PROF. P. M. DUNCAN ON A NEW GENUS OF MADREPORAKTA. 25 



The genus was placed by M.-Edwards and J. Haime in the family 

 Astraeidse, and in a group in which growth took place by means of 

 fissiparity. 



The Hesozoic species described by M. de Promentel and by myself 

 enter the genus thus diagnosed, and the only important modification 

 which I made in revising the genus was to introduce the necessary 

 statement that increase took place by gemmation as well as by 

 fissiparous division of the coraliites*. But the genus was included 

 by me in an alliance, the Goniastrseoid, of several genera, which is 

 characterized by the forms having the coraliites united by their 

 walls more or less completely, and without ccenenchyma serial 

 growth not occurring. 



A very perfect siliceous specimen of the Maryland species gave the 

 following characters : — 



The specimen is generally perfect in its details and once formed a 

 portion of a very large branching colony and was undoubtedly a reef- 

 builder. The fracture across the stem resembles those which occur 

 in large specimens of the genus Madrepora at the present day 

 during violent storms. The outside of the coral is covered with 

 calices in a very perfect condition, and the delicate granular orna- 

 mentation of the septa and of the top of a dome of endotheca which 

 fills up the axis and calicular fossa is still to be seen. An excellent 

 natural section of the coraliites in the axis of the stem has been the 

 result of fracture. The section shows that the axial or parent 

 coraliites have undergone some diminution in the bulk of their walls 

 and septa, and probably this happened during life, for corresponding 

 absorption is seen in many recent forms. On comparing the super- 

 ficial calices with the sections of their parent axial coraliites, very 

 considerable differences will be noticed, and it is evident, after 

 careful examination of this specimen, that had the section alone been 

 present, the description of the details presented by it would not 

 have enabled any palaeontologist to give an accurate diagnosis of the 

 species. On the other hand, were the structures seen on examining 

 the superficial calices to be entirely relied on, mistakes regarding 

 the nature of the endothecal structures and of the dimensions of the 

 columella might have been recorded. 



The lower endothecal dissepiments are rather stout and horizontal, 

 and are well seen in the axial coraliites : and the upper, which are 

 numerous and close near the calices, form perfect oblique or domed 

 floors between the septa, so as to shut out the interseptal loculi 

 beneath them from the surface. These uppermost dissepiments come 

 up to about the same level at the bottom of the calice, and reach 

 up to within a fractional part of a millimetre from the free edges 

 of the septa and columella (PL III. fig. 7). Hence the septa seem 

 to rest upon the upper combined dissepiments and to resemble fine 

 linear growths ; two long ones are often, but not invariably, con- 

 tinuous with the columella and reach across the calice, apparently, 

 but not truly, as one long septum. The study of the section of a 



* " Revision of the Families and Genera of Madreporaria." Journ. Linn. Soc, 

 ZooL vol. xviii. 1884, p. 103, 



