Madreporarian Genus Astreopora. 279 
The Cenenchyma.—The most characteristic feature of the 
genus is afforded by the coenenchyma. It is constructed of 
two elements—the costey which, instead of being lamellate, 
break up into long tapering echinule (fig. 6). The echinulz 
all bend upwards and are united with one another by regular, 
nearly horizontal floors, for the support of which the echinula 
act as pillars. These are thick near the polyp-cavities, but 
thin halfway between them. The synapticular floors are, at 
least in the uppermost layers, perforated by large round holes. 
The different appearances presented by the coralla is largely 
due to the respective developments of these two elements. 
When, as in cases of rapid growth in thickness, the costal 
elements are most developed, the surface is highly echinulate. 
The synapticular floors, on the other hand, are then often 
feebly developed. Again, in cases of explanate growth, the 
echinulz may be feebly developed and the horizontal floors 
strongly developed. Dana proposed to divide the genus into 
two groups (which might “constitute different genera’) 
according to this distinction. But no such division would be 
admissible; the structure is essentially the same. 
Other variations in the coenenchyma are due to irregulari- 
ties in the growth of the two elements, so that the texture 
may here and there appear confusedly reticulate, or foliate, 
or (by the thickening and nodulation of the elements) 
granular. 
On the Relation of Astreopora to Turbinaria.—In my 
recent paper in this Magazine* on the latter genus I 
suggested that it was possible that Astrwopora represented 
a more primitive coenenchymatous coral than either MJadre- 
pora or Turbinaria. Madrepora could perhaps be deduced 
from a simple incrusting form like Astreopora by the 
special growth and lateral budding of some of the coral- 
lites, while, again, Turbinaria might be deduced from Madre- 
pora by the further specialization of this lateral budding. 
The following are the chief reasons which make me now 
think it highly improbable that Astra@opora has any close 
affinity with Turbinaria:— 
(a) The epitheca in Astrwopora persists and plays an 
important part in the development of the corallum. As we 
have seen, the earliest buds develop round the parent polyp 
contained within a common saucer-like epitheca. This 
epitheca follows and more or less supports the edges of the 
expanding corallum. 
* Ser. 6, vol. xv. June 1895, p. 499. 
