lime and granular silica; so are the spaces formerly occupied by the walls and exotheca.

Variety 3.-The walls remain, and consist of saccharoid calcareosilica; but the rest of the sclerenchyma is deficient. The interspaces are empty, and are encroached upon by the saccharoid mineral. Many corallites of the Barbadian Astræans have their walls and cœnenchyma turned into saccharoid-looking spar, all the rest being deficient. These honeycombed-looking masses consist of semicrystalline carbonate of lime, with much silica; and doubtless much of the destruction of details was prefossil, for the species has very delicate septa and costæ*.

8. Calcareo-siliceous casts. Variety 1.-The sclerenchyma is deficient for some lines' depth from the calicular surface, but is present elsewhere, and consists of carbonate of lime, with more or less silica. The interspaces are filled with light reddish-brown silica for some depth below the former calicular margin; but below the spots where the sclerenchyma begins to be seen they are filled with dense carbonate of lime, which is of a light-brown or light-red colour.

Variety 2.-A few septa are all that remains of the sclerenchyma. The interspaces are filled with dense carbonate of lime: this salt fills up the reticulations in the walls of perforate corals, and is more or less mixed with silica, and varies greatly in colour.

Variety 3.-The columella and the septal ends attached, remain alone on the free surface of the corallum; below, all the sclerenchyma exists. The interseptal loculi are filled with saccharoid carbonate of lime and silica.

The largest casts of calices as yet known are those of the Astræans of the Marl of Antigua. The interspaces are seen as dense lightbrown silica; the septa, calicular walls, and costa are wanting for some lines' depth+; and even in transverse sections, made a little distance below the surface, the filled interseptal loculi are distinguished from the vacant spaces where septa have once been. The vacant spaces of the septa, in the casts, simulate interseptal loculi; and the dense silica of these, which is free from carbonate of lime, may be mistaken for septal structure. The siliceous interspaces are not continued down far into the corallum, for carbonate of lime soon exists in great proportion in them; and where the silicification of these interspaces becomes defective, there does the silicification of the septa begin to be intense, and the appearance of a cast then ceases. The intensity of the silicification on the surface is very evident; and that there was infiltration of the interseptal loculi by carbonate of lime and other salts prior to the loss of the septa is so also. It is most remarkable that the whole surface of a large corallum should be covered by casts of its calices, and that the sclerenchyma should be deficient for some depth; and it is difficult of explanation, unless the prefossil condition of the coral be admitted into the consideration of the subject.

The calicular ends of rapidly growing corallites are more delicate, * The Trinidad (St. Croix) corals are generally found thus fossilized. + Quart. Journ. Geol. Soc. vol. xix. pl. xiii. fig. 16.

more permeated by the juices of the polypes, and infinitely more fragile than the deeper parts; they are covered during life by the polype-tissues which determine their nutrition and growth, and these processes are singularly active in luxuriant corals. In slowgrowing corals this calicular fragility is not seen, and the lower parts of a corallite are nearly stony in hardness.

The larger Astræans, the Isastræans, and the Stephanocœnians were, probably, rapid growers, and unfavourable to the development of casts.

In some specimens the prefossil influence of decomposing membrane on tender calices is seen by the fossilization of deformed corallite-ends. It seems reasonable to account for the non-silicification and consequent non-preservation of the superficial hard parts by admitting the destructive influences of a decomposing membrane upon them, their molecular adhesion being slight, and their constitution mainly of carbonate of lime.

The sclerenchyma thus affected might be strong enough to last out the filling up of the interseptal loculi with carbonate of lime, but was not durable enough to withstand the secondary chemical changes which terminated in the replacement of the carbonate of lime in the loculi by silica.

Lower down the corallum, and out of the range of the superficial polype-tissues, the denser sclerenchyma was gradually replaced by a homogeneous dark flint, and lost but little of its details during the process; still a little loss is noticed, especially in prismatic corallites.

To account for this loss, of granules on the septa, of portions of the septa themselves, of endothecal dissepiments, and of more or less of the corallite-wall, we must recognize the influence of the decomposition of the delicate tissue which lines these parts, even deep in the corallum.

The loss of details must have occurred before the silicification of the interspaces (of their carbonate of lime), because the space formerly occupied by the deficient hard parts becomes filled up by an extension of the infiltrated carbonate, which is rarely so perfectly silicified as the rest*.

The most remarkable changes induced by loss of sclerenchyma, and persistence of the mineral infiltrated into the interspaces of a coral, are observed in the case of some perforate corals at Antigua and Jamaica. The corallum resembles that of an Astræan with separate corallites, and dentate costa in regular series +; really the dentations are casts of the regular spaces between the network of the original corallite, whose walls and tissue have been lost. These casts are not perfectly siliceous, but contain much carbonate of lime; so that the destruction of the sclerenchyma, which developes a cast, occurs before silicification is complete. The later Tertiary corals are constantly found without their septa, or with their columellæ and part of their septa attached, the corallite-wall having been worn away: this is produced by prefossil wear and tear, as is also the loss of the details of some dendroid Astræans in the Jamaican limestone. * Quart. Journ. Geol. Soc. vol. xix. pl. xiv. fig. 7. t Op. cit. pl. xiv. fig. 4c.

It is singular, and as yet not explained, that the epitheca or general envelope of many, but not all, of the genera is wanting in the fossil specimens. Doubtless many of the siliceous masses in which corallites can be detected have been rolled since their fossilization, but there are no evidences of epitheca in very perfect Astræans which have not been rolled. Finally, it may be taken as a rule that the most perfect replacement of the carbonate of lime of the corals by silica is found in the oldest beds.

§ V. CONCLUSION.

