180 
Fam. Corallimorphidae. 
Though I fully agree wiih Stephenson (1921) that the Proto- 
stichodactylinae (Corallimorphidae, Ricordeidae (Watzl 1922) and 
Discosomidae) are more related to the Madreporaria than all other 
Actiniaria, in as much as a number of Madreporarian characters 
are accumulated in these families, I cannot, however, accept Ste- 
phenson’s proposition to remove the Protostichodactylinae from 
the Actiniaria — an opinion already pronounced by Krempff 
(C. R. Acad. Sc. Paris. 139. 1904) though partly based on other 
arguments than those of Stephenson. Our knowledge of the 
anatomy of the Madreporaria is namely so imperfect that it is im- 
possible, at least at present, to say where the Protostichodactylinae 
should be placed in the system of Madreporaria, and I think that 
if we remove the Protostichodactylinae we must also transfer the 
Gonactiniidae and the Ptychodactiidae to the Madreporaria, as these 
families are likewise related to this group. For the explanation of 
the question, whether the Protostichodactylinae belong to Actiniaria 
or to Madreporaria, it seems to me important to make it clear 
whether the Protostichodactylinae have lost their skeleton or never 
developed one. If the skeleton of this group is reduced we must 
place them among the Madreporaria; in case that the Proto¬ 
stichodactylinae never had a skeleton I see no reason to remove 
them from their old place. Though we shall probably never be able 
to confirm with certainty which alternation agrees with the phylo- 
genese of the Protostichodactylinae, I cannot find anything advocating 
that a skeleton has been reduced here. In case that all Proto¬ 
stichodactylinae were real deep-sea forms I should be inclined to 
suppose that a reduction of a skeleton had taken place here, as in 
Leptopenus among the Madreporaria, but as most Protostichodactylinae 
(except Corallimorphus and the nearly allied Isocorallion) are strongly 
marked littoral forms, I cannot find the reason for a reduction of 
a skeleton. Thus I favour the view that the Protostichodactylinae 
never developed a skeleton and consider them as forms descending, 
as the Ptychodactiidae and the Gonactiniidae, from a common an- 
cestor with the Madreporaria and having passed a development 
parallel with this group. A supposition of a parallelism here offers 
no difficulties as among the Actiniaria parallel series often appear 
(Carlgren. Wiss. Ergebn. Schwed. Siidpolar-Expedition, Bd. 6, L. 5. 
Stockholm 1911. p. 26). 
