422 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



upon living colonies of Siderastrsea the tentacles of the expanded polyps at once close upon them 

 and prevent their escape. 



The detailed arrangement of the tentacles presents many ditlerences in the various species 

 studied. As seen externally, the twelves tentacles of Madrejyora and Poritex (Fls. I and IV) admit 

 of no proper distinction into an inner and an outer series, though varying somewliat in size, and 

 may therefore l)e described as acyclic or monocyclic; the a2)ical polyps of the former genus 

 bear only a simpUi cycle of six equal tentacles. Although forming only one cycle, the twelve 

 tentacles in l)oth genera represent two orders, constituted of six entotentacles and six alternating 

 exotentaclcs. The tentacles of such genera as Orh'icclln, S(Ae)t<(i<tr»'a. OcuVimi, Chuhirora, and 

 A'iffdiKjiti. whose asexual method of reproduction is by gemmation, usuallv exhibit a regular 

 hexameral multicyclic arrangement, with the formula (I, t), 12, 24, etc. Very often the tirst and 

 second orders are arranged so as to form oidy one inner cycle of twelve memljers, with which 

 tlic twclv(> members of the second cj'cle alternate; the third cvcle of twenty-four alternates 

 with both these, and so on, according to the number of cycles developed. Where only twenty- 

 four tentacles arc present they usually appear as an inner and an outer cycle. 



Very often the hexameral sequence of the tentacles is not complete, especially in Cladocora 

 and Ant ran (I id. The total luimber of tentacles in mature polyps of Oladoeura arhuwula 

 varies from thirty-two to thirty-six, whereas the complete hexameral plan would require forty- 

 eight as the next number after twenty-four has been reached. In describing below the 

 development of the later tentacles of polyps, it is found that the organs do not arise a complete 

 cycle at a time, but in simple or d()ul)lc pairs on each side of the median axis, and in manv 

 species a tentacular cycle once commenced is not always completed before the polyp attains its 

 full size and growth ceases. In such a case it is clear that any intermediate number of tentacles 

 between the commencement of a cycle and its completion may be present. 



Whenever an entocoelic tentacle appears, a corresponding exocoelic member usually arises, 

 either simultaneously or shortly after, so that the number of exotentaclcs comprising the outer 

 cycle is always equal to the sum of the entotentacles of all the inner cycles. Hence in endeav- 

 oring to establish the cyclic scheme of anv hexameral polyp, in which the number of tentacles 

 may be intermediate between twenty-four and forty-eight, or forty-eight and ninety-six, 

 incompletion must be looked for in the two outermost cycles, not in the outermost only. A 

 polyp of CJndocara with thirty-two tentacles bears the cyclic formula 6, 6, 4, 16; one with thirty- 

 six tentacles the formula 6. 6, 6, IS; with forty tentacles 6, 6, 8, 20, where the numbers 16, 18, 

 and 20 represent the exotentaclcs. The exoccelic cycle thus increases by the same amount as 

 the outermo.st entoccelic cycle. The order of appearance of the tentacles demonstrates that it is 

 impossible to establish hexameral completion for all the cycles until the outermost is reached, and 

 then relegate any omissions to this, as is usually attempted. Ka\\ omission due to hexameral 

 incompletion ati'ects both the exocoelic and the last entocoelic cycles. 



The members of an^' tentacular cycle are neaidy always alike in size, but the tentacles 

 usually show a diminution in length in passing from the inner to the outer series, a condition 

 expressed by the term entacmjeous. The organs in Madrepora and Pontes are exceptional in 

 that they vary in size in a very definite manner in the same cvcle, while Orhicella radiata offers 

 a marked exception to the entacmreous order. The polyps of this species do not readily expand 

 their tentacles, and hence are not always favorable for observation. When fully extended the 

 different tentacular cycles are found to be widely apart, and the twelve members comprising the 

 first cycle are much smaller than those of the next, and even less than the tentacles of the 

 outermost cycle. The tentacular plan of a polyp of O. radiata, with thi-ee hexamerous cycles, is 

 represented on the next page. 



Usually the different cycles which constitute the crown of tentacles are closely arranged in 

 a nari-ow marginal zone, so that basally the members of one cycle partly embrace those of the 

 next. O. rad'tata^ just mentioned, is again exceptional in that a wide discal interspace separates 

 one cycle from the next, the tentacular crown being unusually broad, encroaching upon the 

 peristome. The same feature is also characteristic of the polyps of Sidfrastrira and Agaricia; 



