xliii 



As regards the compound eye of insects, Dr. Grenadier 

 adopts the mosaic theory of Miiller, — that is to say, he does 

 not consider that each facet produces an image, as is the case in 

 our eyes, but corresponds to a single point in the field of view. 



In certain species, between the posterior end of the crystalline 

 cone and the front of the perceptive apparatus, is a narrow con- 

 striction, which is sometimes considerably produced, so that the 

 formation of an image would seem to be physically impossible. 



Again, the formation of an image would require a power 

 of accommodation for different distances, but he could find 

 no trace that any such power exists. 



Another objection is the extreme difficulty which would exist 

 of combining so many different images into one idea, though it 

 must be admitted that at first sight this difficulty (though to 

 a minor degree) exists even in the case of simple eyes, the 

 number of which varies considerably. Spiders have six to eight ; 

 some aquatic larv£e twelve ; while the Oniscoidea, assuming that 

 these eyes are aggregates of simple eyes, as Miiller supposed, 

 have as many as twenty to forty. These, however, take in 

 different parts of the field of vision. 



The principal reasons which h^^ve led Dr. Grenacher to decide 

 in favour of Miiller' s theory of mosaic vision are as follows : — 

 In certain cases there is no lens, and consequently there can be 

 no image ; in some it would seem that the image would be 

 formed completely behind the eye, while in others again it would 

 be too much in front. Another difficulty is that any true 

 projection of an image would in certain species be precluded by 

 the presence of impenetrable pigment, which only leaves a 

 minute central passage for the light-rays. In all cases moreover, 

 without any exception, even the sharpest image would be useless, 

 from the absence of a suitably receptive surface ; since both the 

 number and mode of combination of the elements composing that 

 surface seem to preclude it from receiving more than a single 

 impression. He concludes, therefore, that the image theory must 

 be definitely abandoned. 



If these views are correct, we come to the interestmg result 

 that while the image produced on the retina of the ocellus must 

 of course be reversed as in our own eyes ; in the compound eyes, 

 on the contrary, the vision must be direct. That the same 

 animal should see some things directly, and others reversed ; and 



