The shell-substance, if calcareous, has one of the two types, porcellanous or vitreous, that we have already mentioned, but _Polytrema_, a form of very irregular shape, though freely perforated, is of a lovely pink colour. In the calcareous shells sandy particles may be intercalated, forming a transition to the Arenacea. In these the cement has an organic base associated with calcareous or ferruginous matter; in some, however, the cement is a phosphate of iron. The porcellanous shells are often deep brown by transmitted light. {63}[Illustration: FIG. 13.—Shells of _Foraminifera_. In 3, 4, and 5, _a_ shows the surface view, and _b_ a section; 8_a_ is a diagram of a coiled cell without supplemental skeleton; 8_b_ of a similar form with supplemental skeleton (_s.sk_); and 10 of a form with overlapping whorls; in 11_a_ half the shell is shown in horizontal section; _b_ is a vertical section; _a_, aperture of the shell; 1-15, successive chambers, 1 being always the oldest or initial chamber. (From Parker and Haswell, after other authors.)] Despite the apparent uniformity of the protoplasmic body in this group, the shell is infinitely varied in form. As Carpenter writes, in reference to the Arenacea, "There is nothing more wonderful in nature than the building up of these elaborate and symmetrical structures by mere jelly-specks, presenting no traces {64}whatever of that definite organisation which we are accustomed to regard as necessary to the manifestations of conscious life.... The tests (shells) they construct when highly magnified bear comparison with the most skilful masonry of man. From the same sandy bottom one species picks up the coarsest quartz grains, unites them together with a ferruginous cement, and thus constructs a flask-shaped test, having a short neck and a single large orifice; another picks up the finer grains and puts them together with the same cement into perfectly spherical tests of the most extraordinary finish, perforated with numerous small pores disposed at pretty regular intervals. Another species selects the minutest sand grains and the terminal portions of sponge-spicules, and works them up together—apparently with no cement at all, but by the mere laying of the spicules—into perfect white spheres like homoeopathic globules, each showing a single-fissured orifice. And another, which makes a straight, many-chambered test, the conical mouth of each chamber projecting into the cavity of the next, while forming the walls of its chambers of ordinary sand grains rather loosely held together, shapes the conical mouths of the chambers by firmly cementing together the quartz grains which border it." The structure of the shell is indeed variable. The pylome may be single or represented by a row of holes (_Peneroplis_, _Orbitolites_), or, again, there may be several pylomes (_Calcituba_); and, again, there are in addition numerous scattered pores for the protrusion of pseudopodia