32 Dr. Reid on the Development of the Medusae. 
of the body, is so thin that at first sight these processes appear 
to be attached to the external surface. Fig. 16 is a greatly en- 
larged view of one of these bifid processes. Each of these pro- 
cesses forms two hollow floating tubes, communicating with the 
stomach or internal cavity by a common orifice (fig. 16 a), and 
having the edges of their external surfaces covered with nume- 
rous flliferous capsules (fig. 16 5). The stomach is large and 
extends nearly to the margin of the body or disc. Outside the 
position of the four bifid processes, and on the lower surface of 
the inferior wall of the body, there is a circular band, slightly 
elevated, more granular and opake than the portion of the body 
placed within it, having prolongations passing off from its outer 
edge to the intervals between the eight bifid lobes or rays that 
spring from the margin of the body, and others along the centre 
of the lower surface of these bifid lobes, as far as the ocellus 
placed at the point of bifurcation of each lobe (fig. 15). When 
the animal contracts the marginal lobes in swimming, this circle 
becomes narrower, more distinctly defined, and approaches nearer 
to the mouth. In certain states of the animal the prolongations 
from the outer edge of this circle to the intervals between the 
eight bifid rays are longer than represented in fig. 15. When 
the animal is examined in certain positions and with glasses of 
weak power, this circle, and the sixteen prolongations extending 
outwards from it to the intervals between the rays, and along 
the lower surface of the rays themselves, assume pretty nearly 
the appearances represented by Steenstrup as vessels ; and as I 
have been unable to satisfy myself of the presence of any vessels 
there, I am inclined to believe that he has been misled in this way. 
I have occasionally observed the appearance of a thread-like 
nervous circle around the mouth, sending a filament along each 
of the rays towards the ocelli, on approaching which it bifurcated ; 
but not having been able to make these out at other times, 
under circumstances that appeared favourable for their detection, 
I am not prepared to affirm that a nervous system is present. 
At the point of bifurcation of each of the marginal lobes or 
rays there is placed, as Sars has described, a little eminence, hy- 
pothetically designated by Steenstrup an ocellus (figs.l5c &17 g). 
This ocellus forms a mammillary process, consisting of three 
distinct structures (fig. 17 a). The apex is chiefly formed of a 
considerable number of very minute crystals, and a small part of 
its base is more opake and more granular than its larger middle 
portion. From a greatly enlarged view of the crystals occupying 
the apex of the ocellus, given in fig. 18, it will be observed that 
the upper are shorter and thicker than the lower ; in fact, while 
a few of the former are almost as thick as they are long, some of 
the latter arc almost needle-shaped. On fixing the polarizing 
