18 MADREPORARIA. 
the horizontal elements are thin and long, in Pl. XVIII. fig. 9 they are thin and short, the 
trabecule in both being well developed. 
These are only a few samples of the possible variations in the character of the skeleton 
due to variations in the different developments of the skeletal elements.* 
The Septa.—The typical formula, with some of its variations, is seen in fig. 3, B to H, 
although these show only the rough ground plan of their different fusions and pali formations, 
and, except in E, leaving out all indications of the septal granules. 
We may, perhaps, say that the septa always fuse, but that the fusions which usually 
take place visibly, that is, high up in the aperture of the calicle, may also occur so deep 
down as to be lost to sight in the columellar tangle. There is one consideration of weight 
which inclines me to this view, as against the suggestion that the condition in which none of 
the septa fuse (fig. 3, A) is primitive. If this were really primitive, the distinction between 
primary and secondary cycles would be more clearly seen : the primaries would project further 
towards the centre than the secondaries. This distinction of the cycles, shown in the diagram, 
is, however, not, or only very slightly, seen, and there is a reason why it should not be if the 
septa are in reality sloping steeply downwards, so that their inner or axial edges are involved 
in the columellar tangle deep down in the base of the fossa. 
Further evidence for this secondary origin of the condition in which all the septa appear 
to be free, is seen in the fact that irregularities in the fusions always appear as the calicles 
deepen—that is, as the septa remain more and more incomplete round the aperture, and only 
project deep down. This is, however, a point to which further attention will have to be paid. 
We shall return to it again below in connection with the pali, and in the Introduction to the 
West Indian Group (Part IT.). 
The most persistent fusions take place between the lateral septa, and are marked by two 
large pali on each side of the directive plane. The dorsal directive is always free, while the 
ventral triplet of septa may either fuse (C, F) or its components remain separate, as shown in 
the diagram (B, D, E, G, H). 
One other condition of the septa may be noted. When they are long and wedge-shaped, 
and slope inwards to a fine point with little or no development of pali, these long septa 
have mostly very granular edges; the granules, including the wall granule, are then 
frequently square, and diminish in size from the wall ridge down to the minute point-like 
granule, or small palus, bordering the edge of the central fossa—see, for instance, Pl. VI. fig. 6, 
while another case of inward sloping with diminution of the pali is seen in Pl. I. fig. 8. 
The Pali.—These have been already referred to in the last paragraph in connection with 
the septa. They arise usually where the septa fuse, though not wholly because of the fusion, 
* Both Dana (Zooph., pl. liii. figs. 7-12) and Mr. Stanley Gardiner (Proc. Zool. Soc., pl. xxiv. 
figs. a—n) have called attention to the different forms and heights of the wall—the former in merely 
the briefest outlines, the latter with more detail. But in neither case was the analysis of the 
fundamental structure far enough advanced at the time to admit of the differences pointed out 
being of practical use in the face of so many bewildering variations. 
