ON VAEIATION IN THE ASTHAN&IACEJE. 395 



On some Structures liable to Variation in tlie Subfamily Astran- 

 giacese (Madreporaria) . By Sxttabt O. Eidlbt, M.A., F.L.S., 

 Assistant in the Zoological Department, British Museum. 

 [Read 21st February, 1884.] 

 (PiATE XVI.) 



It is one of the chief functions of Zoology (as distinguished from 

 Comparative Anatomy) to bring together the (essentially) like 

 and to separate the (essentially) unlike. The question, What 

 constitutes essential similarity and dissimilarity ? is one of the 

 earliest problems which confront the zoologist, and, if not satis- 

 factorily solved, will prove a stumbling-block at every turn ; it 

 is, moreover, as applied to the lower terms of our classification, 

 viz. genera and species, a question with which a priori considera- 

 tions have very little to do, and which can only be satisfactorily 

 solved, as a rule, by study of a tolerably extensive series of the 

 very genera and species to which the problem is to be ajDplied. 



Perhaps in no group is it more necessary to bear these facts in 

 mind than in the Madreporarian Corals. The constantly decrea- 

 sing value placed upon certain structures, viz.tabulse, coenenchyma, 

 and epitheca, and upon the number 6 as an index to the septa 

 for group-distinction, illustrates the truth of this. In the fami- 

 lies Turbinoliidae and Oculinidse the characters derived from 

 the presence or absence of the numbers of the pali, from the 

 structure of the columella, and (but to a less extent) the condi- 

 tion of the costae, form leading points for generic and specific 

 diagnosis. In that group of the great family Astrseidse, as at 

 present constituted (Milne-Edwards and Haime), which probably 

 shows the closest general resemblance to these families, viz. the 

 Astrangiacese, the columella, costse, and pali form lobes represent- 

 ing physiologically the pali, inasmuch as they frequently exhibit a 

 similarly advanced development, and have been naturally resorted 

 to for purposes of classification *. But the object of this paper 

 is to show that, whatever may be the value for classification of 

 the corresponding parts in the Turbinoliidse and Oculinidse, they 

 must be employed with great caution in the zoological study of 

 the Astrangiacese. 



* See Duncan, Proc. Zool. Soo. 1876, p. 439; Yerrill, Bull. Mus. Oomp. 

 Zool. i. p. 47, Trans. Conn. Acad. i. p. 525 et seq. ; Milne-Edwards and Haime 

 Hist. Nat. Oor. ii. pp. 606-620. 



LINN. JOURN. — ZOOLOGY, VOL. XVII. ' 30 



