448 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



The adult j^olyps of Cltidocora nrliisculu always contain six pairs of complete mesenteries, 

 constituting a first cycle, and six alternating pairs wiiich remain inc(ynipleto and form a second 

 cycle. Representatives of a third cycle are usually developed, ])ut instead of consisting of twelve 

 pairs, one in each (>xoco?le between the previous twelve pairs, only four or six pairs are usually 

 present, all on the same aspect of the older paii's (PI. VI. tig. 49). Earlier stages in the develop- 

 ment reveal that this is probably the sulcar aspect of each s}'stem (p. 458). 



In Siih/raxfnt'ii. radiana six pairs of complete mesenteries are present, along with .six 

 alternating incomplete pairs, and a few pairs belonging to the third cycle may also occur; 

 usually the third-cycle pairs are radially .shorter than those of the second cycle, but at other times 

 they nearly equal them in size. In the larger species, S. tddrvra. though more memiicrs of the 

 third cvcle are present, the whole twelve pairs necessarv to complete th(> cvcle are rarelv present 

 (PI. XXIII, %. 153). 



The polyps of the seven genera described, all produced asexually by the process of 

 gemmation, are thus chai'acterized 1)}' the \ cry r(>gular disposition of the mesenteries in 

 alternating hexamerous c3'cles. The tir,st and second cycles are fully developed in all the adult 

 polyps, while the third cycle may be only partly formed, but so far as it goes the members 

 alternate regularly with the other pairs, according to the oi'der of appearance establisli(>d on 

 p. 455 et. seq. In all the polyps two pairs of directives occur in the first cycle. So far as the 

 mesenterial arrangement is concerned, thei-e seems no difference between a polvp originating as a 

 bud and one deri\ed from a sexually produced larva; both follow tiie normal hexactinian plan. 



Onl\' the members of the first and second orders ever become inserted on the stomodivum 

 in the species studied. The later orders never become complete, but rc^tain a definite size 

 c^haracteristic of the species. 



MESENTERIES IN GENERA REPRODUCING BY FISSION. 



The asexual rejjroduction of the following genera takes place mainly, if not entirely, bv 

 stomodiBal fission: Aijarieia, IsopJiyllia, Dichoca'nia, Favia^ Ifanicina, Mivnndrina, and Colpo- 

 •phyll'ia. In the first four the polyps so produced may become more or less distinct from 

 one another, each with its own .system of tentacles and a column wall; in the remaining genera 

 the separation is incomplete, and meandering discal, tentacular, and columnar systems are pro- 

 duced in place of distinct polyps, and only exceptionally are transverse walls developed, 

 which separate one series of oral apertures fi-om anothei'. Sections have been made through 

 polyps of each of the above genera, and reveal a mesenterial arrangement very diflereiit from 

 that already described for genera where asexual reproduction by gemmation is the rule. 



Transver.se .sections through two polyps of Agarici<i fmglUs are represented on I'Is. XXIV 

 and XXV, while the arrangements of th(> mesenteries of two different polyps of IsophyUlit are 

 diagrammatically shown on next page. Fundamental differences are at once apparent, compared 

 with the mesentei'ial plans already described. No directive mesenteries occur in these nor in any 

 of the other (examples studied Very rai-ely the number of complete pairs may be six, but is usually 

 ii-regidar, while the incomplete mesenteries vary greatly in number, size, and relation to the 

 complete pairs. The hexameral plan is altogether departed from, and each stomodanim may have 

 from ten to twenty -five complete mesenteries a.s.sociated with it. A regular alternation of .second 

 and third cycle mesenteries is found in only one or two jilaces, as at the upper right-hand region 

 of fig. h; here and there a single unpaired mesentery may occur within an exociele. Of the many 

 polyps of each species examined no two display exactly the same number and relationship of the 

 mesenteries. 



A like absence of hexamei-al, or any oth(>r, regularity occurs in the ])olyps of DicJioccenia 

 and 7v^/vr^ (Pis. XIII and XVI). Transverse sections reveal a variable number of pairs of 

 perfect mesenteries from four upward, according to the size of the polyj), while the 

 alternating incomjjlete pairs are rarely the same in number and size in any two exocieles. and no 

 directives occur. 



In sections of mature colonies of Mdiiieinn arcohita only two series of mesenteries can he 

 generally distinguished, complete and incomplete. tlu> latter rarely affording evidence of alternating 



