METHODS OF STUDYING VISION IN ANIMALS 29 



■which, instead, yields a line spectrum. Since the particular 

 method employed to obtain a monochromatic stimulus deter- 

 mines the nature of the source which may be used, it is neces- 

 sary to make special recommendations for each of the several 

 methods of investigating color vision. 



For all color vision experiments it is highly desirable that 

 the source provide a light ^Yhich is natural (like sunlight in 

 quality) constant, controllable, and measurable. 



a. Sun. — Sunlight is perfect as to naturalness (quality) and 

 it certainly should be used extensively for comparison experi- 

 ments in connection with the use of artificial lights. For accu- 

 rate quantitative experiments it is at present extremely unsatis- 

 factory because of (a) uncertainty" of supply, (b) fluctuations in 

 quality and intensity, and (c) uncontrollability. 



b. Nernst lamp. — The Nernst lamp yields a light whose quality 

 is excellent and whose intensity, during the last half of the life 

 of the glower, is fairly constant. What is lost in naturalness 

 by the use of this source, instead of the sun, is more than com- 

 pensated by the gain in constancy of f[uality, intensity and 

 controllability. For qualitative experiments the Nernst glower 

 is admirable, if intelligently handled; and for quantitative 

 work it is probably the best source now available. It can not, 

 however, be used for extremely high intensity work. 



c. Tungsten lamp. — Of the medium power electric lamps the 

 tungsten appears to be the most satisfactory for investigations 

 of color vision which do not demand stimuli of high intensity. 



d. Acetylene. — In the absence of electricity, the acetylene 

 lamp may be used to advantage in the study of color vision. 



e. Carbon arcs. — The crater of the open carbon arc yields a 

 light of eminently satisfactory quality, but the remainder of 

 the light is unsuitable for visual work. A well constructed 

 arc-lamp, burning high grade carbons, on a well controlled 

 circuit may be made to supply a large amount of light of satis- 

 factory quality and in fairly constant intensity, provided the 

 crater alone be used. 



/. Recommendations concerning sources for tests of color vision. — 

 In general, as sources of light for the study of color vision we 

 recommend the following in order of preference (a) sun, (b) open 

 carbon arc, (c) Nernst, (d) tungsten, (e) acetylene. These 

 sources are not, however, equally suitable for all methods. 



