THE THEORY OF ARTHROPOD VISION. 559 
It is evident if we regard the great rod as a receptive end 
organ that the light must either act at some part or through- 
out the whole length of the rod. . 
In the latter case the same objections hold good as in Miiller’s 
original theory, such a condition would only give anything 
like vision with pencils of parallel rays; the same thing is 
true if the inner portions of the great rod are the seat of 
stimulation. 
If the outer extremity of the great rod is the receptive 
structure, it would receive its illumination from the subcorneal 
image, and the light rays from every part of this image would 
fall upon every part of the extremity of the great rod, hence 
the smallest distinct visual point would be the representative 
of the subcorneal image. And as this image is similar in all the 
adjacent facets, there could be no distinct mosaic representing 
its several parts, and hence all acuity of vision would vanish. 
Miller's theory, even in its modified form, is therefore optically 
untenable. 
Exner’s Views.—S. Exner has done great service in his optical 
investigation of the compound eye. He does not ignore, as 
Notthaft and Grenacher do, the refractive structures, and he 
has shown that an image may be formed by refraction in the 
absence of convex surfaces. 
Refractive Cylinders.—S. Exner investigated the subcorneal 
image which exists in the cone of Dytiscus. The corneal 
facets of this insect are very slightly convex, and he states [252] 
that the subcorneal image is but little affected by the medium 
in which the cornea is immersed, a condition which he explains 
by his theory of refractive cylinders. 
Exner found that the entire crystalline cone of Dytiscus 
produces an image at a focal distance which led him to deter- 
mine its refractive index as 1°8 [245]. I found formerly the 
refractive index for the corneal facet of a Hornet (Vespa crabro) 
to be equal to 2°0 ina similar manner. Such high refractive 
indices are unknown amongst organic substances, and indicate 
an error. Exner [285], therefore, investigated thin slices of 
the cone, and found that they have a refractive index equal 
37 
