Motile larvae figure prominently in a number of past scenarios for chordate and vertebrate origins, notably in the writings of Garstang, Berrill, and Romer. All three focus on the motile larva of a primitively sessile tunicate ancestor as a vertebrate progenitor; Garstang went further in deriving chordates themselves by neoteny from a yet more ancient larva of the dipleurula type. Yet the molecular evidence currently available shows convincingly that the part of the tunicate larva that persists to the adult expresses only a subset of the genes required to specify a complete bilaterian body axis, and essentially the same appears to be true of dipleurula larvae. Specifically, both are essentially heads without trunks. Hence, both are highly derived and as such are probably poor models for any real ancestor. A more convincing case can be made for a sequence of ancestral forms that throughout their evolution were active, motile organisms expressing a full complement of axial patterning genes. This implies a basal, ancestral form resembling modern enteropneusts, although a pelagic organism at a hemichordate level of complexity is also possible. A reassessment is thus required of the role played by adult and larval tunicates, and of larvae more generally, in chordate evolution. Tunicates need to be interpreted with caution, since the extreme degree of modification in the adult may have been accompanied by reductions to the larva. Dipleurula larvae may retain some ancestral features (e.g., of apical, oral, and anal organization), but are otherwise probably too specialized to be central players in chordate evolution. Garstang nevertheless remains a key figure in the history of evolutionary thought for his innovative ideas on the relation between ontogeny and phylogeny, and the way in which major innovations in morphology and body plan arise.
Molluscan dorsal-ventral patterning relying on BMP2/4 and Chordin provides insights into spiralian development and bilaterian body plan evolution