THE CRASPEDOTE MEDUSA OLINDIAS AND SOME OF ITS NATURAL ALLIES. 17 



the outer nerve-ring; but this coincidence with a character of the Trachylinae must 

 be due to convergence, and the difference of innervation must not be regarded as 

 absolute for the two types of otocysts. In Olindioides, moreover, the two nerve- 

 rings are seen in several places to be directly continued one with the other through 

 breaks in the mesoglcea separating the two. 



The presence of centripetal canals is another point of resemblance between Olin- 

 diadse and some Trachomedusse ; but it is evidently of no systematic moment, since 

 these canals may be present or absent in closely related genera. 



I have spoken of the marginal nettle-ring; but it must be borne in mind that it 

 is unlike the nettle-ring of the genuine Trachomedusse. For, while in the latter the 

 ring is a very distinct structure continuous throughout, in the Olindiadae it is brought 

 about merely by the close proximity of the velar tentacles, and in such a form as 

 Gonionema, where the intervening spaces are wider, the ring tends to become more 

 or less discontinuous, and in some sections through such a space the nettle-cells are 

 totally absent. We must also remember that in some Leptolinae there is a well-devel- 

 oped nettle-ring, as in Laodice (Brooks, '95). 



The presence of endodermal tentacle roots in the Olindiadse is another feature 

 of resemblance to the Trachomedusse, but we must also remember that these struc- 

 tures may be present in forms which undoubtedly belong to the Leptolinse (Brooks, 

 '95). 



The unsatisfactoriness of Haeckel's system of the medusse has been pointed out 

 by Brooks ('86, '95), but accepting it we must place the subfamily Olindiadse in the Lep- 

 todmedusse. It is impossible to put it under the Cannotidse, to which Gonionema is 

 referred by Haeckel. It appears, to me that it comes with least violence under the 

 Eucopidse, where our medusse may find their temporary resting-place. 



The genus Gonionemoides formed by Mayer ( : 00) is evidently nearly related to 

 our genera, but less so than these are among themselves, and it will perhaps be found 

 advisable to place it in a distinct subfamily. 



The fresh-water medusse Lhnnocodium and Limnocnida have not found a satis- 

 factory place in the system. Allman ('80) and Fowler ('90) think that Limnocodium 

 should be placed in the Leptolinae, while Lankester ('80) and Gunther ('94) bring 

 it under the Trachylinse, in which the latter also places his genus Limnocnida from 

 Lake Tanganyica. It appears to me that these genera present some features which 

 point to their affinity to the Olindiadae, near which they can, in my opinion, be placed 

 most naturally. But as I am not able to examine any material of these medusae, 

 I shall confine myself to some critical remarks on the statements of previous authors 

 concerning points of fundamental importance. 



