634 



HANDBOOK OF PHYSIOLOGY 



NEUROPHYSIOLOGY I 



cells remain small and transparent but secrete distal 

 to themselves a fluid or paste extending to the corneal 

 lens and supposedly aiding in the refraction of light. 



At the proximal end of the dioptric mechanism 

 within the ommatidium is a ring of receptor cells, 

 or two rings, one distal to the other. Two rings may 

 be more primiti\'e than one ring. In Limulus and some 

 others, an eccentric receptor cell lies outside the ring 

 but extends a terminal segment toward the dioptric 

 system in a position central to the ring of other recep- 

 tor cells. More commonly there is no eccentric cell, 

 and the ring of receptors secretes a translucent rod 

 in the core position as a rhabdom which conducts 

 light energy along the optical axis to the more proxi- 

 mal parts of the receptor cells. 



In rendering these dioptric structures visiljle in 

 sections through a compound eye, it is customary to 

 bleach the pigment from the cells which sheath each 

 ommatidium. Exner (68) appears to have forgotten 

 the existence of this pigment mantle while tracing the 

 path of light rays, van der Horst (256) drew attention 

 to the pronounced diaphragmatic effect of the mantle 

 in many compound eyes, limiting the passage of light 

 to so small an aperture that no image could be pro- 

 duced at the receptor level. Under these circum- 

 stances an indi\'idual ommatidium could do no better 

 than serve as a photometer. Compound eyes which 

 are used in daylight operate in this way, with each 

 ommatidium isolated from its neighbors, and any 

 picture of the outside world a synthesized one in the 

 central nervous system, built on the mosaic of photo- 

 metric information coming from the indi\idual om- 

 matidia. 



Notthaft (202) took the extreme view that each 

 ommatidium operated on an all-or-none principle. 

 Either a target was included in its visual field enough 

 to stimulate the receptor system, or not. Almost 

 certainly this view is too severe. In Limulus, where the 

 compound eye may be somewhat degenerate, the 

 optic nerve fibers lack lateral connections and gan- 

 glion cells for a distance from the eye sufficient that 

 electrodes can be applied and the response of indi- 

 vidual ommatidia studied (75, 89, 90, 286). In 

 juveniles, two or more nerve fibers per ommatidium 

 may carry nerve impulses when the eye is illuminated, 

 but in adults only one is conducting, seemingly the 

 one arising from the eccentric cell. A wide range of 

 sensitivity and of response is evident. But the function 

 of the 9 to 1 9 other receptor cells in each ommatidium 

 remains unknown. Both at threshold and under in- 

 tense illumination, the ommatidium discharges im- 

 pulses as a unit. 



The directional sensitivity of single ommatidia in 

 the compound eye of Limulus has been evaluated using 

 the same electrical technique (282). Sensitivity is 

 highest on the optic axis and falls off to a tenth or less 

 for light sources 10 to 20 degrees on any side. The 

 effectiv-e aperture of the ommatidium from a physio- 

 logical point of view is thus to 40 degrees for high 

 sensitivity and to 180 degrees for response to stimuli 

 as much as four log units above threshold intensity. 

 Yet the maximum angular separation of Limulus 

 ommatidia is about 15 degrees, the minimum 4 to 5 

 degrees. Hence the overlap of visual fields of neigh- 

 boring units must be extensive and the acuity which 

 might be predicted (as Notthaft did) on the basis of 

 number of ommatidia is probably not realized. Since 

 the dioptric mechanism of the Limulus ommatidium 

 is somewhat different from that of most other arthro- 

 pods, however, these findings may not apply widely. 

 Acuity may be far better elsewhere in the phylum. 



The compound eye seems particularly efficient in 

 detecting movements in its total visual field. This 

 can be demonstrated under field conditions (34) or as 

 a sensitivity to flickered light in the laboratory (298- 

 301). When plotted on a probability grid, flicker- 

 fusion curves are like visual-acuity curves in being 

 essentially straight lines (298, 299). This may be due 

 to a normal statistical distribution of sensitivities 

 among the ommatidia; or it may arise through the 

 recruitment of progressively more ommatidia in a 

 convex eye as the intensity of stimulus rises. Crozier 

 & W'olf (42) believed that the latter was the limiting 

 factor in the crayfish Camharus. 



The intensity difference required for flicker detec- 

 tion by arthropods is greater than that for the human 

 eye. At optimum intensity the honeybee requires one 

 stimulus to be 25 per cent greater or less than the 

 other (298, 299). For the fly Drosophila the difference 

 must be of the order of 225 per cent (98, 99). For 

 man 1.5 per cent is adequate in good illumination. 

 Hence the visual field of the arthropod eye contains 

 a gray scale with far fewer than the 500 steplike 

 increments between black and white detectable by 

 the human eye. 



Evaluation of stimuli effective with a compound 

 eye is more satisfactory if it can be made from elec- 

 troretinograms rather than kinetic responses of the 

 entire animal. Electrical records of this kind are pos- 

 sible either with a surviving eye (56) or an intact 

 animal (87). Antrum & Stocker (7) learned with this 

 technique that insects show two \ery different ranges 

 in flicker detection. The fly Calliphora, the wasp Vespa 

 and the honeybee Apis responded to rates as high as 



