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



The point of greatest importance in determining the position of these medusae 

 is of course, in the absence of a definite knowledge of the life history, the origin of the 

 otolithic cells, which, according to Lankester ('80) and Giinther ('94, '94a), are produced 

 from the endoderm of the circular canal. Lankester made his observations on the 

 marginal vesicles exclusively by means of optical sections, and it is no discredit to 

 the author to say that none of his figures touching the point in question can be con- 

 sidered as decisive. Giinther ('94), on the other hand, carried on his observations on 

 actual sections prepared from materials killed with osmic acid, and by means of the 

 best optical appliances, and he reproduces several figures intended to show the endo- 

 dermal origin of the otolithic cells. It must, however, be remarked that the 

 figure (Fig. 1) he gives us of the earliest stage that came under his observation in 

 the development of the otocyst represents a somewhat advanced stage, when its 

 constituent parts have all been formed, and these parts have only to enlarge, multiply, 

 and undergo a little differentiation to arrive at the definitive condition; and it is 

 exceedingly improbable that any additional endoderm cell should at this stage wander 

 out to reinforce their precursors, even if these were derived from that source (com- 

 pare Figures 17 and 18 in the present paper). The absence of the intervening meso- 

 gloea must not be taken by itself as a proof of the morphological continuity of the two 

 cell layers, and it is necessary to exercise constantly self-criticism not to be misled 

 into taking appearance for reality. The question of morphological continuity or dis- 

 continuity in such a case comes very near to splitting a hair, and it is only by the closest 

 scrutiny that one can draw a safe conclusion. 



It may be added in passing that the structure of the otolith which Lankester 

 and Giinther regard as unique represents only a developmental stage in other forms, 

 and it is possible that these authors have observed only comparatively young otocysts, 

 and that in older ones the otolith loses its cellular structure. Should the ectodermal 

 origin of the otolithic cells be granted, the structure of the marginal vesicles of Lim- . 

 nocodium and Limnocnida is exactly like that of the same organs of the OlindiadaB, 

 except for the peculiar prolongations into the velum of the vesicles in the first-named 

 genus. 



The presence of tentacle roots, though of secondary importance, is another fea- 

 ture common to the Olindiadse and the two fresh-water genera. 



The young stages of Limnocodium described by Lankester ('80, '81) are sup- 

 posed by him to have developed from the egg-cell, although he did not observe any 

 mature females, but only males. On examining them, however, one is struck with 

 some features in their anatomy that point strongly to their origin as medusa buds 

 either from a hydroid stock or, more probably in this case, from young female medusae. 



