AND THE MOKPHOLOGY OP THE EYE IN INSECTS. 413 



nients in the embryo some time before the hatching of the egg ; but before that event 

 they are withdrawn into the interior of the embryo by a process of invagination during 1 

 the formation of the fore gut, which lies entirely in front of these structures. 



The fore gut forms no part of the alimentary canal of the imago, but is shed with the 

 larval integument in the first stage of the formation of the pupa. This then puts on the 

 same appearance as the embryo had whilst it still lay inclosed in the egg, by the un- 

 folding of the imaginal disks ; and these, again, form the anterior part of the body. 



After the unfolding of the imaginal disks, the development of the compound eye 

 proceeds in the same manner as in those insects in which it is developed from the cells 

 of the hypoderm. 



I believe that the pedicle, which supports the disk from which the dioptron is deve- 

 loped, is merely a portion of the connective-tissue capsule that incloses the ganglion 

 before the evolution of the disks in the formation of the pupa. The optic nerve and 

 retina are formed at a later period. 



These facts throw light on the relationship between the ordinary compound eye and 

 the remarkable encapsulated eye of the entomostracous Crustacean, Leptodora hyalina, 

 described by Dr. Weismann f . This is deeply seated in the median line of the transparent 

 head, in immediate relation with the cephalic ganglion, and it has no apparent connexion 

 with the skin or surface of the animal. 



In describing the variations of the retina I have already adverted to their develop- 

 mental significance. I have especially studied the developmental relation of the segre- 

 gate and continuous forms of retina in the Dragon-flies, and there are numerous indi- 

 cations that what I have actually observed in these insects is a common phenomenon 

 amongst the Arthropods. 1\\ the earlier stages of the Dragon-fly larva the compound 

 eye is already functional, at least as far as its central facets are concerned. In this con- 

 dition there are very few facets. Those of the more central portion of the eye have a 

 distinct retinula to each, and each retinula has a distinct nerve-bundle (figs. 43 & 45). 

 As development progresses, new segments are added to the periphery of the dioptron, 

 and the number of retinula? increases. 



In some sections the continuous retina of the perfect insect (rt ) is already seen upon the 

 surface of the optic lobe (fig. 44), and the fibres of the nerves from the partial retina? (rt*) 

 can be seen passing between its rudimentary elements. The existence of a decussating 

 nerve beneath the continuous retina is very apparent. The retinal ganglion (fig. 44, g) 

 is now deeply seated in the interior of the optic lobe. 



Beneath the ganglion, both in this and in later stages, there is a remarkable ganglionic 

 centre (marked k in my figures). It is in relation with the retinal ganglion by a 

 second bundle of decussating fibres. It is, I believe, the representative of those 

 remarkable kidney-shaped bodies described by Mr. Newton in his admirable description 

 of the eye of the Lobster % . 



As development progresses, the continuous retina gradually enlarges, and approaches 



t Zeitschr. f. w. Zool. Bd. xxiv. 



t E. F. Newton " On the Eye of the Lobster," Quarterly Journal of Mie. Science, 1873. 

 SECOND SERIES. — ZOOLOGY, VOL. II. 61 



