MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 541 



genera, is >ut a mere fragment of the group Eor this purpose, even when combined with the 

 knowledge available from the labors of others. It should, however, be taken as a contribution 

 around which other researches may accumulate." 



The only fundamental distinction which the facts as yet seem to warrant is the recognition 

 of the two great groups, Entocnemaria and Cyclocnemaria. The material studied is insufficient 

 to determine the exact taxonomic value of the canalicular system and the formation of synapticula. 

 Their small morphological value has been noticed, but it is not certain as to how far other 

 important structural features may be associated with their presence. From the conditions 

 already described in Siderastrsea and Agaricia, there would seem to be some connection between 

 the occurrence of synapticula and peculiarities in the tentacular system. I have therefore for 

 the present retained the divisions of Perforata and Aporosa, and Fungacea, until further results can 

 be obtained which will enable their precise significance to be understood. 



From the few generic representatives studied, it would lie premature to attempt a diagnosis 

 of the polypal characteristics of the families included, and I have therefore omitted such entirely. 

 Xo doubt some of the generic characters given should rank as of family importance. I have 

 attempted to define the genera in terms of the polyp, at the same time giving the species from 

 which alone the characters have been drawn. To the definition of each genus 1 have added the 

 diagnosis given by Duncan (1885) in the " Revision," so that in each case the skeletal and polypal 

 descriptions can be compared side by side. It will be understood that the latter definitions are 

 founded entirely upon the types studied, and consequently may require modification as the soft 

 parts of other representatives become known. Before any genus or species of coral can be fully 

 known, it is necessary, of course, to possess descriptions of both the skeleton and the polyp, but 

 the scope of the present work is wholly confined to the latter, and does not presume a complete 

 systematic account of the species. 



I have fully described only a single representative of each genus, though in many cases 

 other species have been investigated. In the interests of systematic studies upon the West 

 Indian Madreporaria, it is highly desirable that a comparative description should be made of the 

 polyps of as many representatives of each genus as possible, in order to determine the range of 

 variation, and material for such is now in hand. 



" Certain results, bearing upon the morphological classification of the Madreporaria, have presented themselves 

 since the above was written, and are given in a series of four papers, "The Morphology of the Madreporaria," appear- 

 ing in the Annals and Magazine of Natural History, 1902-3. The distinctions already found between corals repro- 

 ducing by gemmation ami by flssiparity obtain a greater phylogenetic significance from the fact that apparently the 

 one or the other method of growth is altogether characteristic of any particular species. There seems to be no 

 intermingling in a species a- regards the two methods of asexual growth; it is shown that the few instances of 

 simple fission hitherto considered to occur in corals which are usually gemmiferous, are really examples of a peculiar 

 method of budding, which I have termed "fissiparous gemmation." The distinction between fissiparous and 

 gemmiferous corals must he regarded as of greater taxonomic importance than is accorded it above. Further, the 

 mesenterial increase beyond the protoenemic statre, occurring in Poritrs and Madrepora, is shown to be associated with 

 fissiparous gemmation, hence the process can not be compared with the metacnemic growth in corals generally. 

 Studies upon other than West Indian representatives of the genera are necessary before the full morphological value 

 can be assigned the great differences between the mesenterial plan of Porites and Madrepora and that of other corals. 



Vol. 8— No. 7 10 



