STUDIES IN THE RETINA. 37 



ments of the parts of the elements give us an entirely differ- 

 ent kind of transition form from that which the earlier 

 workers looked for. But, as we shall see below, in young eyes 

 transition forms more like the kind expected do occur in 

 great numbers (fig. 12, c). 



I have not succeeded in ascertaining whether or not the 

 tip of the cone shown in the earliest stage (fig. 8, a) is the 

 true tip. The aspect of b certainly suggests that the tips 

 break off, and when we find the cone not only in the Anura, 

 but also in the salamander (see below), running straight 

 down to the pigment layer, it is fairly certain that the same 

 is the case here. The cone " grows " by the enormous deve- 

 lopment of the basal vacuole, which forces out the ellipsoid 

 (which was here densely stained) with the conical mass 

 distal to it; c shows a double cone; d, e, f show different 

 lengths of the vacuole and enlargements of the tip ; from / 

 the outer limb of the rod grows and forces back the ellipsoid, 

 which changes its shape. Some of these developing outer 

 limbs (from g) seemed to be thrust much deeper than others 

 into the pigment cells, at times almost displacing the latter 

 — an appearance very suggestive of their active growth in 

 length. 



The narrowing of the vacuole from c to d, e,f is not in- 

 variable. It is probably purely accidental that those with 

 compressed vacuoles were most numerous in the sections. In 

 such a confused mass of elements forcing their way out, and 

 mutually squeezing one another, naturally many different forms 

 will be assumed by individuals. Further, these long narrow 

 vacuoles would be likely to be less injured in section-cutting 

 than those most swollen, and would occur more frequently in 

 the preparations. The inner limbs of rods g and h are also 

 probably chance forms. 



It will now be understood from the series why I do not 

 think that there is much value in the conical shape of the 

 rod. In the swelling of the outer limbs seen in e and/, to 

 the size shown in I, it seems hardly likely that the conical 

 shape would be preserved, especially if the outer limbs of the 



