382 CORALS and Paramuricea). Though not strictly epithecal, we may also consider in this connection the remarkable root-like processes (" radici- form prolongations") which are so characteristic of many of the Rugosa (e.g. Omphyma, Goniophyllum, Rhizophyllum, Pholidophyl- lum, Ptychophyllum, &c.) These processes may attain a length of several inches, and they consist of a process of the epitheca and wall, inclosing a diverticulum of the visceral chamber, and in some cases subdivided by dissepiments. In the simple species (such as Omphyma) they serve to attach the corallum to foreign objects ; but similar processes occur in various compound forms (Eridophyllum, &c.), and serve to unite the various corallites with one another, being thus of the same nature and function as the tubular connect ing processes of the Tabulate genus Syringopora. The epitheca of the Rugose corals is usually marked with numerous fine encircling striae, and with longitudinal lines, grooves, or ridges. The latter, however, are usually regarded as not being true costcc. They differ, in fact, from tue costae of the Zoantharia scleroder- mata in not being placed opposite the septa, but alternating with these structures, and thus corresponding with the interseptal loculi. Not uncommonly (as in many species of Zaphrentis, in Strep- telasma, &c.) there are two of these pseudo-costae, which are pre-emin ently developed and run along the dorsal surface of the corallum, the other ridges con verging to wards thiscentral pair in a pinnate manner. This appearance seems to be due to the fact that the growth of the cor- rallum was effected by the intercalation of successive septa along the sides of the primary septum to which this principal pair of pseudo- costae corresponds. In Pholidophyllum the costee are double, and are Covered by double longitudinal rows of small imbricated scales of an epitheca] nature. In genera in which the wall is absent and the corallites are united together by the confluence of the septa (e.g. in Pachyphyllum, Smithia, Phillipsastrcea), it is impossible to deter mine whether true costae are present or not. True costae are also certainly present in the genus Holocystis. The calice in the Rugosa varies greatly in form, shape, depth, &c. It is usually more or less circular, or oval, sometimes semicircular or subtriangular in outline (Rhizophyllum.), sometimes quadri lateral (Goniophyllum). It may be very shallow, or very deep, and its edges may be completely everted (e.g. in Ptychophyllum pattllatum). It may be at right angles to the axis of the corallum, or it may have any degree of obliquity. In the simple and curved or horn -shaped coralla, it is very common for the calice to be very oblique to the axis of the coral ; and in these cases the convex side of the corallum (dorsal side) is the longest, and the concave side (ventral side) is the shortest. The septa of the Rugosa are substantially similar to those of the Zoantharia sclerodermata in form and structure, forming, when well developed, a series of vertical calcareous laminae which are primitively double, spring from the inner surface of the wall, and radiate towards the centre of the visceral chamber. They are rarely or never perforate, and they vary much in the extent of their development. Sometimes they are almost wholly aborted, being only recognizable as so many faint striae on the inner surface of the calice (as in the typical examples of Cystiphyllum). At other times, they extend inwards only an extremely short distance from the wall, as in the genus Amplexus. In other cases, again, they are well developed towards the centre of the corallum, but have no connection with the outer wall, from which they are separated by dissepimental vesicles, as in the genus Lonsdaleia. In other instances, finally, they have their normal arrangement, being attached externally to the wall, and extending inwards to, or near to, the centre of the visceral chamber. In rare cases the eepta may be all of nearly the same size ; more commonly they are markedly different in size. Most of the forms in which the septa are well developed, show a distinct tetrameral arrangement of the septa, though it does not appear possible to assert positively that the primitive and first developed cycle of septa consists of only four elements. In many cases, however, the septa of the adult are a multiple of four, and their quadripartite disposition may be plainly manifested by the fact that four of the septa are pre eminently developed and form a conspicuous cross (Stauria), or by the presence of four calicine depressions which have a similar cruciform arrangement (Omphyma). In the typical Rugosa the septa, though undoubtedly not simultaneously developed, are never theless of only two sizes, a larger and a smaller, alternating regularly with one another ; and they cannot, therefore, be distin guished according to their dimensions into a series of regular cycles. The small or secondary septa also may be occasionally absent. The primary or larger septa, be their development what it may, are for the most part equal in size ; but, notwithstanding this fact, corallum often shows a very distinct bilateral symmetry, the due apparently to the primitive tetrameral disposition of the septa. 