FAMILIES AWD GENERA 0"P THE MABBBPORABIA. 137 



established a new generic diagnosis. Later on Etallon, Eeuss, 

 nd myself decided that the genus must enter the Eungida ; and 

 since that time it has been placed in thoPoritidse by Milaschewitsch 

 and E. Tomes, and replaced by the latter in the Fungida. The 

 careful researches of E. Pratz enable me to place the genus in 

 Idle transitional group between the Astrseidse and Pungida — in 

 the family Plesiofungidse. 



Milaschewitsch, noticing the perforate condition of the septa 

 of Thamnastra^ans from Nattheini, was led away to believe that 

 this condition was invariable in the genus. Moreover, he thought 

 that a perforate septum necessitated a Porosa alliance. There 

 are perforate septa in Gyphastrcea, an aporose form, and solid 

 septa in most species of Madrepora, a perforate genus. Per- 

 foration may be noticed in some English Thamnastrseans that have 

 been much weathered, but in parts o£ the colony only. The con- 

 dition is not universal in specimens from Nattheim, and is the 

 result of destructive siliceous fossilization and weathering. 



The trabeculse of the septa do not always leave spaces between 

 them like lattice-work ; such a condition is local and exceptional. 

 At the same time the thinnest part of the septum is between the 

 nodular masses of the trabecules. There is no doubt that 

 the septa are united by growths which are not exaggerated 

 granules, and which are synapticula. These not only inter- 

 fere with the open condition of the interseptal loculi, but make 

 the position occupied by any thing like a mesenteric fold very 

 shallow. The existence of endotheca is decided, and especially 

 in some Oolitic species. It was noticed by Milne-Edwards and 

 Jules Haime, and figured in my Monog. Brit. Poss. Corals, 

 Pal. Soc. 1872, Pt. iii. pi. 4. fig. 9, and has been recognized in 

 continental forms. 



The presence of a wall bounding the corallites has been debated. 

 I do not find a true wall, but that synapticula are often placed 

 in vertical series and act as a wall, which, however, is not com- 

 plete. Often no trace of such a limiting series can be found. 

 On the other hand, the existence of a wall has been stated by 

 De Promentel. The columella is variable in the genus, and 

 has been used to establish subgenera. Finally, the colony is 

 sometimes convex, plane, and nodulose, or decidedly branched. 

 The structures in these forms are identical, and therefore the old 

 plan of making the branched forms iuto a new genus or subgenus 

 has been abolished. 



