06 
secreted calcareous matter around themselves, and thus formed these root- 
like tubes.” 1 In all probability these radiciform processes are hollow and 
tubular throughout the whole of the Cyathophylloidea, and not exothecal. 
Lindstrom has also demonstrated the presence of such anchoring filaments 
connected with the interior of the calicos in some of the Operculata, such as 
llhizophyllum gotlandicum, H. elongcitum, and possibly in Goniophyllum ; in 
11. dong at um these tubes are tabulate, 2 and in 11. attenuatum become 
actually stoloniferous, giving rise to new corallites. 3 
In some cylindrical or conical corals these radiciform processes are 
developed on all sides of the corallite, but in others only on one; in the 
latter case this probably represents the underside, or that to which the 
corallite is inclined. 
8. Fistula ?. — These are present in two of the three sub-fasciculate 
species and one of the ob-conical forms, forming a prominent feature in T. 
Lonsclald, even more so than in its two varieties, and are present in all parts 
of the corallum, but placed at irregular distances from one another. It is in 
T. congregationis that these connecting processes are seen in their fullest 
development at regular stated intervals between corallites, tier above tier. 
The young corallites, or buds, in T. princeps are united by fistulae — a very 
interesting point in the morphology of these corals. 
The processes are more or less horizontal, usually swollen at their 
points of projection from the corallites, diminishing in diameter in their 
median course. They are tubular, and, like the radiciform processes, arc not 
only prolongations of the visceral cavities of the corallites they unite, but 
arc partially filled with concentric stereoplasma. In T. Lonsdalei the septal 
spines continue quite into the mouths of the fistulae, and in one instance a 
spine was observed in the open tube course (PI. XXVI, fig. 7). 
The presence of fistulae in Tryplasma recalls the structure of the 
Eridophyllum group on the one hand, and the systematically distant 
Syringopora on the other. There is, however, this difference between the 
fistulae of Tryplasma and the processes connecting the corallites in 
Eridophyllum — that, whilst the fistulae in the former are projected indis- 
criminately from all sides of a corallite, there does seem to be a marked pro- 
trusion of processes from one side only of each corallite in a given corallum 
1 Lindstrom — Geol. Mag., 1866, III, p. 406. 
2 Lindstrom — Bihang K. Sv. Vet. -Aka 3. Handl., 1882, VII, No. 4, t. 2, f. 15 and 16. 
3 Lindstrom — Bihang K. Sv. Vet.-Akad. Handl., 1682, VII, No. 4 , t. 3, f. 17. 
