Fig. 3

A, B Alternative ways to organize and innervate body musculature, modified from Ruppert’s Figs. 6–8 (see [9]). A represents a vermiform ancestral coelomate that moves (or burrows) by means of propagating waves of contraction that travel along the body. The muscles responsible (m1) are an intrinsic part of the outer mesothelium of the coelom (c), and are innervated by intraepithelial nerves (representative nerve fibers are shown in section as small open circles). B The situation in chordate somites, where the myotomal muscles (m2) lie along the inner surface of the coelom. To innervate these requires a fold or invagination of the neurogenic epithelium to bring the nerves (arrows) within contact range of the muscles. In effect this produces a rudimentary nerve cord that can then be shallow or deep as required. The region indicated by asterisks is a continuation of the coelomic cavity, as in chordates the muscles extend into the coelom in this way. My assumption is, that if Pikaia is a chordate, its musculature would necessarily be arranged in a similar fashion, along the inner surface of the coelom. Tissues are color-coded here and in subsequent figures: neural tissues and non-neural ectoderm in blue, mesoderm in pink (for epithelia and connective tissue) and red (for muscles), endoderm in yellow. C The specific example Ruppert uses to illustrate his point: the collar cord (cc) of an enteropneust, modified from Fig. 3 in [9]. The main coelomic cavity (c) belongs to the mesocoel, while the perihaemal coeloms (phc) are diverticula arising from the metacoel that project forward on either side of the medial blood vessel (bv). As explained in the text, the perihaemal coeloms are then ideally positioned to serve as models for the first somites