20 ACTINIARIA 



mesenteries (the checked development of the dorsolateral mesenteries of the second cycle) 2) the presence 

 of a single, well-developed siphonoglj^Dhe and 3) the arrangement of the tentacles : the inner endocoel- tentacles 

 are shorter than the outer exocoel-tentacles (compare Carlgren 1904 p. 544). There is no doubt that the 

 above-named genera are nearly related to each other, but it is a question, if these genera alone ought to be 

 placed in a particular family. Leaving Oractis out of consideration, the position of which I will further discuss 

 later on, we find a bilateral symmetry, though of a somewhat different type in the Edwardsids, in Pentactinia 

 — where the number of mesenteries is the same as in Peachia, but the ventrolateral mesenteries of the second 

 order not developed, while in Peachia the corresponding, dorsolateral mesenteries are missing, in Parahal- 

 campa and Limnactinia, and at last probably also in Siphonactinopsis, the mesenteries of which are twice 

 as many as in Peachia. A single, ventral siphonoglyphe is also present on most, perhaps all Edwardsids, 

 though it is only a little differentiated, furthermore in Pentactinia, Harenactis, Mesacmaea, Scytophorus and 

 Parahalcampa. The same arrangement of the tentacles as in Peachia we observe in several Edwardsids, 

 namely in the subfamily Edwardsiinae, while in theMilne-edwardsiinae and the other genera the inner tentacles 

 are longer than the outer ones, or all tentacles of about the same length. Only the arrangement of the mesen- 

 teries is thus specifically characteristic of Peachia, Haloclava and Eloactis, in as much as there are 10 pairs 

 of mesenteries, while the dorsolateral mesenteries of the second order are missing, an arrangement which 

 possibly they have in common with Siphonactinopsis, though there are twice as many in the latter. However 

 much the arrangement of the mesenteries varies in the Athenaria I will call the attention to the fact that 

 Scytophorus has onlj^ 14 mesenteries, and therefore I do not think it justifiable to estabHsh a separate family 

 for Peachia, Eloactis and Haloclava on account of the number and position of the mesenteries. The maintain- 

 ing of the family Monaulidae and the establishing of several families owing to the arrangement of the mesen- 

 teries are the logical consequences of these facts. 



If we keep to the arrangement of the tentacles, it would be much more justifiable to place the 

 subfamily Edwardsiinae together with Peachia, Haloclava and Eloactis in a family Edwardsiidae, and the 

 subfamily Milne-edwardsiinae together with the other acontia- and sphincter-lacking Athenaria in another 

 one. As it is, however, possible that this conformity in the arrangement of the tentacles in the Edwardsiinae 

 Peachia, etc. may depend on a convergence ^, caused by their having parasitic larvae (observed among the 

 genus Edwardsia and Peachia), I provisionally place the subfamilies Edwardsiinae and Milne-edwardsiinae 

 as before in a single family, Edwardsiidae, while the other acontia- and sphincter-lacking Athenaria 

 are referred to a family which I call Halcanipoididae after Halcampoides, the most primitive genus. (Halcam- 

 poididae, proposed as a subfamily by Appell5f (i8g6), is synonymous with my family HalcampomoriDhidae 

 which must be dropped, according to the international rules). Thus I provisionally refer to the Athenaria 

 the same families as in 1900, only adding the new family Limnactiniidae, and with the difference that the 

 name Halcampomorphidae is exchanged for Halcanipoididae. Consequently the arrangement of the genera 

 in the different families is as follows: 



' I seize the opportunity to rectify an error, slipped in while my paper (1900 b) was being printed. Page 544, the first note has 

 "nicht wahrscheinlich" for, in my manuscript "recht wahrscheinlich." 



