On the Structiue iind Affinities of Muestra purasites Krolin etc. 55 



consideratiou has convinced me that tliis view is erroueous, that the 

 tvvo types of tentacle have had totally independent orig-ins and that 

 tbey are characteristic of the two subfamilies of Cladonemidae ve- 

 spectively. 



The b rane hing- type of tentaci e occurs in its simplest condition 

 in Eleiitheria, the development of the more complex forms has been 

 studied in the ])olyp [Stauridi-iirii] and in the young medusae of Clado- 

 nenia, and in Dendronenta. In all, the axis of the tentacle and of 

 its branches is either "a continuous tube" lined with endoderm, or 

 endoderm cells, and in the development of the Cladonema tentacle, 

 the branches appear to arise on the axial side. 



The semi p innate type of tentacle is characterised by the pos- 

 session of cnidophors which in all known forms consist solely of 

 ectoderm and are developed upon the abaxial side of the tentacles. 



This is certainly the case in Mnestra, Gemmaria, Zanclea and 

 Pteronema and will very probably also be found to be the case in 

 Cteuaria, when that remarkable form is reinvestigated. The various 

 stages through which the cnidophor-bearing tentacle may have 

 passed in the coiirse of evolution are still to be found among the An- 

 thomedusae. We assume that in the primitive type of tentacle, the 

 cnidoblasts were fairly uniformly distributed over the entire surface. 

 The cnidoblasts then seem to have become segregated into batteries 

 as in Sarsia. The cnidoblast batteries then became grouped along 

 the abaxial side of the tentacle [Ectopleura minerva Mayer), then 

 elevated upon contractile peduneles [Cfemiria [?]), and finally they 

 became enclosed in cellular envelopes as in Gemmaria, Muestra etc. 

 It is interesting to note that Gemellaria implexa, the hydroid form 

 of Gemmaria, has capitulate tentacles. 



Nematocyst streaks (cnidacts) and nematoeyst tubes. 



There are two types of structure observal)le among the Clado- 

 nemidae which have never been proved to be homologous and which 

 should therefore not be confused with oue another. In one type 

 the cnidoblasts of the exumbrella are arranged along radial streaks 

 (cnidacts) and lying on the outer surface of the umbrella are in- 

 stantly available for purposes of offence. Cnidacts occur in severa! 

 genera of medusae (e. g. Ectopleura with 8) other than the Clado- 

 nemid genera Zauclea, Muestra and Ctenaria. 



In Gemmaria and Cte/taria there is a seeoud type of nemato- 



