FORAMINIFERA 381 amples of the group, such as ftonionina (19 in fig. 1), and is scarcely tracenble in the small folt/ytomclla crispa(lGin fig. l),one of the most common of British Foraminifera, though in the large I olysttrmclla of tropical seas the cunal-system presents a regular development not approached elsewhere. This is specially displayed by "inter nal casts " of this shell (fig. 21), that have been fortunately obtained by the deealeification of specimens in which the sarcodic contents of the chambers had been replaced by mineral deposit, so as to give an exact model of the original animal. The Nummuline type is most characteristically represented at the preseut time by the genus Opcrculina, which is so intimately re- Fio. H.Operculina laid open, to show its internal structure: , marginal cord seen in cross section at a ; b, b, external walls of the chambers ; c, c, cavities of the chambers; c , c , their ular prolongations; d, d, septa divided at d , d , and at d", so as to lay open the interseptul canals, the general distribution of which is seen in the septa <?, e the lines radiating from e, e point to the secondary pores ; y, g, non-tubular columns. luted to the true Nummulite by intermediate forms that the two types cannot be separated by any definite boundaries, notwithstand ing the dissimilarity of their most characteristic specimens. This genus is represented in temperate seas by small and feeble examples; but it attains a much higher development in tropical seas, the largest existing Opcrculina, however, only attaining the size of the smaller fossil Nummulitcs. As the organization of Operculina affords the clue, not only to that of the great JS ummulitic series, which, at the commencement of the Tertiary epoch built up an immense mass of limestone that girdled a large part of the globe, but also to that of the far more ancient Eozoon, it will be described in some detail. The shell is a compressed nautiloid spire (fig. 22), the breadth of whose earlier convolutions increases in regular pro gression, but of which the last convolution (in full-grown specimens) usually flattens itself out like that of Pcncroplis (5 in fig. 1). As a rule, the shell-wall of each whorl of the spiral completely invests all the preceding whorls ; and each chamber-cavity (fig. 22, c, c)is continued from tlie margin towards the centre by two "alar pro longations," one on either face of the spire (c , c }, which are filled during life by extensions of the sarcodic segments that occupy the successive chambers. The size and extent of these " alar prolonga tions" differ much in different varieties of Operculina, and often in different parts of the same shell, being, as a rule, least marked in proportion as the spire flattens itself out. There is often no ex ternal indication of them, the successive whorls of the spire being visible from the centre to the margin ; but in vertical sections the spiral lamina of each whorl is still seen to be continued over the whole of the preceding whorl, even to the centre of the spire, though it may adhere to it everywhere except near its margin. The chambers are separated by double septa, which are shown in trans verse section at d, d ; these do not extend internally quite to the margin of the preceding whorl, but leave a fissured aperture, by which each chamber communicates with the one behind and before it, and the last chamber with the exterior, as seen at a ". The "spiral lamina," which encloses the chambers and their "alar prolonga tions," is perforated by minute closely-set tubuli of about rciroth of an inch in diameter; the orifices of these aro seen on the internal wall of the chambers to be disposed with great regularity, as in fig. 23 ; and they seem normally to run parallel to each other (fig. 24), so Fl " 2 3--ItermiI surface of wall of two cham- .,i e bers, o, a, of Nummulite. showing the orifices of as to open with equal re- j ts minute tubuli; 6, 6, the septa containing gularity on the external canals; c, c, extensions of these canals in the surface. But in their intermediate skeleton ; d. d, larger pores, passage from the interior towards the exterior of the shell these tubuli sometimes approach and sometimes diverge from one another. so that the pseudopodia which pass through them issue forth in closely-set bundles, separated by intervening spaces of non- tubular shell. These imperforate portions are harder than the porous shell, and often project as ridges or tubercles, forming a more or less regular "sculpturing" of the surface. They frequently correspond to the septa beneath, which are themselves composed of non-tubular shell-substance, and are overlaid by ridges of the same vitreous imperforate material. Nothing can be more inconstant, however, than this sculpturing, even on different parts of the same shell ; so that it cannot be accounted as of the least value as a specific character. The outer edge of the spire is bounded by a peculiar band, of different structure from the rest, which may be distinguished as the "marginal cord" (fig. 22, a, a). This band, instead of being perforated by minute parallel tubuli, is traversed by a set of larger canals, inosculating with each other, and forming part of the "canal-system," the peculiarity of its character being clearly shown in vertical sections of well-preserved fossil Nnm- mulites (fig. 24, b, b). The " canal-system consists in the first Fia. 24. Vertical section of tubulated chamber-walls, a, a, of j ummtilite; b, b, marginal cord; c, cavity of chamber; d, d, non-tubulated columns. place of a pair of "spiral canals," which run along the edges of the marginal cord, communicating with each other through its system of inosculating passages ; while from each of these there pass off into spaces left between the two layers of each septum (as shown at d , d , fig. 22) a scries of "intcrseptal canals," the dis tribution of which marked out by causing them to imbibe coloured fluid is seen in the septa c, c. The general arrangement of this canal-system will be better understood by reference to that of Pvlys- tomclla shown in its "internal cast" (fig. 21), the position of the "spiral canals," however, as well as the distribution of their branches, being very different. The typical form of the Nummulite is lenticular (fig. 25, is), somc- Fic. 25. A, piece of ffummulitic Limestone from the Pyrenees, shoving Num- mulites laid open by fracture through the median plune; B, vertical section of A ummulitt; c, Orbitoidei. times more nearly approaching the globular, whilst in other cases it is as compressed as that of Opcrculinn,- great differences of this kind presenting themselves between individuals, which, on account of their general conformity of structure, must be accounted specifi cally identical. It is among the compressed forms that the largest size is attained, some of these measuring 4 inches across, while the globose forms seldom much exceed an inch in diameter. The distinction which separates the typical Nummulitcs from the typical Opcrculincc consists in this, that in the former the earlier whorls are completely invested by the alar prolongations of the later, the successive turns of the spiral lamina being separated overtheir whole surface from those internal and external to them )>y intervening spaces, as shown in fig. 26, so that the spire is but little, if at all, discernible on the surface. On the other hand, we never -find the later whorls of true Nummulites flattening themselves out margin ally, like those of the typical Operculinte ; and there even seems rea son to believe that when the spire has attained its full growth, it tends to close itself in. Curious differences are scen.in the course and arrangement of the " alar prolongations" ; and it has been pro posed to group the numerous modifications of the Nummulitic type in accordance with these differences. Thus, in some of these, the alar prolongations extend in a regular radial direction from the margin to the centre of each whorl of the spire, being separated by corresponding radial prolongations of the marginal septa. In other instances, again, they follow a sinuous course, the septa prolonged from those of successive chambers still remaining distinct from one another. From these, again, we pass to a group in which the septa
