COBBOLD, ON A NAKED-EYED MEDUSA. 3 
transmitted light, appears more opaque than the filament, m 
consequence of its greater thickness, and the abundance of 
those highly refracting fusiform particles already described. 
The limitmg membrane of an otolitic vesicle was discernible, 
but there were apparently no vibratory movements within the 
cavity. 
The ocelli, eight m number (2 x 4), are placed round the 
circular margin of the disc, at intervals, between every third 
tentacle—an arrangement somewhat peculiar. The unassisted 
eye failed to detect their presence ; a very slight enlargement, 
however, rendered them visible. Magnified 60 diameters, 
each ocellus was seen to consist of a transparent vesicle con- 
taining a round nucleus at the base, and in addition, five 
bright yellow, highly refracting globules (fig. 8). The latter, 
larger than the nucleus, varied in size respectively, the diffe- 
rence being uniform and gradational. Under a quarter-inch 
lens these variations in size were more obvious, the bulk of 
the central and superior globule bemg paramount. The wall 
of the sac was now seen to be double, the ocellus bemg sup- 
ported by a cellular thickening of the ining membrane of the 
circular gastro-vascular canal (fig. 10). When under exami- 
nation, the nucleus broke up, and many of the tissues, else- 
where, disintegrated, while the animal was still living in an 
enfeebled condition. 
The marginal vessel is about the width of the filamentary 
tentacle, and to the unassisted eye its walls appear transpa- 
rent and homogeneous. Two kinds of corpuscles, large and 
small, are contained within the canal; of these we shall speak 
more particularly when referring to the circulation. 
The shelf-like veil is directed inwards at a right angle to 
the axis of the disc, and, though broad and conspicuous, offers 
no structural indications. 
The sub-umbrella is placed rather higher than midway be- 
tween the marginal ring and the convex surface of the disc. 
The depth of the concavity lessened during contraction, but 
not uniformly so, it bemg observed that the upper part re- 
mained unaffected, to the extent of a third of its area, from 
the summit downwards, the circular limit of this rigid portion 
forming, as it were, a point d@appui for the development of 
contractile action throughout the remainder of the membrane. 
No muscular tissue, properly so called, could be detected. 
The proboscidiform peduncle has all the features common 
to the genus. It is about the twentieth part of an inch in 
length, quadrangular, and provided with four simple or 
slightly fimbriated triangular lips (fig. 5). The contained 
stomachal cavity was thrown into various shapes during the 
