128 



The species, as may be seen from the accompanying table, 

 is extremely variable, but its main characters would seem to 

 be as follows : — 



**Corallum conical or wedge-shaped, generally compressed, 

 usually with a distinct scar below, 2 to 7 mm. in length and 

 showing 12 to 24 septa. Sides of the corallum commonly 

 with curved transverse bands, corresponding to intervals of 

 growth, often at the narrow sides opposite the ends of the 

 calicle carried out into small wings. Wings sometimes 

 replaced by hollow root -like processes near the scar, perhaps 

 also with additional roots at the sides; in large specimens 

 wings generally absent. The calicle is elliptical, the top of 

 the long axis not more than i to 3 mm. below that of the 

 short. Relation of axes very variable, about 2 to i. 



" The centre of the calicle a deep fissure, the larger septa 

 •ending almost perpendicularly against it, filled in below by 

 trabeculae in medium-sized specimens from 20 septa and in 

 large from 24. In all free specimens septa of cycles I. to IV. 

 present, cycles V. and VI. depending on the size of the 

 individuals. 



"Height of free corallites from 4-35 mm.; long axes of 

 same from 9-37 mm." 



The septa, as in most or all species of the genus, have 

 radially set fine ledges with spines at intervals, and the 

 larger against the axial fossa are often much broadened at 

 their edges where the trabeculae come off 



Although there are only three specimens in the collection, 

 which appear to be absolutely the same as F. rubruvi, this 

 name has the priority. All the young corallites are of course 

 ■fixed. They break off generally when their calicles have 

 attained a length in their long axes of about 9 mm., the free 

 corallites being about 5 mm. high. Some, however, are 

 attached considerably longer, the largest attaining a height 

 of 23 mm. Three of the speci.nens could not ever have be- 

 come free, the central stalk being still perfect and surrounded 

 on all sides by rootlets, numbering 7, 8 and 8. One of the 

 specimens is attached to a small piece of decaying serpulid 

 tube, and a second to a mere fragment of coral. Both sup- 

 ports seems to have been free, and suggest a possible reason 

 for the throwing down of extra rootlets. The edges of the 

 septa of these specimens further do not show through the 

 epitheca. The calicles of two are scarcely compressed, and 

 these two, taken alone, would undoubtedly have been placed 

 in the former genus Rhizotrochus . Each rootlet communicates 

 with two interseptal spaces on opposite sides of a septum. 

 The latter bisects the rootlet where it joins the large corallite, 

 and thence continues into the rootlet for some distance as a 

 ridge on its lower side. 



