8 THE CRASPEDOTE MEDUSA OLINDIAS AND SOME OF ITS NATURAL ALLIES. 



bulbular velar tentacles. All suspicion that the bulbular and the filamentous forms 

 may be fundamentally different has been dispelled by the presence of numerous 

 intermediate stages. The bulbular velar tentacles as well as the ectodermal thick- 

 enings that connect them are then the nisi formativi of the stinging cells, whence they 

 travel in opposite directions, up along the peronia into the exumbrellar tentacles, 

 or down into the filiform velar tentacles. The wandering nettle-cells are much less 

 nmnerous in the exumbrellar than in the velar tentacles. 



The marginal organs are all otocysts, and are present in large numbers, two at 

 the base of the endodermal root of each exumbrellar tentacle. In younger exam- 

 ples they form a single row along the umbrella margin and are situated close to the 

 circular canal (PI. I, Fig. 3), but in larger ones they no longer keep the same front, 

 and, generally speaking, the older ones are pushed up into the jelly of the umbrella (Fig. 

 4). They are either spherical or ellipsoidal, and some of the larger ones measure 

 no less than 160 micra across. The vesicle is lined by a flattened epithelium, which 

 presents a local thickening at a place turned towards the circular canal, where prob- 

 ably the sensory cells are found. The otolith is surrounded by some granular proto- 

 plasm containing nuclei, and is attached to the wall of the vesicle by means of a short 

 stalk consisting of some granular, well-staining protoplasm containing nuclei. The 

 otolith itself is highly refringent, and, when fully formed, consists of somewhat irregu- 

 larly concentric layers arranged around a centre. It appears to be very hard and is 

 apt to be detached from sections. 



Haeckel says that the structure of the marginal vesicles of Olindias shows it to 

 belong to the Trachomedusse, but as both Olindias and Olindioides as well as the 

 closely related genera before mentioned appeared to me to present several features 

 which are not commonly met with among the Trachomedusse, I have paid special 

 attention to the development of the otocyst, which, in the absence of any knowledge 

 of the life history, must be considered the most reliable criterion by which to decide 

 the natural relationship of the veiled medusae. 



I have shown in Figvires 15 to 19 (Pis. II, III) some stages in the development 

 of the otocyst. The endodermal roots of the eximibrellar tentacles start from the 

 circular canal at the place where the two kinds of endodermal cells already mentioned 

 meet each other, and the first rudiment of the otocyst is seen in process of formation 

 close to this root. It consists of a small segregation of ectodermal cells hardly dis- 

 tinguishable from the rest, closely applied to the endoderm of the circular canal at 

 the point where the two kinds of cells meet (PI. II, Fig. 15). At this stage it would 

 be very apt to be overlooked, were it not for the presence of the root of a very young 

 exumbrellar tentacle. There cannot be any reasonable doubt that the rudiment 



