It only remains to add that there is sometimes a further complication. If the rock be very porous and permeable by water, it may happen that the original shell is entirely dissolved away, leaving the interior cast loose, like the kernel of a nut, within the case formed by the exterior cast. Or it may happen that subsequent to the attainment of this state of things, the space thus left vacant between the interior and exterior cast--the space, that is, formerly occupied by the shell itself--may be filled up by some foreign mineral deposited there by the infiltration of water. In this last case the splitting open of the rock would reveal an interior cast, an exterior cast, and finally a body which would have the exact form of the original shell, but which would be really a much later formation, and which would not exhibit under the microscope the minute structure of shell. [Illustration: Fig. 2.--Microscopic section of the silicified wood of a Conifer (_Sequoia_) cut in the long direction of the fibres. Post-tertiary? Colorado. (Original.)] [Illustration: Footnote: Fig. 3.--Microscopic section of the wood of the common Larch (_Abies larix_), cut in the long direction of the fibres. In both the fresh and the fossil wood (fig. 2) are seen the discs characteristic of coniferous wood. (Original.)] In the third class of cases we have fossils which present with the greatest accuracy the external form, and even sometimes the internal minute structure, of the original organic body, but which, nevertheless, are not themselves truly organic, but have been formed by a "replacement" of the particles of the primitive organism by some mineral substance. The most elegant example of this is afforded by fossil wood which has been "silicified" or converted into flint (_silex_). In such cases we have fossil wood which presents the rings of growth and fibrous structure of recent wood, and which under the microscope exhibits the minutest vessels which characterise ligneous tissue, together with the even more minute markings of the vessels (fig. 2). The whole, however, instead of being composed of the original carbonaceous matter of the wood, is now converted into flint. The only explanation that can be given of this by no means rare phenomenon, is that the wood must have undergone a slow process of decay in water charged with silica or flint in solution. As each successive particle of wood was removed by decay, its place was taken by a particle of flint deposited from the surrounding water, till ultimately the entire wood was silicified. The process, therefore, resembles what would take place if we were to pull down a house built of brick by successive bricks, replacing each brick as removed by a piece of stone of precisely the same size and form. The result of this would be that the house would retain its primitive size, shape, and outline, but it would finally have been converted from a house of brick into a house of stone. Many other fossils