REPORT ON CORALS— DEEP-SEA MADREPORARIA. 197 



in outline, with their walls rising to a height of about 3 or 4 mm. above the basal mass. 

 Septa often irregular, owing to the constant fission of the calicles, but in the rounded 

 calicles disposed with more or less regularity in six systems and four cycles, with some 

 septa of a fifth cycle. Quaternary septa prominent and much larger than the tertiaries, 

 which lie far back on the sides of the fossa, sometimes with their inner margins confluent 

 in the deeper parts of the fossa over the margins of the tertiaries. Fossa moderately 

 deep, with a spongy mass at its bottom not at all prominent representing the columella. 



Tins species differs from Heteropsammia michelini in no point of importance excepting 

 in being compound. Heteropsammia michelini in many specimens shows a tendency in 

 the calicles to divide into two, and Heteropsammia cochlea (Spengler), from Ceylon, 

 bears two calicles, sometimes three imperfect ones. I have examined specimens in the 

 British Museum of this species, and it obviously forms a simple stepping-stone between 

 Heteropsammia michelini and the present form. It would be absurd to place the present 

 form and Heteropsammia cochlea in a genus apart from Heteropsammia michelini 

 because of their compound nature. The characters of MM. Milne-Edwards and Haimes' 

 genus must be modified to include compound forms as well as simple. I can find no 

 evidence in the adult coralla of the present species of any remains of a spiral shell within 

 the basal mass. If a shell were originally present, as is quite possible, it has become 

 entirely absorbed. The walls of the spiral chamber occupied by the Sipunculid are com- 

 posed of bare hard coral tissue. In decalcified spirit specimens, moreover, no trace of 

 any membranous tissue was seen corresponding with the spiral cavity. The Sipunculid 

 was left hanging in a simple spiral cavity excavated within the spongy coenenchymal 

 mass of soft tissue. Semper, in his account of the Heterocijathi and Heteropsammias 

 of the Philippines, comes to the same conclusion as to the absence of a spiral shell in 

 Heteropsammia and all species of Heterocyathus, excepting Ileterocyathus parasiticus. 

 He says, " the Sipunculid lives always only in a cavity formed by itself in the base of the 

 coral." 1 Semper described in the same paper two new species of Heteropsammia obtained 

 by him off the Philippine coast, but curiously enough seems to have met with no com- 

 pound specimens. He remarks on the interest of Verrill's Heteropsammia geminata 

 from Burmah (Amer. Jour, of Science and Arts, vol. xlix., 1870), which has two calicles. 

 Milne-Edwards' two genera Heteropsammia and Stephanoseris are clearly most closely 

 allied, and should be merged into one. The presence of synapticulse in the latter genus is 

 a matter of small importance. 



Extreme length of the base of the largest specimen, 20 mm. Extreme breadth of the 

 base, 15 mm. Extreme height of the corallum, 16 mm. Diameter of one of the nearly 

 circular calicles, 8 mm. 



Numerous specimens dredged. Off Samboangan, Mindanao Island, Philippine 

 Islands. 10 fathoms. 



1 C. Semper, Ueber Generatioiisweclisel bei Steinkorallen, &c, Z. fiir Wiss. Zoologie, Bd. xxii. 1S72, s. 255, u. 264. 



