440 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



and Pontes astrixoides likewise often exhibit a bright pink color ininiediatel_y below the surface 

 of the corallum. In the latter it is still obvious on dried specimens twelve months after collection." 



Where the soft tissues of the polyjis themselves are not densely colored, a green or reddish 

 skeleton below must partly determine the general effect by reflection through the polypal wall. 

 Lacaze-Duthiers (1897, p. 5) found the coralla of all the Mediterranean corals examined ]>y him 

 to be perfectly white when deprived of their soft tissues, hence concludes that the coloration 

 is wholly polyi)al. 



The surface of many polyps often presents a brillant emerald-green color, arranged in streaks 

 or patches. It is found mostly on the disk as a peristomial ring, but occasionally on the oi'al 

 aspect of the tentacles. Such has been observed in Sidera^streea radians^ Orhicella annularis, 

 Agaricia aguridtes, and ColpophyUia, as an addition to the more usual colors. The phenomenon 

 is wholly superticial, not intracellular in origin, for histological examination fails to reveal any 

 structure which can be assigned as its cause. The l)right green and reddish colors of the other- 

 wise colorless and transparent Phyllangia amerlrana and Astrangia soUtaria are also of this 

 character, and pigment graiudes are here absent from the ectoderm, as well as zooxanthelhe 

 from the endoderm. 



The causes of coloration in living corals may be thus summarized: 



(1) Ectodermal . — Pigment granules in ectodermal cells: aggregated in more or less isolated 

 patches (Isop/ii/llia, Mxandrlna,, etc.); generally distributed {Porites, Dichocmnid). Superficial. 



(2) Endodermal. — Yellow cells or zooxanthelhi? (nearly all coi'als); green granules {Dicho- 

 cmnia); pigment cells (Porites). 



(3) Skeletal. — Perforating green and red alg£e. 



MESENTERIES. 



In all recent morphological studies, which have for their object the determination of the 

 natural relationships of the Anthozoa, a greater signiricance is attai'hed to tlie order of appear- 

 ance and adult arrangement of the mesenteries than to any other polypal characteristic. These 

 sei've not only to limit the j)rimary divisions of the grou]) — Alcyonaria, Antipatharia, and 

 Actiniaria, but in the Actiniaria the chief subdivisions — Hexactinia', Zoanthea', Cerianthea', are 

 likewise determined by the mesenterial features; and even for minor classiticatorA' purposes the 

 variations in the organs are often of great service. It will l)e found that a similar importance 

 nuist l)e accorded the mesenteries in the ]\Iadreporaria, and dependent ujwn their arrangement 

 is that of the tentacles and septa, as well as other less important relationships. 



Among the adult polyps here studied, the simplest condition of the mesenteries is that 

 occurring in the genera Jfadrtj/ora and Porites. This is represented in transverse sections in 

 the figures on Plates I and HI, and, for the stomodieal regions, diagrammatically in the accom- 

 panying fig. 3. The latter will also serve as an illustration of the terminolog}' to be employed, 

 and also for certain cognate ujorphological considerations, which may be first considered. 



In practically all the polyps of the two genera mentioned only twelve mesenteries occur, 

 arranged in six pairs with regard to a median axis, which includes the longer diameter of the 

 stomodauun. The four pairs reaching the stomodanxm are known as complete mesenteries, 

 while the two remaining pairs, not united with the stomodwum, are incomplete. The Roman 

 numerals I-VI correspond with the established order of appcai'ance of the pairs of mesenteries 

 in coral larvie to bo descri))(>d later. 



The twelve mesenteries, characteristic of the adult Madrepora and Porites. occur also 

 in the larval stage of genera of which the adults possess more than six pairs, and are 

 strongly marked off' by their mode of origin from the mesenteries subsequently developed; 

 they serve, in fact, as the starting point for various types of mesenterial arrangement of the 

 highest importance in Zoantharian studies. In 1900, I found it desirable to introduce some 



"The subject of coral boring Algir liat^ been more fully dealt with in a paper: "Boring Algse as Agents in the 

 Disintegration of Corals." Bull. Amer. Mus. Nat. Hist., Vol. XVI, liioi'. 



