90 THE APODID^E PART i 



istic elements in the Crustacean eye, we find them to 

 be (i) the crystal cones, and (2) the retinulse, z>., 

 definite groupings of a certain number of retinal and 

 pigment cells (see Fig. 23). In the first of these we 

 have a new structure, whose development must be 

 accounted for ; the second may safely be assumed to 

 be merely specialised hypodermal sensory (i.e., visual) 

 cells ; we have simply to account for the " retinula- 

 tion " of these cells, as Lankester calls the grouping 

 of them into retinulae. 



We assume, then, that first of all the crystal cones 

 were but slight irregularities in the thickness and 

 refractiveness of the developing cuticle. Under 

 these irregularities, i.e., under those which either con- 

 centrated the light or otherwise favoured its passage, 

 the visual cells would naturally tend to group. We 

 say naturally, because it is clear that, under places 

 through which the light but feebly penetrated, the 

 visual cells would be rendered useless. In process of 

 time, certain definite irregularities of the cuticle would 

 be selected and further developed as lenses, &c., for 

 collecting the light. We find in the eye of Limulus 

 the particular form of cuticular development which 

 may have given rise to the crystal cones of Apus (cf. 

 Fig. 22 with Fig. 23). This fact is particularly 

 interesting because we have already seen that 

 Limulus has retained the original position of the brain 

 in the bent Annelid. And here we find the same 

 animal supplying a form of eye which shows clearly 

 a possible origin of the Crustacean crystal cones. 

 We have only to assume that such conical processes 



