428 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



another. Mnd Ofciipv nearly the whole of the exposed poly pal area (PI. XXII, fig. 1.5(i). In the 

 living eonditioii each appears to arise either directly over or near the centripetal termination of 

 the septum with which it corresponds. The C3'clical arrangement is difficult to establish, and in 

 many instances this would be impossible without the assistance from the septa which can be seen 

 below through the soft tissues. 



On full expansion the inner tentacles are found to consist of a short c}' lindrical stem, which 

 bifurcates a little l)eyond midway, each half bearing a spheroidal enlargement at the apex; the 

 outermost tentacles, however, are simple, consisting of a short stalk, terminated by a knol)-like 

 swelling (PI. XXII, fig. 1.51). Thus in Sidi'ranfra'K thei'c is a true dimorphic condition of the 

 tentacles, apparently the only instance of such among the Madreporaria. if we except the 

 differences between the radial and axial polyps of M(iih-ci«irn. In the course of the development 

 of the young polyps (p. 533) it has V)een. ascertained that the inner tentacles are at first simple, then 

 afterwards another moiety arises over the same mesenterial chamber, and finally a common stem 

 is produced, which bears the two halves at its extremity and raises them above the disk. 

 Ontogenetically, therefore, the bifurcations represent distinct and separate formations, and only 

 later constitute an entire tentacle. 



Subsequent examination of sections confirms what woidd be expected from the external 

 relationships, namely, that the l)ifurcated inner tentacles are all (Mitoccelic in position, while the 

 simple outermost tentacles communicate with the exocoeles. In the nearlv mature polyps of a 

 colony, however, some of the entocoelic tentacles may be simple, Ijut such are merely examples 

 in process of development. The exoccelic members are never double. 



On retraction of the polyps the disk and tentacles remain uncovered, the column wall in 

 Siderastrsea being incapable of overfolding. The tentacles are now represented by minute, 

 simple and double tubercular enlargements, scattered over the greater part of the polypal wall. 

 Microscopic sections reveal that the stems are no longer determinable as such, having become 

 involved in the discal tissues, while the knobs remain as mere ectodermal thickenings (PI. XXIII, 

 tigs. 154. 155). The apex of the exoccelic tentacles occurs as a simple swelling of the disk, 

 directly overlying its cori-esponding septum, while the two knobs of the entoccelic tentacles are 

 disposed one on each side of an entocoelic septal ridge, the two halves connected by a tissue 

 similar to that of the disk, which manifestly represents the stem of the expanded tentacle. 

 The ectoderm of the knobs includes a peripheral layer of long narrow nematocysts, and is thus 

 easily distinguished from the rest of the disk. 



Both S. f(t(1!(tni< and S. siderea are further characterized l)y th(» tentacles being apparently 

 arranged in only approximate cycles, and by the occurrence of a comparatively wide interspace 

 between one cycle and another. The imperfect cyclic disposition results from the pre.sence 

 of tentacles intermediate in position l)etween the true cycles, and on the a 'tual polyp it is 

 often very puzzling, if not impossible, to say to which cycle some of the tentacles should be 

 relegated. Polyps are found with from five to seven or eight tentacles, which, so far as their 

 position is concerned, must be regarded as belonging to an inner cycle, and the remaining 

 members seem to come in irrespective of any cyclic plan. In mature polyps of both species 

 three more or less complete alternating cycles of tentacles are actually present, in addition to 

 the outer single-knobbed cycle. The members of the latter biMng situated near the polygonal 

 periphery of the polyp are rarely included within a circle. 



When studied in conjunction with the underlying septa an apjiroximate tentacular regularity 

 can be established, as in fig. 150. The innermost cycle comprises six double-knol)bed tentacles, 

 separated by a wide interspace from the members of the second and third cycles, and these latter 

 cycles are separated from the outermost cycle of single-knobbed tentacles. It is manifest 

 from the figure that tlu^ tentacles correspond with the se{)ta, and not all the twelve meml)ers 

 necessary to complete the third cycle occur. S. radlaiix appears to never complete its third 

 cycle of mesenteries, tentacles, and septa, while in ^S". sidet'&i it is occasionally reached or even 

 exceeded. In fig. 150 only one meml)er is wanting to completi> the third cycle of entotentacles. 



The apparent irregular disijosition of the tentacles in this genus becomes explicable on a 

 knowledge of the development of the mesenteries and their corresponding septa, or rather the 



