7. Anthozoa (incl. Hydrocorallia). A. Zoantharia, 23 



they are formed in the endoderm and not ingested. A circular endodermal 

 musculature is usually present. It is better developed distally giving rise in 

 some species (e. g. Orbicella cavernosa and annularis) to a typical diffuse 

 sphincter. This is even more strougly developed in Isophyllia dipsacea 

 recalling the "restricted" sphincter of the Actinian Macrodactyla. The amount 

 of the development of the sphincter muscle appears to be dependent on the size of 

 the polyp, those of 0. and Is. being among the largest studied. The expansion 

 of the polyps begins soon after sunset and is maintained during the greater part 

 of the night. The tentacles are usually in two alternating cycles, the inner 

 one entoccelic and the outer exocoelic. In Agaricia and in fully developed 

 apical polyps of Hadrepora the latter is wanting. The nematocysts are in 

 irregular patches along the tentacle and in a knob-like battery at the apex. 

 Complete tentacular introversion may occur and the tentacles may even 

 apparently disappear becoming involved in the greatly expanded margin of the 

 disc. In Madr. and For. there are only 6 to 12 tentacles. Polyps at the apex 

 of long established branches of Madr. have 6 entoccelic tentacles but polyps 

 at the ends of short rapidly growing branches have rudiments of 6 others 

 (exocoelic) and in regions of vigorous growth polyps may be seen with the 

 full complement of 12 tentacles. It appears that on any polyp assuming the 

 axial condition the 6 exocoelic tentacles tend to completely disappear. In 

 Siderastrtea radians and sid&rea the exoccelic tentacles are simple but the ento- 

 ccelic are bifurcate due to the fact that two originally separate tentacles are 

 now borne upon a common stem. In the larvas the 12 protocnemes are 

 generally developed at the time of fixation or shortly afterwards and the 12 primary 

 tentacles appear either simultaneously or one cycle in advance of the other. 

 In the latter case the 6 entoccelic tentacles usually appear first and are followed 

 by the 6 exoccelic but in Sid. radians this order is reversed. In the earliest 

 budding polyps of Madr. and For. 2 median and 4 lateral tentacles are seen, 

 later ones show 8 or 10, the tentacles thus arise in median and then in 

 successive bilateral pairs. Siphonoglyphs are not found in Madreporarian 

 polyps, the ectoderm of the stomodseum is uniformly ciliated and gland cells 

 and nematocysts are present. The colour of corals is due to pigment granules 

 in ectoderm cells (Mceandrina, For. etc.), Zooxanthellse (nearly all corals) or 

 green granules (Dichoccenia) or pigment cells (For.} in the endoderm, or per- 

 forating red and green algae in the skeleton. The 12 protocnemes of buds 

 and of larvae arise as 6 bilateral pairs in a definite sequence which is probably 

 the same throughout the Madreporaria and conforms with that found in most 

 Actiniaria. Two pairs of directives are formed from the 3 rd and 4 th pairs of 

 mesenteries. The first 4 pairs of mesenteries unite with the stomodaeum 

 (Edwardsia-atage) in the order of their appearance and a long interval elapses 

 before the 5 th and 6 th pairs become complete. The first cycle of secondary 

 mesenteries arises in unilateral pairs in the primary exoccels and the second cycle 

 of metacnemes in successive isocnemic pairs in the exoccelic chambers between 

 the pairs of the two previous cycles. In polyps reproducing by fission the 

 mesenteries arise as isolated exoccelic pairs in regions of most forward growth 

 and each and all the pairs may become complete. In Madr. the mesenterial 

 filament is divisible (as seen in transverse section) into 3 areas the middle 

 one of which corresponds to the "Drtisenstreif" of Actiniae and the two lateral 

 to the "Flimmerstreifen". The filaments of many corals undergo peculiar 

 modification becoming, within restricted limits, almost wholly glandular. There 

 are no cinclides but the filaments may be extruded through temporary perfo- 

 rations. The filaments in the Madreporaria seem to appear independently of 