I believe that there is no truth in the hypothesis which asserts that the silicification of the West Indian fossil corals was the result of a volcanic outburst, which poured siliceous solutions over the depressed reefs of the Miocene age. The corals of the Inclined Beds and of the Chert of Antigua are silicified, so are the Woods and Mollusca; but the lower beds of the Marl contain both perfectly silicified Astræans and calcareo-siliceous corals also, whilst the upper part of the formation yields fossils in all stages of siliceous and calcareous mineralization. The neighbouring Barbudan limestone has no siliceous corals, yet it is of the same age as the Antiguan Marl. The San Domingan fossils are usually calcareous, but here and there siliceous specimens are met with in the same beds with the others. There is no evidence to prove that coral-growth ever ceased from the Miocene time to the present in the Caribbean Sea, although there has been a change in the facies of the fauna. These facts and remarks are very antagonistic to the hypothesis, and rather tend to prove that the silicification of the corals has been a slow process, which has had no other origin than in those chemical operations which are still in action, and that their greater or less intensity in certain favourable localities has produced siliceous fossils amongst those affected by the calcareous form of mineralization. Wherever a highly aërated sea, containing silica in solution, acts on calcareous fossils at considerable depth, and, therefore, under considerable pressure, there would appear to result a chemical transposition; and the presence of crystals of quartz, of homogeneous flint, of the hydrates of silica, and their coloured varieties is due to chemical influences which bear a relation to the geological changes in and about the reefs. It must be remembered that there is a small amount of silica in corals; and it will be observed in microscopic sections of Antiguan corals, in which silicification is incomplete, that the silica is usually deposited in molecules in the centre of the calcareous mass, and not on its superficies. The process of siliceous transposition is doubtless very slow; it is not always perfect, for the silica would appear to have an affinity for bodies formerly organized, and often to destroy the former tissues. Thus there is abundant carbonate of lime in which transposition is not going on around the siliceous fossils, and the persistence of some animal or vegetable organized tissue, decomposing more or less slowly, appears to determine the presence of certain forms of silica; moreover, the details of corals are often so

destroyed by the deposition of homogeneous black flint, that this destruction must have commenced after the coral was imbedded.

In the course of this communication facts and hypotheses have been determined and propounded, which have led to the following conclusions:

:

1. Silica, whether homogeneous, granular, or crystalline, does not appear to be deposited at first in the interspaces of corals, but replaces a salt of lime which was infiltrated partly in solution or partly in a state of mechanical suspension in a compound fluid. The replacement does not commence until the salt of lime has acquired a certain density, and it occurs first of all in the central parts of the loculi in the form of granular points*.

The combinations of the silica with water, the production of its hydrates, and the evolution of coloured varieties would appear to be determined by the presence of decomposing animal and vegetable matters, and of salts of iron, at the time of the entombment of the corallum in the bed of the detritus of the reef.

2. The opal lines, the opaque lines in agates, the broad wavy milky clouds in otherwise transparent flint, and the porcellanous opaque silica are found to have a very close relation with the form of silicification, in which much destruction of the hard parts of the corallum is noticed.

3. Every grade of a silicification which destroys the textures of corals and reduces them, at last, to pure homogeneous black flint, like that of the upper Chalk, is to be distinguished. Moreover, crystalline quartz, chalcedony, and every variety of transparent or opaque silica constantly replace the carbonate of lime or the usual coral-salts of part or of the whole of a corallum, without preserving the structural details.

These destructive siliceous fossilizations, not peculiar to the West Indian beds alone, form one of the causes of the deficiency of the geological record.

MAY 11, 1864.

Thomas Carrington, jun., Esq., Chesterfield, Derby; J. B. Even, Esq., C.E., Mem. Soc. Géol. Fr., 36 Rue Montagu, Brussels; The Rev. John Henry Timins, M.A., of Trinity College, Cambridge, Vicar of West Malling, Kent; and Henry Woodward, Esq., F.Z.S., of the British Museum, and 144 Leighton Road, N.W., were elected Fellows.

The following communications were read:

1. On a SECTION with MAMMALIAN REMAINS near THAME.

By T. CODRINGTON, Esq., F.G.S.

IN excavating a cutting near Thame, on the Railway between Thame and Oxford, a section remarkable in itself and yielding a large

* The phenomena of orbicular silex in the silicified corals will form the subject of a future communication.

Fig. 2.-Section with Mammalian remains, near Thame.

quantity of Mammalian remains has been exposed during the past summer.

Externally the undulation of the ground cut through by the Railway does not differ from that around it. It is detached on all sides, the Thame river and two small tributaries nearly surrounding it. The highest part, which is not far from the cutting, is not more than 25 feet above the water in the nearest stream, and from it the ground falls gradually with a rounded outline.

Fig. 1.-Plan showing the position of the section with Mammalian remains, near Thame.

The position of the cutting with relation to the existing streams is shown in the plan (fig. 1). The section (fig. 2) has the vertical heights exaggerated twice.

Resting on the Kimmeridge Clay is a bed of coarse gravel averaging 2 feet in thickness, abutting against a bank of undisturbed clay on the west, and gradually dipping eastward. The materials of the gravel are-angular chalk-flints (about 50 per cent. of the whole), quartz, hornstone, and ironstone, in pebbles or fragments more or less rounded, rolled lumps of chalk, and Tertiary pebbles. The order in which they are here mentioned is that of their relative abundance. Blocks of Sarsen or grey-wether sandstone, in size from half a cubic yard downwards, occur in the gravel.

The pebbles of quartz, and the hornstone and ironstone, may have been derived from the neighbouring Lower Greensand beds; but a fragment of mica-schist, which I obtained from the gravel, seems to point to the Northern Drift as the source of at least some of the materials.