318 DR. A. VATJGHAN ON THE FAUNAE SUCCESSION [May I906,, 



uniformly-expanding cone, whereas C. rushiana is, for the greater 

 part of its length, purely cylindrical. 



In many respects C. contorta approaches near to 0. costata, M'Coy. 1 

 Compare PL XXIX, fig. 5. Both have the same form and dimen- 

 sions ; in both the septa are simple and thickened ; both have 

 sparse tabular intersections in the interseptal spaces ; and both 

 have a dense central area. 



The differences are, however, well marked: — C. contorta has- 

 numerous septa (thirty to forty), and the septa are strongly bent. 

 C. costata has a small number (twenty-five or less) of straight 

 septa. In C. costata the epitheca is strongly costated ; in C. contorta 

 the costation is suppressed. C. costata is said to have a solid central 

 axis ; whereas, in C. contorta, the central area is made up of a 

 dense ring surrounding a columellarian lath, the interspace being 

 vesicular. 



Evolution of Cyathaxonia. — It seems possible that the 

 Cyathaxonia? of the upper beds are derived by direct descent from 

 the tubular variant of Zaphrentis cf . Phillipsi, by the strengthenings 

 of the septa and of the central tube. 



The purely-radial type of septation, as exhibited in Cyathaxonia 

 rushiana, probably indicates parallel development with Densiphyllum 

 (see below) due to convergence. The central columellarian lath is 

 also, most probably, an instance of convergence with the Litho- 

 strotions and Clisiophyllids of the Upper Avonian. 



The absence of Cyathaxonia from the South-Western Province 

 may be accounted for, by the fact that Zaphrentis aff. Phillipsi 

 does not there show any tendency towards the production of a 

 central tube. 



Densiphyllctm. (PI. XXIX, fig. 6.) 



This genus is here employed in the sense in which it is inter- 

 preted by Thomson. 2 The corals included here have the cornute 

 form which is common to both Zaphrentis and Cyathaxonia ; they 

 have the septation of Cyathaxonia, but differ from this genus in 

 the absence of a well-defined central axis. 



The most typical group of the genus is not represented in the 

 Rush sequence, and the majority of the specimens collected from 

 these beds may be more correctly termed Densiphylloid Zaphrentes* 

 Densiphylloid Zaphrentes are occasionally met with throughout 

 the sequence ; but the earlier examples are mere variants of 

 Zaphrentis cf. Phillipsi : whereas the upper forms are more 

 specialized, and show relationship with Cyathaxonia costata y 

 M'Coy. The figured specimen from the Cyathaxonia-Beds is of 

 this type, and is a Zaphrentis in which the septation is purely 

 radial and almost devoid of septal break ; the wall is extremely 

 thick, and the septa reach almost to the centre, where their thickened 

 ends are webbed together by small tabular expansions. An 

 example of a different type is noticed under Amplexi-Zaphrentis. 



i ' Brit. Palseoz. Foss.' 1855, p. 109, & pi. iii C, fig. 2. 

 2 Proc. Phil. Soc. Glasgow, vol. xiv (1882-83) p. 445. 



