119 



As is said above tlie shape of the pneumatophore varies exceedingly. The appendao-es 

 are numerous. Next to the main tentacle (length of the longest 20 cm.) there are several smaller 

 ones. The main tentacle was always wanting in the distal part of the specimen. Reproductive 

 appendages were present in all stages of development. 



Of the exact position of all the other appendages nothing could be said, as contraction 

 of the float makes it difficult to consider the position the appendages have taken, to be the 

 definite normal one. It seems to us that the position of appendages and their various wavs of 

 arrangement are not at all to be traced in fully-developed specimens which are preserved in 

 spirit. The best way to know more about it seems to us to follow the development of the 

 appendages from the earliest larval stages up to the mature ones. 



Huxley 59 found two larval stages, Haeckel 88 b describes also some of them and so 

 does Chun 97 a. The latter informs us that he will publish a monograph o{ Physalidae and we 

 are looking forward to an exact description of the development of the appendages in Phvsalia 

 which will certainly fill a great void. 



Some authors suppose there is no definite arrangement of appendages in Physalidae, 

 but it seems to us they cannot make an exception in the general structure of Siphonophora, 

 which always shows some kind of definite arrangement. 



We abstain from giving a detailed enumeration of the Siboga material as we should 

 have to describe one specimen after the other, and this would create too much confusion : 

 preserved Physalia never being normal. 



Among the 20 specimens of Physalia ntriculus of Cat. 20 there were some which showed 

 many beautiful gonodendra, attaining considerable size. We thought it interesting to detach one 

 of them and tried to separate the different groups, so as to get a definite idea of its structure. 



Chun 97a gives a preliminary description of the structure of gonodendra in Rhizophy- 

 saliae of which we propose to give a summary as it seems to us to be useful when comparino- 

 our material with his observations. 



In Rhizophysa (Chun 97 a) a mature gonostyle shows one main branch which o-ives off 

 many side-branches. Each of these side-branches consists of one sexual palpon ("Genitaltaster") 

 one medu.soid gonocalyx and 8 — 9 gonophores. A gonodendron of Physalia difters from the 

 same in Rhizophysa through the presence of polyp-like appendages, which formerly Chun thouo-ht 

 represented the proximal part of the medusa as they resembled exactly the basal part of the 

 medusiform gonocalyx and these might have become detached. 



However they have nothing to do with these structures as they also appeared in youno- 

 stages of gonodendra. Chun calls them "Gallertpolypoide". They are situated at the proximal 

 part of the furthest side-branches. They consist of tubes of which the proximal third part is bent 

 and lies against ("angeschmiegt") the main stem and is attached to the latter by a muscular 

 lamella. A distal opening was never to be found according to Chun. 



Chun describes further on (p. 69) the development of a gonodendron in Rhizophysa, which 

 we will not repeat so far as the description of the earlier stages is concerned. As soon as the 

 " maulbeerformige Gestalt" has been reached each of the semispherical buds constitutes one 

 side-branch of the gonostyle. All these side-branches keep the same degree of development • not 



