

 generally aiopted, red, green (and yelloio). This choice seems there- 

 fore unfortunate. Wilson is of the same opinion. When red and green 

 color-blindness are the kinds of complete or typical partial blindness 

 which are most generally seen, it would seem that the difiBculty might 

 be considerably diminished, if, in place of using the actual colors, those 

 should be selected best suited to these kinds of color-blindness, although 

 they might not suit the third kind of typical partial color-blindness or 

 violet-blindness, which, according to the experience acquired up to the 

 present time, is much more uncommon. 



Undoubtedly, the principle wehave endeavored to establish would not be 

 radically enforced, but the practical result at least would be comparatively 

 nearly accomplished. As the color-blind has but two principal colors, or 

 two classes to which he can refer all the colors, it is evident that to select 

 two colors that he can recognize and distinguish without the least hesita- 

 tion, it would be necessary to selectone from each class. In this way, it is 

 always possible to bear in mind that each kind of color-blindness will 

 always be able to find two colors distinctly defined, but not more than two. 

 It is therefore necessary first to ask how far two colors for signals could 

 satisfy the demands of railways and the navy. As regards railways, 

 it is claimed, and it may be conceded, that, in case of necessity, and 

 perhaps without too great inconvenience, two colors might be made 

 to answer. It is certain that three colors are a great improvement 

 upon two. Let us admit, however, that two colors would answer, and 

 that it were desired to sacrifice the advantage of three colors for an- 

 other advantage, namely, that of retaining in the service of railways the 

 color-blind, there will still remain the necessity of making a good selec- 

 tion of these two colors. This is more easily said than done. The choice 

 must be so made that one color may be selected from each of the two 

 groups in which all the colors are classed according to the system for the 

 color-blind. Now, it is found, as we have already seen in the instances 

 of the principal colors of the red and green blind, that, amongst the seven 

 colors of the rainbow perceptible to the normal observer, four, namely, 

 red, orange, yellow, and green, belong to one class, and three, especially 

 blue, indigo, and violet, to the second. Consequently, one of the colors 



