This typical articulation of the two coelom-sacs begins very early in the lancelet, before they are yet severed from the primitive gut, so that at first each segment-cavity (us) still communicates by a narrow opening with the gut, like an intestinal gland. But this opening soon closes by complete severance, proceeding regularly backwards. The closed segments then extend more, so that their upper half grows upwards like a fold between the ectoderm (ak) and neural tube (n), and the lower half between the ectoderm and alimentary canal (ch; Figure 1.82 d, left half of the figure). Afterwards the two halves completely separate, a lateral longitudinal fold cutting between them (mk, right half of Figure 1.82). The dorsal segments (sd) provide the muscles of the trunk the whole length of the body (1.159): this cavity afterwards disappears. On the other hand, the ventral parts give rise, from their uppermost section, to the pronephridia or primitive-kidney canals, and from the lower to the segmental rudiments of the sexual glands or gonads. The partitions of the muscular dorsal pieces (myotomes) remain, and determine the permanent articulation of the vertebrate organism. But the partitions of the large ventral pieces (gonotomes) become thinner, and afterwards disappear in part, so that their cavities run together to form the metacoel, or the simple permanent body-cavity. The articulation proceeds in substantially the same way in the other vertebrates, the craniota, starting from the coelom-pouches. But whereas in the former case there is first a transverse division of the coelom-sacs (by vertical folds) and then the dorso-ventral division, the procedure is reversed in the craniota; in their case each of the long coelom-pouches first divides into a dorsal (primitive segment plates) and a ventral (lateral plates) section by a lateral longitudinal fold. Only the former are then broken up into primitive segments by the subsequent vertical folds; while the latter (segmented for a time in the amphioxus) remain undivided, and, by the divergence of their parietal and visceral plates, form a body-cavity that is unified from the first. In this case, again, it is clear that we must regard the features of the younger craniota as cenogenetically modified processes that can be traced palingenetically to the older acrania. We have an interesting intermediate stage between the acrania and the fishes in these and many other respects in the cyclostoma (the hag and the lamprey, cf. Chapter 2.21). (FIGURE 1.163. Frontal (or horizontal-longitudinal) section of a triton-embryo with three pairs of primitive segments. ch chorda, us primitive segments, ush their cavity, ak horn plate.)