VOL. XVI. J PROCEEDINGS OF THE NATIONAL MUSEUM. 131 



1893. 



The idea brought forward in this work was ehiborateil more fully by 

 Richard Hcrtwig in liis report on the (Uiallenger Actiniaria ('82), in 

 which the structural peculiarities of the various forms are employed, 

 not only to distinguish the principal groups, but also to detine in an 

 accurate manner the various families of the llexactiniie. In some par 

 ticulars the Idea was carried a little too far, owing to the absence at 

 that time of anatomical studies of a large series of forms, Ilertwig be- 

 ing' obliged to rely entirely on his (»wn observations in deciding as to 

 the relative importance of a character. Omitting the Alcyouaria and 

 Antipatharia from consideration, Hertwig recognizes six tribes of Acti- 

 niaria, which correspond in taxononiic value to Andres's famdies. These 

 tribes are (1) Hexactinia^, (2) I'aractinia', (.J) Monaulea^, (4) Edwardsise, 

 (5) Zoauthea% and (0) Cerianthea', and all are characterized by the ar- 

 rangement of the mesenteries. Three ot these orders corresi)oiid to 

 families of Andres' classitication; other three of Andres' families, viz; 

 Thalassianthiua^, Stichodactylinie, and Minyadinie are grouped with 

 his Actinina; to form the tribe Ilexactinite, while two other tribes, not 

 represented in Andres' system, are instituted for forms presenting an 

 arrangement of the mesenteries not previously recognized. In com- 

 paring the systems of Hertwig and Andres, however, it nuist be re- 

 membered that the tAvo w(u-ks were so nearly contemi)oraiieous that 

 the respective systems were entirely independent one of the other. 

 Andres, it is true, had the advantage of the earlier work of the 

 brothers Ilertwig ('79), which no doTibt influenced considerably his 

 id(^as as to the relationships of certain of the groups, but had no cog- 

 nizance of Richard Ilertwig's later observations. 



The introduction into the classitication of the Anthozoa of a system 

 based upon anatomical peculiarities, instead of oiu'. resting entirely on 

 variable characters, readily subject to modification in accordance with 

 external conditions, was very important. There yet remained to be 

 taken the further step of adding to anatomical characters the informa- 

 tion derived from embryological investigation, a step the importance 

 of which the Hertwigs had recognized and contributed to, to a certain 

 extent. Some of the necessary intormation was contributed later by 

 Boveri ('90) and myself ('91), and as the result of these observations 

 1 drew up a classification of the Anthozoa founded upon structural and 

 embryological characteristics. p]itlu}r of these classes of facts, taken by 

 itself, is liable to lead to ei-rors; it is only by combining both that a 

 true knowledge of the ]»hylogenetic relationships of the various grou])S 

 can be obtained. For instance, relying entirely on embryological data, 

 the Hexactinia; could be se|)arated into three distinct groups, one in- 

 cluding those forms in which tlie mesenteries ap])ear according to the. 

 succession described by Lacaze-Duthiers; a second, in which the 

 mesentery succession is that described by the Hertwigs ('79); aiul a 

 third, in which it is that described by iladdon ('87), H. V. Wilscm 

 ('88), and myself ('91). I have shown, lioMcver ('91 a), that the third 



