MEMOIRS OF THE NATIONAL ACADE^IY OF SCIENCES. 495 



of the early stages become the entosepta of the later or adult stage, or whether the exosepta 

 remain exosepta throughout their existence. It is clear that whichever method is followed will 

 lead to very different fundamental results in formulating the law of septal sequence. 



The former condition appears to be verj- generally assumed. According to this, the six 

 primary exosepta are considered to represent the second order of entosepta found in the developing 

 coral in which the primary and secondaiy orders of mesenteries have appeared. The outermost 

 C3^cle of twelve septa, constituting the third cycle, would be exosepta, and represent new forma- 

 tions, alternating with the twelve older septa. The twelve pairs of tertiary mesenteries on their 

 appearance would include these thii'd-cycle septa within their entocoeles, and thus the exosepta of 

 a previous stage would become the entosepta of a later stage. The next cycle of septa would 

 consist of twenty-four exocoelic members alternating with the first, second, and third cycles, and 

 on the appearance of the quaternary mesenteries these would likewise become entosepta, followed 

 by a fifth cycle comprising onlv exosepta. 



My investigations so far give support to the view that the exosepta remain exosepta 

 throughout, and that new entosepta arise with the new pairs of mesenteries. The results, how- 

 ever, are not yet sufficiently complete to fully estaiilish the method. Such a relationship agrees 

 with what actuall}' takes place during the growth of the different C3cles of tentacles; an exoten- 

 tacle i-emains an exotentacle throughout the life of the polyp, those appearing early being dis- 

 placed by the later entotentacles. 



ASEXUAL REPRODUCTION. 



By far the majority of corals are colonial in habit, and in nearly all cases the colony is 

 produced as a result of the non-sexual or vegetative reproduction of an original, sexually-formed, 

 simple polyp. It follows that in any study of the Madreporaria much attention must necessarily 

 be given to the various methods of polypal increase other than by larvtv. 



Some few instances are known in which colonies are formed b\' the direct union or aggrega- 

 tion of individuals originally distint't. For these G. von Koch (1890, p. 376) employs the term 

 " AsrsTresrated Colonics," to distina'uish them from colonies produced bv the budding or fission of 

 a single polvp. Von Koch describes such aggregated colonies in the usually simple coral, Balan- 

 ophyU'ia V'/rrucaria, and shows that in all probability they have been produced from originally 

 free and distinct larva} which settled near one another, their skeletons afterwards becoming fused 

 in a common deposit. Lacaze-Duthiers (1899) describes the production of somewhat similar 

 aggregations, ''boquets," in certain specimens of CaryopliyUia obtained from Port Vendres. 

 In the process of fixation of the larva; of Sidermtrxa radian>< I have actually observed such 

 aggregations taking place, and have followed the formation of colonies therefrom as far as the 

 production of the tentacles and earlj' stages in the appearance of the skeleton." 



The foregoing, however, are only exceptional instances of colony formation, and need not lie 

 further noticed in any general consideration of the subject. 



The manner in which polypal increase is brought about, whether by budding, or by fission, 

 or l)oth combined, varies greatly in different species, and is the main clause of the immense variety 

 of form assumed by colonial corals. So far as the resulting types of growth can be studied by 

 observation of the skeleton alone, they have been described by writers such as Dana, Milne 

 Edwai-ds and Hainie, and Duncan, and an extensive terminology has arisen in connection there- 

 with. Dr. A. Ortmann (1890). in his paper "Die Morphologic des Skelettes der Steinkoralleu 

 in Beziehung zur Koloniebildung," has summarized the methods of colony formation in a very 

 detailed manner, and at the close of the account applies the results in part toward a c-lassification 

 of the Madreporaria. Von Koch, in various contributions, has also given nuich attention to 

 the process of coral budding, especially as re\'ealed by means of serial sections of the corallum. 

 As the earliest stages in the production of any new polyp nuist necessarily take place within the 

 soft tissues of the parent, the study of these should yield results of fundamental importance. 



So far as the species of corals under observation are concerned, all the methods of asexual 

 polypal reproduction can be reduced to two tj^pes, which result in very different morphological 



"" Aggregated Colonies in ^ladreporarian Corals." Amer. Nat., vol. xxxvi, 1902. 



