MEMOIRS OF THE NATIONAL AC^ADEMY OF SCIENCES. 493 



disappear downward so do the septa; or conversely, as the pairs of mesenteries increase in num- 

 ber the septa follow. This is well shown in the series of sections of the bud of Sohnmtrsea 

 on PI. XII. In the transverse section, represented in tig. 89, nine protocnemes are present, 

 and the same number of alternating septa, the latter indicated by the septal invaginations. Only 

 one representative of the incomplete lifth and sixth pairs (V) occurs at the stage, and with it is 

 associated a smaller septum. This mesentery disappears a few sections below, and then the 

 septa become octaraeral, in correspondence with the four pairs of mesenteries i-emaining; as 

 others of the mesenteries disappear the septa are found to follow, and in the last section obtained 

 (fig. 90) only six mesenteries and six septa occur. 



Whatever be the arrangement of the septa in adult corals, all the evidence seems to indicate 

 that either six or twelve protosepta constitute the fundamental plan for all Madreporaria, 

 whether recent or extinct. Thus, by means of serial sections, Prof. G. von Koch (1889) has proved 

 that the coral CanjophylJia nigom, which Moseley (1881), from his studies of the adult, tirst 

 described as octameral, is reallj' hexamei'al in its early stages. At first, six septa of the primary 

 cycle are present, then six septa of the second cycle, and it is only with the appearance of a third 

 cycle that modifications are intnjduced which lead to the production of the adult octameral 

 condition. Lacaze-Duthiers (189-1) refers to the eight or ten systems of septa exhibited by 

 Flahelhiiii, yet the early stages are typically hexamerous. 



Lindstrom's results in I'egard to the coral Duncunla are also of the same character. 

 Pourtales (1871) regarded this in the tirst instance as a living meml)er of the order Madreporaria 

 Rugosa. With regard to the tetramerism of its septa, Lindstrom, as quoted by Pourtales, 

 writes: 



" There seems to oe no I'eason to class this species, Duncania, among the Rugosa, which 

 commonly are considered to have four septa of the first order. In making a thin section of the 

 apex of a Duncania I distinctly saw six septa of the first order, which met in the center." 



There is also evidence, from the results of Ludwig and Pourtales, that some of the Palaeozoic 

 Tetracoralla exhibit a like hexameral primary condition." 



Solenastrxa Jnjude-'< is hexameral as far as the third cycle of septa, but the additions beyond 

 this take place in such a manner as to wholly destroy such symmetry. The septa of adult polyps 

 of Ma?iicma and Favia, which reproduce by oral fission, exhibit no hexameral plan, yet the early 

 stages have been found to be of this type, and such can probably be assumed of fissiparous corals 

 generally. 



The instances are sufficient to show how little importance can be attached to the disposition 

 of the septa in the adult coral, when discussing the broad relationships within the group. The 

 whole history of any form must usually be known before its true nature can ])e determined, 

 or any great importance assigned its adult peculiarities. The soft parts afi^ord the surest guide 

 for morphological comparison, from the greater certainty with which the relationships can be 

 determined, owing to the man_y structural details available for correct orientation. 



It has already been shown that, as regards the appearance of the twelve protocnemes, there is 

 evidence of practical uniformity throughout the Actiniaria and Madreporaria; the several diver- 

 gences in the later mesenterial sequence, distinguishing the great grouj^s, make this their starting 

 point. As regards the septa, also, there is good reason to expect that the six or twelve proto- 

 septa will be found characteristic of both living and extinct corals, and that all the numerous 

 types of metaseptal sequence likewise make this their point of divergence. The distinctive 

 characters of the principal divisions of the Zoantharia are not manifest from the beginning, but 

 from the completion of the jirotocnemic stage. 



METASEPTA. 



The endeavor to establish the order of appearance of the septa beyond the primary stage 

 presents many difliculties; no uncertainty can possibly arise as far as the protoseptal stage, but 



"The subject is further discussed in a paper: "Relationships of the Rugosa (Tetracoralla) to the living Zoan- 

 thew." Johns Hopkins Univ. Circ, vol. xxi, no. 15.5; also, Ann. Mag. Nat. Hist., ser. 7, vol. x, May, 1902. 



Vol. 8— No. 7 7 



