30 H. M. BERNARD. 



layer of cylindrical bags was formed, new ones endeavouring 

 to force their way out between them would have to be wedge- 

 shaped or conical in order to succeed. This would especially 

 be the case with any sac-like body forcing its way out between 

 a compact mass of amphibian rods; its shape would necessarily 

 be conical, the basal portion, between the inner limbs of 

 the existing rods, which (inner limbs) are certainly softer 

 than the outer, could here be swollen or bag-like ; while 

 between the outer limbs of the rods, which there is every 

 reason to believe are, at least compared with the inner limbs, 

 turgid and firm, the new protrusion would have to be tapering 

 and thread-like, and only able to swell again as a vesicle 

 when it projected beyond the tips of the rods — that is, in the 

 pigment layer. 



If this line of reasoning is correct, it leads us to conclude 

 that the cones, at least in the eyes of the Amphibia, are not 

 the important fixed morphological elements which all the 

 earlier investigators declared them to be, but merely stages in 

 the formation of new rods. Daring as the conclusion may at 

 first sight appear, there is nothing very improbable in it. If 

 new rods are required, as they must be as the retina grows, 

 they would certainly be formed as they are in the earlier 

 development of the eye, viz. as protrusions from the retina. 

 Such protrusions, forced down between the existing rods, 

 would, as far as we can see, necessarily be conical. We 

 know of no other conical protrusions but those usually called 

 cones, and hitherto regarded as fixed elements with special 

 functions. Hence these cones must themselves be the stages 

 in the development of fresh rods which the argument requires. 

 Turning to our figures and diagrams, we find this line of rea- 

 soning fully confirmed, for qualitative and quantitative com- 

 parison of the different forms of the cones themselves shows 

 clearly enough that they are different stages in the protrusion 

 of matter beyond the limiting membrane. In this connection, 

 too, we can best call attention not only to the large vacuole 

 in the basal section of Cg, a vacuole sometimes enormously 

 distended, but also to the small vacuoles (omitting the 



