INVERTEBRATE PHOTORECEPTORS 657 



eyes of lepidopterous larvae; these organs were called '•composite ocelli" 

 by Landois (1866) but were thought to be ommatidia by Pankrath (1890), 

 Redikorzew (1900), and Sanchez (1926). Corneli (1924) has identified 

 corresponding structures in larvae of the hymenopterous family Tenthre- 

 dinidae, and Constantineanu (1930), in some larval dipterans. Schmitt- 

 Auracher (1923) reported pigment migration in ommatidia of Ewproctis 

 caterpillars. Werringloer (1932) has contended that the compound eyes 

 of ants are unlike those of other insects both in histology and in embry- 

 ology. Neither Schmitt-Auracher's observations nor Werringloer's con- 

 tention has received support by later investigators. 



Attempts to find phylogenetic relationships among insect compound 

 eyes have led to papers by Chatin (1876, 1878) and Lankester and Bourne 

 (1883). Patten (1890) tried to prove that the ommatidium was a modi- 

 fied hair-bearing "sense bud," but his whole idea was based upon false 

 premises: that the nerve fiber entered the crystalline cone and that the 

 cone was the photosensitive element (Patten, 1887b). Lowne (188-1) 

 also held this view, although he recognized that the diaphragming effect 

 of the pigment cells (pointed out by Will in 1840, 1843) argued against it. 

 -Viallanes (1892) followed Schultze (1867) in maintaining that the cones 

 were chitinous and nonnervous and that the retinula cells grouped around 

 the end of the cone must be the photosensitive part. Some of these 

 difficulties in interpretation were not fully cleared up until the studies of 

 van der Horst (1933). Debauche (1942) suggested possible homologies 

 between retinula cells and sensory elements of other receptor systems. 

 The association between tactile hairs and ommatidia, pointed out by 

 Hertweck (1931), lends itself to broad generalizations such as those of 

 Patten and Debauche. 



Except for specialization of "turban" eyes, such as were described 

 above in May flies and some other insects, the normal number and 

 arrangement of compound eyes in this arthropod class are two, one on 

 each side of the head. In the whirligig beetle Gyrinus the eye on each 

 side is completely divided, with one part facing into air, the other under- 

 water and facing the bottom. Some other beetles have divided eyes, but 

 the cause is obvious in each instance: excessive encroachment on the eye 

 area of the head by enlargement of the antennal base. 



Experimental study of insect vision as a means of learning how the 

 compound eyes function seems to have begun with the work of de Serres 

 (1813, 1814), who used black varnish to cover eyes and parts of eyes and 

 observed behavior changes. De Serres also introduced the term "false 

 pupil," although his description of it was not so complete as that of 

 Ewing (1826). Duges (1830) concluded that binocular vision was possi- 

 ble in insects, but how useful it could be was far from clear. Dor (1861) 

 suggested that the more extensive visual field of even one compound eye 

 might compensate for lack of accommodatory ability, so that, with many 



