Eyes of Molluscs and Arthropods. 727 



The diiference between the refractive index of the nuclei and that of 

 the suiTounding" bodies is g-reater than that between the sheaths of the 

 rods and tbeir axial eores ; bat tlie light passes withont any disturbing- 

 effect through several layers of larg-e nuclei and thick cell walls, so 

 that it is not probable that the rods, which are much more homogene- 

 ous, would cause any breaking: up of tlic rays of light. 



As we construe the facts, the result would be that the rays of light 

 from any externa! object are brought to a focus upon a definite niveau 

 in the layer of rods, and there aet upon a certain layer of cross, retini- 

 dial fibrillae, produciiig at that point an impression of an Image ; the 

 divergent rays then impinge upon the concave argentea, and, by re- 

 flection, come again to a focus in exactly the same place as before, and 

 therefore upon exactly the same set of cross fibrillae. In all parts of 

 the rod layer . except at the common points wliere both incident and 

 reflected Images are formed, no di st inet impression will be produced 

 except that of light. 



Chapter VII. Reniarks upon the Classiflcatiou of Eyes. 



The facts and deductions stated in the foregoing pages tend to 

 show that all eyes consist of ommatidia, upon the construction and 

 arrangement of which depends the perfection of the eye. Therefore any 

 Classification of eyes should be founded upon the modification of the 

 ommatidia. But it is certain that many complex eyes have originated 

 independently, in very limited groups of animals. Therefore we can 

 classify eyes, ( I ) as organs which attain essentially the same ends by 

 various means. »Such a Classification would entirely ignore the genetic 

 relation of the animals possessing the various types of organs. (2) We 

 may classify eyes — having principally in view the genetic relationship 

 of their possessors. On account of the frequent, independent develop- 

 ment of eyes in closely related forms, the latter method cannot be 

 logically applied except in restricted groups, and even then with con- 

 siderable uncertainty. 



Grenacher (39) is inclined to lay great stress upon the fact that 

 the retina in Cephalopods and Heteropods is single layered. His ar- 

 guments in favor of his opinion are remarkable ones. He says, p. 216: 

 »Am prägnantesten gebe ich vielleicht dem scharfen Gegensatz, in 

 dem ich mich darin zu meinen Vorgängern finde, dadurch Ausdruck, 

 dass ich überhaupt eine «Schichtung« der uns hier beschäftigenden 

 Retina nicht anerkennen kann, und den Gebrauch dieses Wortes als 



