126 ACTINI.E OF ALBATROSS EXPLORATIONS M'MURRICH. 



thiuji'. In clioosiiii;' the lelativo coiLsisteiicy of the various forms as a 

 basis for his division of the Zoantharia Miliie-Edwards iiaturnlly falls 

 into certain of the errors which such a chissification entail?, and which 

 had been Ininded down i'roni ciulier days, as for instance the grouj)- 

 ing of the Millepores with the INIadreporaria. In this, however, there 

 is neither loss nor gain, since none of his predecessors, with the cou- 

 spicuous exception of K;ipp, had suggested the reference of these forms 

 to their proper position. The principal error of the classilication, as 

 already pointed out, lay in the attaching of too great importance to 

 the presence or absence of a corallum. and in the disregard of the sim- 

 ilarity of the soft parts of Madreporaria and Actinaria so definitely 

 stated by Dana. 



Milne-Edwards's classification had a marked infiuence upon later 

 writers, most of whom adopted his larger divisions, the principal mod- 

 ifications introduced by them attecting the arrangement and definition 

 of the lesser groups. An exception to this, however, was the classifi- 

 cation of Gosse ('GO) who adhered to the arrangement laid down by 

 Dana, but went a little further than that author in dividing certain of 

 the tribes of Actinaria into families, thus: 



SiibonU'r Actinaria — 

 Tiilie I. Astra^aoca : 



Family I. 2*Ietri(lia(la' = forms with componnd tentacles. 



n. Sagartia(Lp = with simple tentacles, adherent base, and colunm 



pierced by cinclides. 

 111. Antheada^=: column smooth and iiiiptrt'orate. margin simple. 

 IV. Actiui,nda> = margin beaded. 

 y. Bnnoditla':=; column warted. 

 XL llyauthida'= base non-adherent, rounded, simple. 

 VII. Minyadida' = base nou-adberent inclosing an air chamber. 

 Tribe II. (^iryophylliacea : 



n. Without a corallum. 

 Family 1. ("apueada^-=simph'. 



II. Zoanthida'=;comi)ound. 

 fS. AVith a corallum, certain corals divided into four families. 



The coralligenous Astra'acea Gosse does not classify,noneof the gen- 

 era being Jkitish, nor does he divide the Madreporacea or Antipathacea 

 into families, for the same reason. The Lucernaria<he he excludes from 

 the Actinaria, recognizing their affinities to the JMedusfe. 



On comparing this classification with that of Milne-Edwards, it will 

 be seen that, independently of the associati<m of non-coralligenous and 

 coralligenous forms, there is a very ditferent grouping of the genera. 

 The family Metridiada^ (a name badly chosen) is equivalent to IMilne- 

 Edwards's Thalassianthina' and Phyllactina', the Sagartiada' are the 

 Aetinines perforees, and the Bunodida- the Actinines verruciueuses, 

 both raised to the rank of families. The Actinines vulgaires are 

 divided into three families, two of wliich, the Autheada^ and Aetiniad;e 

 belong to the Astranicea, while the third, the Cai)neada', is referred, 

 with the Zoanthidae, to the Caryophylliacea', while JVIilne-Edwards's 



