LIMNOCODIUM (CRASPEDACUSTES) SOWERBII. 367 
seat of origin ; but the cortical cells rest upon the tissue of the 
transparent marginal ring and have some kind of attachment 
to that ring, though I am not able to point out any definite 
nerve-fibres. That the attachment is very slight is shown by 
the fact that sometimes a marginal body (refringent body) be- 
comes broken off from its seat of origin, and falls into the 
tubular capsule. I have met with such a detached refringent 
body enclosed in a constriction of the peripheral extremity of a 
capsular tube (PI. XXXI, fig. 20) . This detached refringent 
body was remarkable for the fact that it had undergone an ab- 
normal development, and was actually a c^st with an internal 
cavity, into which projected the large refringent cells, whilst it 
was clothed externally by the cortical cells. 
That the refringent bodies in themselves, irrespective of the 
tubular capsules, correspond morphologically to solid tentacles 
is proved by an occasional abnormality. In place of a normal 
refringent body I have found rarely a tentacle-like body one of 
which is drawn in PI. XXXI, fig. 18. This small tentacle was 
in exactly the same position as a refringent body. It is seen to 
have in its axis one large and apparently two smaller granular 
cells similar to those which are protruded from the endoderm of 
the ring- canal to form the medullary cells of the normal re- 
fringent body. ‘At the same time the ectoderm which clothes 
these axial cells is free and superficial. Instead of forming a 
capsule^ the ectoderm cells have simply invested the axial growth 
as they do in the case of a tentacle. 
Further, I have observed another abnormality of an exactly 
complementary character, namely, the commencement of the 
tubular ectodermal capsule without any refringent body to oc- 
cupy its proximal extremity (PL XXXI, fig. 19). The exist- 
ence of these two elements separately one from the other, jus- 
tifies us in regarding them as two distinct structures which have 
united to form the complete ‘^refringent bulb and sac of our 
new Medusa. 
Have we anything parallel in Leptomedusae or in Tracho- 
medusae for the two distinct elements, the modified tentacle and 
the velar ectodermal capsule, above described ? Assuredly there 
is no parallel at present described among Leptomedusae. 
Among Trachomedusae, however, we find a very exact parallel. 
Although in no other Trachomedusan does the axial ])ortion of 
the modified tentacle retain so simply the character of the car- 
tilaginoid axial tissue of a tentacle, and although in other Tra- 
chomedusae an otolithic concretion is formed by the axial cells, 
whilst an otolith is most decidedly not formed in Limnocodium, 
yet in the essential character of the refringent body, and in the 
remarkable ectodermal capsule, Limnocodium agrees most pre- 
