CORALS FROM MURRAY, COCOS-KEELING, AND FANNING ISLANDS. 59 



This criticism is just, and because I was aware of the defect so rightly 

 emphasized I designated the classification "tentative." The principle on 

 which M. Dollfus insists must be the guiding one. It is my conviction that 

 any attempt to base a classification of corals on living forms alone is fore- 

 doomed to failure. The forebears as well as the children must be studied. 

 Since the genera Orbicella, Favia, and Favites are as clear-cut in the Oligo- 

 cene and upper Eocene of the West Indies as they are in the modern faunas 

 of Cocos-Keeling Islands, their phylogeny will probably not be ascertained 

 by a study of living species. Orbicella and Favia persist in the living West 

 Indian faunas, but Favites is not known in the Atlantic Ocean. 



As regards the delimitation of species, Mr. Matthai and I are nearly 

 in agreement; but as he has not followed the accepted rules of nomenclature, 

 many of the names used by him are invalid. In some instances, the names 

 applied by him are antedated by those proposed by Dana; but until he had 

 examined the types, he could not be sure of Dana's species. I have made 

 most of the needed revisions in the systematic discussion of the species. 



There is one point which can scarcely be too strongly emphasized, 

 especially as it is so often neglected by workers on corals. Genera must be 

 based 07i type species. The neglect of this matter by modern workers is 

 remarkable, as Fischer de Waldheim designated a type species for Hydno- 

 phora and Milne Edwards and Haime endeavored to ascertain or to designate 

 one for nearly every genus considered by them. I am either giving the 

 previously determined type species, or, when none has been determined, I 

 am designating one for every genus herein considered. 



The photographs used in illustrating this paper are from various sources. 

 Dr. Wood Jones has furnished several; a few, from negatives made through 

 the kind oflfices of Professor J. Graham Kerr, of the University of Glasgow, 

 illustrate the types of Ellis and Solander; the U. S. National Museum has 

 supplied most of the photographs of Dana's types; and the U. S. Geological 

 Survey has contributed six photographs of Dana's types and the two photo- 

 micrographs of Cceloseris mayeri. All the other illustrations were supplied 

 by the Carnegie Institution of Washington. There are about 340 of these, 

 of which 80 were made by Miss L. B. Gallaher in the photographic laboratory 

 of the U. S. National Museum, and about 260 were made by Mr. W. O. 

 Hazard, of the U. S. Geological Survey. 



I wish to thank Dr. George Otis Smith, Director, and Mr. David 

 White, Chief Geologist, of the U. S. Geological Survey, for the interest 

 they have taken in this paper, and the encouragement and assistance they 

 have given me. It has been prepared as a part of my duties as an officer 

 of the Survey, which has also furnished the requisite clerical assistance 

 and the services of a preparator in caring for and handling the collections. 



