Showing chambers, tubuli, and canals. Compare this and Fig. 49 with Figs. 28 and 29.] [Illustration: Fig. 49.--_Portion of Shell of Calcarina._ Magnified, after Carpenter, (_a_) Cells. (_b_) Original cell-wall with tubuli. (_c_) Supplementary skeleton with canals.] To the Nummulites it also conforms in its tendency to form a supplemental or intermediate skeleton with canals, though the canals themselves in their arrangement more nearly resemble Calcarina, which is represented in Fig. 49. In its superposition of many layers, and in its tendency to a heaped-up or acervuline irregular growth it resembles _Carpenteria_, _Polytrema_ and _Tinoporus_, forms of a different group in so far as shell-structure is concerned. The large and curious sandy Foraminifer from the Pacific dredged by Alexander Agassiz, and named by Goës, _Neusina Agassizi_, may also be mentioned as presenting some points of resemblance.[38] It may thus be regarded as a composite type, combining peculiarities now observed in two groups, or it may be regarded as a representative in the Nummuline series of Polytrema and Tinoporus in the Rotaline series. At the time when Dr. Carpenter stated these affinities, it might be objected that Foraminifera of these families are in the main found in the Modern and Tertiary periods. Dr. Carpenter has since shown that the curious oval Foraminifer called _Fusulina_, found in the coal formation, is in like manner allied to both Nummulites and Rotalines; and still more recently Mr. Brady has discovered a true Nummulite in the Lower Carboniferous of Belgium. This group being now fairly brought down to the Palæozoic, we may hope finally to trace it back to the Primordial, and thus to bring it still nearer to Eozoon in time. [Footnote 38: _Bulletin Mus. Comp. Zoology_, vol. xxiii., No. 5, Dec., 1892.]