412 ME, B. THOMPSON LOWNE ON THE COMPOUND VISION 



nerve-terminals are believed to be developed, are those of an undifferentiated epiblast; 

 wbilst in the Arthropod, as in the vertebrate, the great neural tract has already beeri 

 differentiated before the groundwork of the compound eye is laid in the cells of the 

 hypoderin Metschnikoff*. The hypodermic cells in the region of the dioptron become 

 greatly elongated, and undergo both longitudinal and transverse subdivision; but as 

 the tracheal vessels and stellate connective-tissue cells permeate the hypoderin at a verj 

 early stage in the development of the dioptron, it is extremely difficult to determine how 

 much of this structure is derived from the hypodcrm and how much originates ii 

 mesoblastic elements. The earliest stages of development are far more difficult to follow 

 in the Diptera. This arises from two principal causes : first, the " imaginal disks " froir 

 which the integumental structures of the imago are developeel are closely united to tht 

 nervous system ; and secondly, the whole development is very rapid, anel the parts ar< 

 exceedingly soft and easily destroyed. 



The compound eye of the larva of Corethra, in the earliest stage in which I have yei 

 observed it, closely resembles the same structure (as figured by Metschnikoff t) from Aphii 

 rosce. It consists of a discoid group of pyriform cells, with their long axes at righl 

 angles to the surface of the disk. The rounded outer ends of these cells he immediatelj 

 under the transparent non-facetted cuticle, which represents the cornea, Their innei 

 extremities are prolonged, and are intimately connected with the deeply pigmented 

 sheath of the optic nerve. 



Dr. Weismann J has figured the disk from a still younger larva, and shows the nerve 

 ending in a layer of cuboid cells, which, from observations I have made on the very 

 rudimentary eye of a Lepidopterous larva, form, I believe, a distinct layer beneath the 

 disk from which the refractive structures are developed. My own observations have 

 led me to conclude that at this stage the nerve-fibres end in a bed of granular nucleated 

 protoplasm, not unlike a large motorial end-plate (fig. 73). Each of the pyriform cells 

 of the disk has a bundle of rod-like bodies of an orange-brown colour near its outer end. 

 These eventually develop the cone and spindle. 



The rods are larger in the cells of the centre of the disk than in those of its periphery ; 

 and the whole cell shows symptoms of division into four longitudinal parts. 



The close relationship of this disk to the cephalic ganglion, both in Aphis and Corethra, 

 throws light upon the still more remarkable developmental history of the eye in the true 

 Hies, so well described by Weismann. 



In these the disk is attached to the optic ganglion by a narrow pedicle, and has no 

 connexion with the other integumentary structures in the pupa. It is one of the 

 remarkable cellular expansions which are known as " imaginal disks." The head and 

 thorax of the imago are formed from a series of these disks, which are really involutions 

 of the epiblast. These are drawn into the interior of the embryo whilst it is yet inclosed 

 in the egg. 



The procephalic lobes and all the appendages of the head are distinctly seen as rueli- 



* Zeitschr. f. w. Zool. 1871 ; Monthly Mic. Journ. 1872. t Zeitschr. f. w. Zool. Bd. xvi. 



t Zeitschr. f. w. Zool. Bd. xiv. 



