The metamerism of the amniotes agrees in all essential points with that of the three lower classes of vertebrates we have considered; but it varies considerably in detail, in consequence of cenogenetic disturbances that are due in the first place (like the degeneration of the cœlom-pouches) to the large development of the food-yelk. As the pressure of this seems to force the two middle layers together from the start, and as the solid structure of the mesoderm apparently belies the original hollow character of the sacs, the two sections of the mesoderm, which are at that time divided by the lateral fold—the dorsal segment-plates and ventral side-plates—have the appearance at first of solid layers of cells (Figs. 94–97). And when the articulation of the somites begins in the sole-shaped embryonic shield, and a couple of protovertebræ are developed in succession, constantly increasing in number towards the rear, these cube-shaped somites (formerly called protovertebræ, or primitive vertebræ) have the appearance of solid dice, made up of mesodermic cells (Fig. 93). Nevertheless, there is for a time a ventral cavity, or provertebral cavity, even in these solid “protovertebræ” (Fig. 143 _uwh_). This vesicular condition of the provertebra is of the greatest phylogenetic interest; we must, according to the cœlom theory, regard it as an hereditary reproduction of the hollow dorsal somites of the amphioxus (Figs. 156–160) and the lower vertebrates (Fig. 161–163). This rudimentary “provertebral cavity” has no physiological significance whatever in the amniote-embryo; it soon disappears, being filled up with cells of the muscular plate. Fig.167. Head of a shark embryo. Fig. 167—Head of a shark embryo (_Pristiurus_), one-third of an inch long, magnified. (From _Parker._) Seen from the ventral side.