THE THEORY OF ARTHROPOD VISION. 561 



It further corresponds with the layer which I believe to be the 

 sentient retina (see PI. XL., Fig. i). 



The Erect Image in the Eye of Lampyris. Although Exner 

 rejects my view of the position of the retina, he gives a figure 

 of a section of the eye of Lampyris splendidula [252 Taf. i. 

 Fig. i] in which he represents the dioptron and retina exactly 

 as I do, only he terms the dioptron a vitreous. 



Exner further found an upright image of external objects 

 immediately beneath my membrana basilaris, and he not 

 only saw, but photographed this image. It is reproduced in 

 his book [252, frontispiece], but in its printed form it is re- 

 versed ; the reversal, however, is due, as he explains, to the 

 method of photo-printing employed. 



Exner's Theory of Super-position of Images. Exner supposes 

 that all the rays which fall from a point upon all the lenses of the 

 compound eye form, after refraction, the envelope of a caustic 

 curve, the cusp of which forms a point in the image and lies 

 upon the retina. He regards his so-called vitreous as a perfectly 

 transparent medium in which the caustic surfaces formed by 

 the intersection of the refracted rays lie. This theory might 

 be conceivable if such a vitreous existed in the compound eye, 

 but, unfortunately for the theory, no such vitreous exists in any 

 insect I have examined ; if it exist in Lampyris spendidula 

 the compound eye of this insect would be utterly unlike that 

 of any other Arthropod ; and there are indications in Exner's 

 figure that his so-called vitreous differs in no marked manner 

 from the great rods, as they appear in the recent eye, with the 

 pigment but little developed. 



With regard to his mathematical demonstration, he assumes 

 that the surface of the cornea is a segment of a sphere. Now, 

 so far as my observations go, it never is a segment of a sphere. 

 He further assumes that the light-rays leave the cone as a 

 parallel-rayed pencil, and his circles of confusion have approxi- 

 mately a diameter equal to the radius of the corneal facet, 

 roW f an inch or more. Such pencils would produce no 

 definite picture even if the super-position were perfect, which 

 is far from being the case when the cornea deviates from a 



372 



