54 MADREPOKAEIA. 



C'alicles (lig. 'J) deep, subcircular, 2 '5-3 "5 mm. across. Walls nearly vertical, fenestrated, 

 composed of a light delicately filamentous reticulum, ragged seen sideways, full of meshes 

 seen from above, and with fine septal spikes projecting here and there from the margin, though 

 a symmetrical radial arrangement is not visible either in the strise on the top of the wall, or 

 in projecting ridges round its edges. A little way down within the calicle the septa become 

 visible as tliin rows of delicate filamentous teeth wliich eventually meet, interlace, and, 

 putting out lateral projections, form a large, exquisitely delicate, filamentous columellar tangle. 

 Some 6-l<^ sejita are early j^rononnced, Init they as a rule slope down with ragged edges on to 

 the columella, without the formation of any rosette of pali, although theii- teeth may show a 

 tendency to bend up. Only in the lateral shallow calicles does this prominence of certain 

 septa result in the formation of the rosette. In these lateral calicles the walls are thick and 

 so delicately and ii-regularly reticular as to be almost woolly, the septa are in complete 

 numbers, and jagged and pointed, and the columellar tangle is large and the pali rise as large 

 irregular portions of its own jagged reticulum. 



The vertical section shows highly perforated septa, open reticular walls, and delicate 

 tabula; (see PI. III. fig. 2). 



There are three specimens grouped together which have characters running into one 

 another so perfectly that it is necessary to class them together in spite of their diti'erences. 

 Fig. 9, PI. II. and fig. 2, PI. III. relate to specimen a; fig. 1, PI. III. is from specimen h. 



The differences between a. and h are accidental. Specimen h had died down at the top 

 (? killed by some foreign body), but fi'om all round the dead patch the colony is creeping in to 

 cover it over. Some of the new calicles bending round on films of epitheca can be seen to be 

 continuous vdt\\ those which bordered the dead patch, having apparently just escaped the fate 

 of those more centrally placed. There is a sort of slight cushion-shaped growth all round the 

 dead patch, and in the deep calicles of this growth the septa are seen to be much coarser and 

 stouter than in a (see 1*1. III. fig. 1), and the same is the case with tlie wall reticulum. These 

 calicles in fact approach the deep fuanel-shaped type with tliick sloping septa and obscure 

 columella. But the dead calicles which were being covered up had thin septa and delicate 

 reticular walls and columellar tangle as in a (PI. II. fig. 9). ^Vnd the lateral calicles are also 

 like those of a. Here then vm have distinct evidence of structural variation being due to an 

 accidental departure from the normal method of growth. 



The thu'd specimen, c, is a very young colony some 3 cm. in diameter, but showing the 

 edges of still earlier growtlis, the smallest of which was little more than 1 cm. in diameter. 

 This specimen seems only to differ in the reticulum being not quite so delicate and woolly. 

 The skeletal elements are more disposed to be knobbed, and in the lateral calicles the pali are 

 rather more solid. 



In all three the septal apparatus has essentially the same chai'acter, consisting of rows of 

 fine filamentous teeth which put out lateral pi-ocesses ; these interlace to form a large delicate 

 columellar tangle. The rosette formation is confined entirely to the lateral calicles. 



The woolly character of the walls, which can only be seen by a lens, is well shown in tlie 

 magnified fig. 9, PI. II. 



a. Warrior Ileef. Coll. W. Saville-Kent. Zool. Dept. 92. 12. 1. 148. 



b. " Great Barrier Eeef." Coll. W. Saville-Kent, Zool. Dept. 92. 12. 1. rm. 



r. " Torres Strait." " Coll. A. C. Iladdon. ZnoL Dept. 1902. 9. 9. 5. 



