GENERAL REMARKS, 9 
surface of the corallum may be generally taken as a guide to the density within. In some 
species, however, the thin wall of young corallites is fenestrated—that is to say, it consists of 
longitudinal thicker bands connected together by a delicate and perforate lamina; the thicker 
bands may be considered costez, as they project beyond the intervening parts. In later 
development the wall increases in thickness, and the costs become spinose ; still later, the 
cost become lost altogether, the surface-layer gradually becomes denser and is simply 
clothed with spinules. In such species the costulate and simply echinulate conditions coexist 
in different parts of the same colony, and the denser surface of the older corallites does not 
correspond with the primitive density of the earlier-formed layers of the wall. As growth 
proceeds, however, a further deposition of carbonate of lime takes place beneath the surface of 
the corallum, so that a branch which is quite porous near its apex may become nearly solid 
towards the base. 
Before leaving this subject, it appears desirable to discuss shortly the application of the 
term costee in the case of Madrepora. The porous corallite-wall is essentially composed of 
synapticule, and is therefore not a theca, as it differs both in structure and origm. In 
thecate Madreporaria, and, indeed, in non-thecate types also, it is usual to apply the term 
“cost” to the longitudinal ridges which mark the outer surface of the wall. In the 
majority of cases the costz correspond in number and position with the septa, and are usually 
regarded as the distal extremities of septa which pass beyond the thecate wall. If G. v. Koch’s 
theory of the origin of a theca is correct, and considerable evidence has already been collected 
in its favour, the costz should morphologically be considered as the distal parts of septa. In 
the genus Madrepora the so-called coste undoubtedly do not come under this category. 
They bear no regular relation to the septa either in number or position, and in point of 
time appear before the septa and are also present on the walls of corallites in which never 
more than two septa occur. Under these circumstances it does not appear desirable to 
continue the use of one term for two perfectly distinct structures, and im the systematic 
descriptions I have endeavoured to indicate the condition of the wall by the use of such terms 
as fenestrate, striate, striato-echinulate, rugose, &c. 
It is often stated as characteristic of Madrepora that the directive septa (ef. p. 11) are 
more prominent than the other primary septa; but this is by no means always the case, nor 
is that condition confined to the genus. In Madrepora one frequently finds that in radial 
corallites the outer directive septum is broad and the other five primaries narrow and equal. 
In other species the primary septa of both axial and radial corallites are of equal breadth, in 
which case the bilateral arrangement of parts in the polyp is not indicated by the relative 
importance of the directive septa. Again, in some specimens the six primary septa of the 
axial corallites may all meet together in the axial fossa and fuse together so as to form a false 
columella; in other specimens only the directive septa become confluent. In Anacropora and 
Montipora the relative importance of the primary septa is subject to similar variations to 
those which occur in Madrepora. In Anacropora the corallites are prominent, and the 
branches resemble those of Madrepora so closely that it is not until the absence of an axial 
corallite is observed that the generic distinction is realized. In this genus the directive septa 
(e 
