PHOTORECEPTION 211 



The first is a thick converging concavoconvex or biconvex meniscus 

 lens. 



If the secretion products of the three corneagen cells fuse incom- 

 pletely during the formation of this lens, as is typical of certain ocelli, 

 the cornea acts as a triple lens and forms three images (PL IV). The 

 crystaUine lens, like the cornea, is a thick, converging biconvex lens, 

 either unitary or tripartite to correspond with the cornea. It also forms 

 more or less distinct, real, inverted images. The receptive layer, like 

 that of the ommatidium, consists of distal and proximal retinula cells. 

 Some of the optical constants obtained from the lenses of the six ocelli 

 of the caterpillar Isia Isabella are given in Table 9. The peculiar optical 

 properties of the lens system of ommatidia as described by Exner 

 (1 89 1) are not characteristic of these lenses. 



Whether the ocellus acts as an eye or merely as a photoreceptor 

 depends primarily on : (1) the ability of the dioptric apparatus to form 

 images ; (2) the location of the image plane with respect to the rhab- 

 dom; (3) the ability of the rhabdom to receive any images formed. By 

 determining the image space (that region of space in which all possible 

 positions of the image are situated) from the optical constants and by 

 actual measurement of the length of the vertical retinal elements, 

 Dethier has shown (Fig. 99) that regardless of the distance of an 

 object from the ocellus, the image falls somewhere along the length 

 of the rhabdom. In this manner a degree of accommodation is obtained 

 which would otherwise be impossible with a fixed-lens system. The 

 ocellus is analogous in this respect to a fixed-focus camera. 



Vision in insects bearing simple eyes is the sum of the capacities 

 of all units operating jointly. Hundertmark's (1936, 1937) experiments 

 have shown that larvae which orient to black shapes on a white back- 

 ground are directed towards the black-and-white boundary. Dethier 

 has postulated that the basic principles of mosaic vision apply equally 

 well to lateral ocelH and compound eyes. Since each ocellus gathers 

 light from the area at which it is directed and concentrates this light 

 at some point along the vertical rhabdom, the six pairs of ocelli 

 together form twelve points of light. The ocelli are so arranged on the 

 head that little or no overlapping of the visual field exists, hence each 

 spot represents the intensity from a different area. Combined they 

 provide an exceedingly coarse mosaic of intensities. The paucity of 

 units is compensated for in part by the habit, characteristic of many 

 larvae, of moving the head from side to side while advancing. By this 

 klinotactic-like behaviour a larger visual field is examined, and the 

 recognition of changes in intensity, as at a black-white boundary, is 