1 his is especially shown in the general existen je either of a single septum of larger size than all the others (one of the four primitive septa), or of an extraordinary vacant spa, ;, representing this septum, and known as the " septal fossula." The septal ibssula usually presents itself as a more or less conspicuous depression or groove m the calice, and its position, though apparently constant in any given form, is variable, being sometimes on the convex side of the corallum in the simple forms, sometimes on the concave side, and rarely situated laterally. In general it is a simple space or deficiency caused by the absence or abortion of one of the four primary septa, and it is seen in transverse sections to be occupied by from one to three short septa. Sometimes it is accompanied by a tubular depression of the tabulae at that point. Sometimes there are two smaller lateral fossulae, directed at right angles to the main depression, and representing two others of the primitive septa. At other times there may be four shallow fossulse arranged in a crucial manner (Omphyma), but it is not certain that these correspond with the four primitive septa. Lastly, in the genus Meiriophyllun^ the septa are arranged in four groups, which are separated from one another by vacant spaces or fossuke, though in this case also it is not certain that these spaces are homologous with the true septal fossulae of forms like Zaphrentis. The precise physiological import of the fossula is uncertain ; but its presence gives rise to an irregu larity in the arrangement of the septa which is highly characteristic of the Rugosa. Whilst in Stauria all four of the primitive septa are pre-eminently developed, only three are thus predominant in Anisopyllum, and only one in Hallia. The free odges of the septa, where they appear in the calice, are in general plain and smooth, but they may bear granular tubercles (Palceoajdus), or teeth (Zaphrentis cornicula, Heliophyllum). The sides of the septa are likewise generally smooth, but they may be granulated (Palcco- cyclus), or they may be adorned with arched and ascending striae (Heliophyllum) . The axis of the corallum is very often occupied by a columella, which varies much in structure in different cases. In forms such as Cyathaxonia, Lithostrotion, Koninckophyllum, &c., we have a proper or essential columella, which is developed independently f the septa, occupies the centre of the visceral chamber, and projects as a solid rod into the floor of the calice. In other cases the colu mella is composed of twisted lamellae, which inosculate with one another so as to give rise to a vesicular axis, as in Lonsdaleia and Axophyllum. In other cases, a false columella may be produced by the twisting together of the inner edges of some of the primary septa (as in some species of Cyathophyllum). In other cases, finally, the axis of the visceral chamber may be occupied by a series of more or less complicated structures, which may occupy a consi derable space, and which have sometimes been regarded as repre senting a kind of columella (Clisiophyllum, Dibunophytlum, Rhodo- phyllum, Cyclophyllum, &c). These axial structures, however, can only m a very limited sense be regarded as columellar. The continuity of the interseptal loculi in the Rugosa. is generally more or less interfered with by the development of endothecal dis sepiments ; but in no case are synapticulae present. The dissepi ments vary greatly in character and amount. Sometimes they are wholly wanting (CyathaxonidcK and some species of Amplexus) ; at other times they may "be present in small amount (Lophophyllum,. sornes species of Amplexus, &c.) ; at other times they are exceed ingly abundant. In the species of Cyathophyllum, in Litho strotion, in Koninckophyllum, and in other forms, the dissepiments are so largely developed towards the circumference of the visceral chamber as to give rise to a dense peripheral zone of vesicular tissue. In longitudinal sections of the corallum this vesicular tissue is seen to be composed of very minute lenticular cells arranged in olilique rows directed upwards and outwards. In other forms, such as Campophyllum, Lonsdaleia, &c., a similar zone of vesicular tissue exists, but the cells which enter into its composition are of very large size. In the genus Heliophyllum, and in certain other forms, there are found singular dissepimental structures, which arc attached to the sides of the septa, but do not extend completely across the interseptal loculi. The structures in question constitute a series of plates which are attached by their bases to the sides of the septa, projecting freely into the interseptal loculi, and directed inwards and upwards in an arched manner from the interior of the wall towards the centre of the visceral chamber. These arched ridges are placed at corresponding points on the opposite side of each septum ; they consequently appear on the free edges of the septa within the calice as so many spines, and they communicate to cross-sections of the septa a characteristic cross-barred appear ance. In the Cystiphyllidce, again, the entire visceral chamber is filled with a vesicular tissue of convex and inclined cells, which may be regarded as formed partly by dissepiments and partly by tabulae. Tabulce are in general well developed amongst the Rugose corals, and coexist with well developed septa. In some genera (such as Amplexus and Zaphrentis) the tabulae are "complete," that is to say, they pass completely across the visceral chamber from side to side, thus dividing it into a succession of distinct stories, of which only the uppermost is occupied by the living tissues of the animal. In a greater number of cases (e.g. Cyathophyllum, Lithostrotion, Lonsdaleia, Heliophyllum, &c,), the tabulae are "incomplete," that is to say they do not extend across the visceral chamber, but are confined to a larger or smaller central area. The central tabu
late area may or may not be pierced by a columella ; and the septa