^°1893^''] PROCEEDINGS OF THE NATIONAL MUSEUM. 129 



codermata of Milue Edwaitls. He <livides the jirinii) into seven fami- 

 lies, thus* 



EdwardsiiKP. Zoaiithintp. 



Actiniiiic. CiTiaiithina". 



Sticbod.'ictj'liiue. Miuyailiiuf. 

 ThalassianthiiiiP. 



The names of tlie majority of these groups indicate their limitations; 

 the greatest innovations are the separation of the Edwardsias from the 

 Actininte and the establishment of the Stichodactylina?. This family 

 possesses for its distinguishing character the feature upon which Klun- 

 zinger based his family Discosomidse, i. e., the radiate arrangement 

 of the tentacles, but at tli(^ same time it is made much more comprehen- 

 sive, the Pliyllactina' of Klunzinger being associated with Discosoina, 

 Capnea, Aureliana. Phymantlms and other genera, all of which possess 

 radially arranged tentacles. The Thalassianthinae is, consequently, 

 poor in genera compared witli Klunzinger's Thalassianthidsv, contain- 

 ing only a few forms with large compound tentacles. Four of Andres' 

 groups are certainly well established, namely, the Edwardsinte, Ac- 

 tininai, Zoauthinje and Cerianthime. He was influenced, hoAvever, too 

 much by the arrangement and structure of the tentacles in making the 

 Stichodactylina' and Thalassianthina- etjuivalent to these four; they 

 should more i)roperly be made subgroups of the Actinim^. The same 

 remark api)lies, perhaps, to the Minyadinte, though we are still in ignor- 

 ance as to the structural peculiarities of its m«^ml)ers. The fact that some 

 of the species evidently have their parts arranged on a hexamerous plan 

 favors this view, and the occurrence of others ijossessinga decamerous 

 arrangement can not Ije considered as of great weight in favor of keep- 

 ing them distinct, in view of the same symmetry occurring in the Hal- 

 campida', for instance, and in other sporadic instances in which there 

 can be no question as to the advisability of associating the forms with 

 their hexamerous relatives in the Actinina^. In fact, it seems proba- 

 ble that the Minyadina^- are not even to be given a value equal to the 

 Stichodactylina', but are rather to be referred to the Halcampidte, a 

 family <n' subfamily of the Actinina", 



The disappearance of the Ilyanthidje from the list of families is an 

 important point also. Andres has diminished the importance of trivial 

 characters in accomi)liwhing this, and has emphasized tlie importance 

 of the numerical relations of the parts as a basis for classification 

 in sei)arating from them the Edwardsias and referring the thus 

 restricted groups to the Actininje. 



Andres enters into fi mucli more minute division of his families into 

 subfamilies, many of which are well founded, but it will not be con- 

 venient to criticise them here. 



With Andres the second period in the history of the classification 

 of the Anthozoa may be said to close. The period was marked by a 

 gradually growing tendency to divide the group into a number of 

 Proc, 1^, M. 93= 9 



