STUDIES IN THE RETINA. 31 



spherical refractive body) which are scattered along the thin 

 cone. In fig. 3, a,^ very large one appears. These vacuoles 

 are all suggestive of fluid travelling outward along these 

 cones, and helping to swell the terminal vesicle, or, again, 

 they may be the remains of fluid left in the element after its 

 collapse under the action of the reagent, if such a collapse 

 takes place. (On this see below, p. 42, and fig. 12, a, the 

 second element on the left.) 



On turning now to the rods, to ascertain whether they 

 afford any evidence of their early developmental stages 

 having been cones, our attention is at once fixed upon those 

 with very long inner limbs, which we will call, after their 

 discoverer, Schwalbe's rods. On measuring these I found 

 that they were roughly divisible into two kinds, the one with 

 the inner limbs measuring about 30 /i, and the outer limbs 

 20 ju ; the other with inner limbs measuring 20 fx, and outer 

 limbs 80 ju;^ while, lastly, the inner limb of the normal rod 

 measures 10 ^, and the outer limb 40 //. These are shown on 

 the scale, fig. 4, o\, r^, r^. It was obvious that there must be 

 some kind of connection between these different kinds of 

 rods and the developing cones, owing to this general agree- 

 ment in the measurements, but it was some time before I 

 could see any satisfactory explanation of the phenomenon. 

 On arranging the rods side by side with the cones, in the 

 way shown in the diagram (fig. 4), not only does the scale 

 formed by the sections of the latter serve also to measure the 

 different kinds of rods, but also throws light on the develop- 

 mental process. Three divisions of the cone measure the 

 longest inner limb, two divisions the next, while the inner 

 limbs of the cone Co and fully formed rod are already known 

 to agree in length. Turning to the outer limbs, we find the 

 shortest agrees in length with the long terminal vesicle dis- 

 covered at the tip of the most developed cone, viz. 20 /t. The 

 next measures 30 /i, and that of the fully formed rod 40 ^. 

 We may also note that the shortest outer limb is very fre- 



' Hoffman (I. c.) gives the measurements of the outer limbs as from 

 20 to 25 ^, wliicli agrees closely enougii with my own. 



