How ancient is the adult swimming capacity in the lineage leading to Euchordates?

Current understanding of animal phylogeny has provided new insight on the evolution of the ancestors of the Euchordata (cephalochordates1vertebrates). Modern, adult euchordates possess a capacity for swimming, displacing themselves through water independently of currents. Swimming is achieved by the presence of a fin and axial body movements. However, the classic hypothesis on the origin of euchordates prescribes that their adult ancestors were sessile, and that axial swimming was only present in larvae that developed into sessile adults by metamorphosis, as in ascidian urochordates (Fig. 1). This implied that the free-living and swimming euchordate adults were paedomorphic (Garstang 1928). This hypothesis was compelling for several reasons. First, sessile adult lifestyles are present within every major group of deuterostomes (ascidian urochordates, pterobranch hemichordates, and crinoid echinoderms), raising the possibility that this was the primitive lifestyle of deuterostomes. In fact, sessile pterobranch hemichordates have been classically suggested to be representative of early deuterostomes. Furthermore, pterobranchs possess a motile, ciliated bilateral larva, similar to that of echinoderms. By modification of a similar larva, the notochord and dorsal nerve chord have been hypothesized to have evolved to facilitate habitat selection for the sessile adult, via an actively swimming stage (Fig. 1, Berrill 1955). Second, it is likely that early chordates were ciliary mucus suspension feeders, as revealed by cephalochordates and the ammocoete larva of lampreys, a fact apparently more consistent with suspension-feeding in sessile lifestyles, rather than with free-living active predation. Third, lophophorates are sessile suspension-feeding invertebrates with motile ciliated larvae and a remarkable combination of traits found both in derived deuterostomes and protostomes. It was reasonable to consider that lophophorates could be representative of the lifestyle of the most recent common ancestors shared by the deuterostomes and protostomes, suggesting that even more remote ancestors of the chordates had also been sessile (Fig. 1). Thus, a chordate ancestor with a motile larva and a sessile adult seemed an inescapable conclusion. It is therefore not surprising that the hypothesis of paedomorphosis has been repeatedly mentioned in college textbooks (Romer and Parsons 1977; Carrol 1987; Benton 2000). However, suspension feeding and motile ciliated larvae can also be observed in burrowing, swimming or other nonsessile adult lifestyles, and in a strict sense cannot be considered as evidence for sessile ancestors. Furthermore, new molecular evidence indicates that the phylogenetic positions of ascidia, pterobranchia, lophophora, or echinodermata do not provide support for the presence of a sessile adult or of a process of paedomorphosis in the evolution of chordate ancestors (Wada and Satoh 1994; Wada 1998; Cameron et al. 2000). According to these analyses, chordates are not nested within any of these groups. Moreover, burrowing enteropneusts, rather than sessile pterobranchs, were found to be basal within hemichordates (Cameron et al. 2000), and appendicularians with an adult capacity for swimming, rather than the sessile ascidians, are basal within urochordata (Wada and Satoh 1994; Wada 1998). Thus, appendicularia, rather than ascidia, and enteropneusts, rather than pterobranchs, should be considered more closely representative of early urochordates and hemichordates, respectively. Importantly, according to molecular evidence, hemichordates and echinoderms do not represent successive outgroups to chordates, but rather, they make up a new group (hemichordata1echinodermata) that is the sister group of chordates (Turbeville et al. 1994). This implies that urochordates and euchordates are as relevant as echinoderms and hemichordates in order to infer the lifestyle of the most recent common ancestor of all deuterostomes. Considering this molecular evidence, the corrected inference is that adult early chordates were free-living ciliary mucus suspension-feeders, with axial swimming and a fin with iterated rays. Furthermore, early deuterostomes did not have a sessile adult lifestyle (Wada 1998; Cameron et al. 2000), EVOLUTION & DEVELOPMENT 7:3, 171–174 (2005)