494 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



it is doubtful if the next step in septal development is yet fully understood. Milne Edwards 

 and Haime (1857) first attempted with any degree of thoroughness to determine the law govern- 

 ing the septal sequence of corals, and gave expression to their general results in the well-known 

 figure on PI. A5 of the Atlas accompanying the " Historic." Their investigations, however, were 

 conducted almost exclusively upon adult coralla, the determining factors being the comparative 

 sizes and radial extent of the different septa. The relative size and extent of the septum were 

 conceived to indicate the order of appearance; the largest, most developed septa, were the oldest 

 or first formed, the smallest were the last formed. 



Prof. G. von Koch, in the course of his wide and thorough studies of coral morphology, has 

 investigated the laws governing the order of appearance of sei)ta more fully than any other 

 writer. His results are largely founded upon the order of appearance and relative magnitude of 

 the septa in serial sections of fully developed coralla, a method far more likely to yield reliable 

 results than an examination of only the superficial characters of the calice. He concludes that in 

 the main the law^ of Milne Edwards and Haime expresses the actual facts of the case — a new sep- 

 tum alwavs appears l)etween two older septa, and as a rule a perfect cycle is present before the 

 septa of the next cycle arise. 



Unfortunately, thore is no account available of the order of development of the first meta- 

 septa in their relation to the mesenteries, and yet it is only upon this relationship that the iirob- 

 lem can be satisfactorily solved. Both Lacaze-Duthiers and von Koch describe such early stages 

 in the formation of the corallum, but in no case is the relationship of the mesenteries indicated. 

 One of the facts frecjuently emphasized in the present investigations is that the formation 

 of the septa follows very closely upon the appearance of the mesenteries. In ordinary cases, 

 whatever be the numljcr of mesenterial pairs present, an entoscptum and an exoseptum are 

 associated with each. This is clearly shown in the serial sections of the young bud of 

 Solenastrcea (PI. XII); the polyps of Axtraiujln (fig. 47), rhijlJangia, and Cladocom (fig. -ilt) are 

 also very instructive in this respect. The correspondence in the number of mesenteries and 

 septa at all stages can l)e established with the greatest certainty; if any mesenterial cycle is 

 incompletely developed so is the septsil. In Jl/i/ilcina (tig. 13;^) and Pectinia septa appear as a 

 rule only within the entocoelic chambers, but one corresponds with even the youngest of the 

 mesenterial pairs. 



Transverse sections of Manicina sometimes reveal septal invaginations within the exocoelic 

 chambers, as in fig. 13a, although by far the majority are entosepta. On following the sections 

 toward the distal region, however, mesenterial pairs are usually encountered which correspond 

 with the apparent exoceelic septa seen below. Thus tlie latter are really entosepta which in their 

 downward growth exceed that of the mesenterial pair within wliose interspace they occur. 

 H. V. Wilson (1888) states that the septa of Manlchui may for a time be exoccelic, but it is 

 doubtful if higher sections would not have revealed the corresponding mesenteries. Bourne 

 (1893) also found a few exocoelic septal invaginations in Fuiujid, though only entosepta are the 

 rule; here, again, it is likely that the more distal regions of the polyp would have revealed the 

 corresponding mesentei'ial pairs. 



From an examination of a large number of stages in the growth of various species, it becomes 

 manifest that the seipience of the septa beyond the protoseptal stage follows very closely that of 

 the mesenteric succession. Knowing then the order of appearance of the mesenteries, that of 

 the septa can be determined also. The metacnemes have been shown to appear, not a cycle at a 

 time, but in successive Ifilateral pairs fi-om one aspect of the polyp to the other, and the septa 

 must follow a like se([uence. Although in the end the mesenteries and septa of any one cycle 

 become practically equal in size, yet the early stages render it evident that adult size docs not 

 conform with order of appearance. 



A difficulty arises, however, in connection with the exosepta. It has been established 

 throughout that the exosepta, like the cxotentacles, always constitute the last or outermost 

 cycle; the entosepta form all the internal cycles, the sum of the entosepta corresponding with 

 that of the exosepta. But at all stages in the development of most corals, from the protoseptal 

 stage onwards, exosepta occur. The important question therefore arises, whether the exosepta 



