AND THE MOEPHOLOGY OF THE EYE IN INSECTS. 409 



elaborate work, only represents portions of it in six figures ; and these are all diagram- 

 matic outlines, -with but little detail. 



M. Berger*, who has given the best figures and descriptions, has only worked with 

 comparatively thick sections with low powers, and has not only overlooked the bacilla, 

 but many other very obvious details. 



On the otber hand, it must be confessed that many observers have believed that tbey 

 have traced nerve-fibres from the optic nerve into the great rods or their spindles. Dr. 

 Grenadier has given one figure, and only one figure, in which he shows this relation. 

 It represents three of the segments of the dioptron of a Crane-fly t ; and in one of the 

 three he has shown a nerve passing through the cuticular membrana basilaris, and 

 ending in one of the cells of the spindle. I have occasionally observed an appearance 

 similar to that represented in this figure, and my observations have led me to the con- 

 clusion that the fibres represented are not nervous, but form a portion of the connective 

 framework. In Tipula the bacilla beneath each segment of the dioptron are enclosed in 

 a distinct sheath (fig. 57), which often appears to be continuous with the sheath of the 

 great rod. This gives rise to an apparent continuity between the neural and dioptric 

 structures in many sections, and may have led Dr. Grenadier to believe that he had 

 actually traced the nerve into the spindle. 



There is no other figure in Dr. Grenadier's work which shows any actual passage of a 

 nerve-fibre into the great rods ; but two figures represent nerve-fibres passing up to, but 

 not through, the basilar membrane +. These are both from the eyes of Crustaceans, in 

 many of which the bacilla are very short. Several figures, of the eyes of Crustacea, 

 actually show the bacilla, although the author has not referred to these in the text. In 

 all the other figures, either the neuron is unrepresented, or it is only shown in outline 

 without sufficient detail. 



The retinal layer in Arthropods exhibits two very distinct modifications, which I 

 propose to term the segregate and the continuous retina. 



The segregate retina is characterized by having the bacilla arranged in distinct reti- ) 

 nuke, one for each segment of the dioptron. Each retinula is connected with the 

 ganglion by a distinct fasciculus of nerve-fibres enclosed in a separate pigmented sheath 

 (fig. 57). This form of retina is very frequent in larval insects ; it is less often seen in 

 the perfect state. The Crane-flies and some beetles have a segregate retina in the perfect 

 condition : Telepliorus affords a good example. 



The continuous retina consists of a bacillary layer extending over the inner surface of 

 the basilar membrane, connected with the ganglion by a single large nerve-trunk, 

 the fibres of which exhibit a complete decussation (figs. 41, 51, rt). The bacilla always, 

 however, show a tendency to be collected in small bundles, one corresponding to each 

 segment of the dioptron, except in some of the Dragon-flies, wherein the central portion 

 of the bacillary layer, at least, does not exhibit any division into separate fasciculi. 



The continuous retina is characteristic of the true Flies (Brachycerous Diptera), of some 

 Hymenoptera, Lepidoptera, and Neuroptera. 



* E. Berger, " Untersuch. u. den Bau des Gehirns und der E.etina des Arthropoden," Arb. Zool. Inst. Wien ; torn. i. 

 (1878). t L. c. fig. 44. i L. c. figs. 109, 114. 



