REMAEKS ON THE CLASSIFICATION OF CORALS. 55 



their succe.s.sioii in geologic time. No one has approached the subject in tliis 

 manner. Bernard is working along these lines, but his work is not far 

 advanced. The past classifications have been based on some particular 

 featvires, usualh' of the skeleton, without reference to the whole structure 

 and 1 listory of the organisms. The classification of Milne-Edwards and Haime 

 was based on the gross morphology of the skeleton; that of de Fromentel 

 on the mode of growth. Duncan based his on a combination of the gen- 

 eral skeletal featvires and mode of growth, but evidently did not seek to 

 find what characters were of phylogenetic imjiort. Von Heider and Ort- 

 mami based theirs on the structure of the wall, which has been shown to 

 possess almost no systematic value. (3gilvie was carried away by her 

 researches into the microscopic structure of the septa of corals, mixing with 

 her results the false princii^les used by Pratz, von Heider, and Ortmann in 

 their attempts at classification. She has made but small contribution to 

 formulating a true system of classification. The work of all these, and of 

 man^■ other investigators, has continually added to our knowledge of corals, 

 until now we understand fairly well the whole make-up of the coral skele- 

 ton and much of the relations existing between the skeleton and soft parts, 

 even if there is still a great deal to be learned. As oiu' knowledge of fossil 

 faunas increases, we know more and more of the succession of the various 

 species, genera, etc., in time. Therefore, probably jjefore many years, some- 

 one may be able to give us a classification based on the actual phylogenetic 

 grouijing of the various genera. 



If the above represents the status of the question of classification, it 

 is evident why I have not adopted or proposed any system. 1 will state 

 that I believe the family Turbinolida?. shdlild be divided into several 

 families. Flabellum has not as much in common with Platyti'ochus or 

 Discotrochus as it has with Eusmilia. I have projected a revision of the 

 Turbinolid genera, but it is not in a condition for presentation. 



. There is no means of determining how many difi'erent types of corals 

 are represented by the old Astreid family. 



Under these conditions the only possible thing to do was to describe 

 my material with all the care possible, so as to aid in the future work of 

 rean-anging the various coral genera. 



