Eyes of Molluscs and Arthropods. 633 



mcntioned parts, in various degrees of clearness, are to be found, and 

 by studying them. one may arrive at a tolerably complete coneeption 

 of the existing eonditions. In tig. 7ü, is represented a surface view of 

 the corueal facets with the hypodermis and inner euds of the onima- 

 tidia still attached. Those portions of the calycal sheath which, by 

 the fusion of corresponding parts from adjacent ommatidia, form the 

 membranous tubes at the corners of the Squares (fig. 76 rt.s.) — are 

 exactly similar to those seen in the deeper parts of the eye. Here we 

 have the evidence for the above made assertion that each tube is com- 

 posed of parts of the membranous ends of seven retinulae belonging to 

 four different ommatidia ; for one may see that, as the tube approa- 

 ehes the facets, its walls thicken, and, when it finally abuts against the 

 eorneal membrane, are resolved into seven hyaline thickeuings, so ar- 

 rauged as to form the regulär, four-armed figures shown in the drawing 

 (fig.76,/andtig.77]. Bycareful examination of manyspeeimens, someof 

 which have been treated with caustic potash, a series of forms may be 

 found, varying from those in which the tubulär membrane projects some 

 distance away from the cornea, to those having four arms extendingout- 

 wards as thin membranes (fig. 77 rt.s.), and still others, in which these 

 radiating membranes have become so extended as to be continuous with 

 corresponding ones from adjacent tubes. In those treated with caustic 

 potash, all the more delicate membranes are dissolved, their thickened 

 extremities alone remaining. One then sees that the centre of the tubes 

 is reduced to a small opening, while each of the radiating membranes 

 has resolved itself into two, doubly wedge-shaped thickenings, seven 

 or eight of which are arranged (fig. 77 IV) around a point which lies 

 directly beneath a place where the corners of four adjacent facets touch 

 each other. Of these eight radiating thickenings, two in each group al- 

 ways face the same Square, and — since they unite with those forming 

 the remaining sides of the same — the centrai retinophorae are com- 

 pletely and immediately surrounded by the outer ends of the retinulae , 

 1 — 7, thus justifying our conclusion that, morphologically, the latter are 

 the most intimately connected with the retinophorae. It has been said 

 that the ends of seven retinulae constitute one of the four-armed figures 

 described above, and, if this is so, it is evident that one of the pairs, and 

 in fact a pair separated by an open angle (as e. g. s' and s^^ //; fig. 77)^ 

 must be formed, not by two pieces, but by one only. Although it is not 

 difficult under favorable circumstances to observe the dividing spaces 

 between the segments, yet it is by no means easy to determine 

 satisfactorily, whether two of the pieces in each group are fused to 



