No. 383.] BASIS FOR A THEORY OF COLOR VISION. 835 
them. The ether waves thus vibrate across a series of fibrils 
of different lengths. 
We assume that the length and position of a fibril deter- 
mine the amount of its response to an ether wave of a given 
length, and as each visual cone contains a complete series of 
these fibrils, we may understand how each cone responds to 
the entire series of visible ether waves or to various combi- 
nations of them; and since the rods and cones vary in shape 
and in size in the retinz of different animals, or even in different 
parts of the same retina, there should be a corresponding varia- 
tion in the length and in the number of fibrils they contain, 
and consequently a corresponding variation in the powers of 
color vision. 
I shall first describe the structure and arrangement of the 
retinal cells in some invertebrates, omitting, of course, many 
details that cannot be introduced into a paper of this character. 
Then, assuming that the human retinal cells have a similar 
structure, I shall try to show how we may explain many phe- 
nomena of color vision on that basis. 
More than twelve years ago I described a series of what 
seemed to be nerve fibrils arranged with great regularity in 
the rods and retinal cells of mollusca and arthropods (Mitth. 
aus Neapel, 1886). Although from time to time I have con- 
firmed and considerably extended my original observations, no 
other investigator appears to have made a serious attempt to 
do so. But indirectly the observations have received ample 
confirmation, since my description of the nerve endings in the 
retinal cells are in complete harmony with the best recent 
results obtained from the study of other organs by the methylen- 
blue and the gold-chloride methods. 
We have not found these fibrils in the rods and cones of 
vertebrates, but there is no good reason to doubt their exist- 
ence there, since, as I shall show elsewhere, the retinal cells 
in both vertebrates and invertebrates agree in many important 
details. 
Whether these fibrils in the rods are nerve fibrils in ‘the 
usually accepted meaning of the term need not detain us. here. 
It will be sufficient for our purpose if we can show that these 
