Nature 414, 407 - 409 (2001) © Macmillan Publishers Ltd

HENRY GEE

Henry Gee is a senior editor at Nature.
e-mail: h.gee@nature.com

Some curious fossils from the Cambrian period have been grouped into a new phylum, the Vetulicolia. All of its members are extinct, and their unusual anatomy tempts evolutionary speculation.

Researchers engaged in acts of discovery sometimes have to confront the truly strange and make sense of it. Many of the fossils from the Cambrian Period (about 543–510 million years ago) are cases in point, for they defy comparison with living creatures. In this seeming 'explosion' of life, evolution was toying with many different designs. If all living things share a common ancestor, even the strangest fossil creature must be related to something more intelligible, whether living or long-since dead. Identifying these relatives can be a daunting task.

Such is the problem faced by Shu and colleagues, who on page 419 of this issue1 describe and discuss several fossil animals from the Chengjiang fauna of southern China. This is a remarkable assemblage of exotic forms, comparable with — and slightly older than — the fossils of the famous Burgess Shale of western Canada. Shu et al. describe the exquisite fossils of an unusual, partially segmented creature (see Fig. 1 on page 420) and, in an extended comparison with three other known fossils (Figs 2–5), create a new phylum — a natural group of animals united by a distinctive body plan. They suggest that this phylum, which they call the Vetulicolia, is part of the Deuterostomia, the 'superphyletic' group that includes echinoderms (starfishes, sea urchins and allies), hemichordates (the obscure acorn-worms and pterobranchs) and chordates (vertebrates, including ourselves, as well as tunicates and the amphioxus Branchiostoma). The relationships of these various groups are outlined in Fig. 1.

Figure 1 The deuterostomes, and possible places for vetulicolians in the evolutionary scheme of events. Full legendHigh resolution image and legend (30k)

Interpretation of any problematic fossil is a risky business. Interpreting presumably adult forms of an unusual fossil as deuterostomes is riskier still, because the defining characters of deuterostomes are conventionally embryological and will not show up in adults. However, molecular work suggests there is a close alliance between hemichordates, which have structures called pharyngeal slits, and echinoderms, which do not, at least in living forms2. This conclusion implies that some features seen in adult chordates and hemichordates — in particular the pharyngeal slits — might be characteristics of deuterostomes in general. Shu and colleagues present evidence that vetulicolians, like hemichordates and chordates, have pharyngeal slits; this finding, among others, aligns them with deuterostomes. If this interpretation is correct, vetulicolians represent a new, primitive deuterostome body plan that could shed light on the long-vexed question of vertebrate origins.

The vetulicolian fossils are just a few centimetres long, and are united by several characteristics, including a markedly bipartite body. The anterior half is voluminous and sac-like, with a large opening at the front — regarded as the mouth — and five pairs of lateral openings. Shu and colleagues interpret these features as pharyngeal slits. At least some vetulicolians have a trace of a groove or cleft on the inside surface of the floor of the anterior cavity, which Shu et al. interpret as an endostyle. The endostyle — a characteristic feature of chordates — is a gland-rich gutter in the ventral floor of the pharynx. The mucus it produces lubricates the inside of the pharynx, where it traps and concentrates food particles. It also, incidentally, concentrates iodine. The candidate endostyle of vetulicolians is similar in position to those of tunicates and the amphioxus. The filter-feeding larvae of lampreys are the only vertebrates to retain an endostyle. On metamorphosis, the endostyle becomes the adult thyroid gland. Neither echinoderms nor hemichordates are thought to have an endostyle, which appears to be a uniquely chordate feature.

The posterior half of the vetulicolians is divided into seven segments. A structure interpreted as a gut runs along the entire length of the posterior section, which looks remarkably arthropod-like. Indeed, the authors remark that the entire animal looks much like a small shrimp. But vetulicolians appear to have been limbless, and no arthropod exists that has traded all its limbs for pharyngeal slits.

Where do the vetulicolians fit into deuterostome evolution? The authors' assertion that these creatures were 'basal deuterostomes' — that is, deuterostomes that branched off the main lineage before the echinoderms, hemichordates and chordates became distinct — is properly cautious, given the difficulties inherent in interpreting unusual fossils. To assert that vetulicolians were more or less close to specific deuterostome groups might be asking more of the evidence than it can stand. External commentators, however, like court jesters, have licence to be more reckless.

Looking at the distribution of features in vetulicolians, and mapping them onto a deuterostome phylogeny (Fig. 1), you could make a case that these animals are akin to chordates, rather than deuterostomes in general. This would then make vetulicolians directly relevant to the question of vertebrate ancestry. In short, the latest common ancestor of chordates would have looked rather like a vetulicolian. The only real difference would be that vetulicolians lack a notochord — the axial, stiffening rod that all chordates have at some stage in their life cycle. The amphioxus has a notochord throughout life; tunicates lose it as adults, and in most vertebrates it is supplanted by the vertebral column. Shu and colleagues do not find compelling evidence for a notochord in vetulicolians.

Nonetheless, the vetulicolian body plan is close to what could be regarded as archetypal for the most primitive chordates. In the early 1970s, A. S. Romer3 speculated that the vertebrate body was an amalgam of two separate entities: the 'visceral' and the 'somatic'. The visceral animal corresponds with the internal organs, musculature (characteristically smooth) and associated innervation; the somatic part includes the skeleton, the musculature of the body wall (generally striated), the central nervous system and the sense organs.

Romer used this 'dual-animal' model as a device to explain the course of vertebrate evolution. He argued that primitive chordates, such as tunicates, are all viscera — little more than a sac-like pharynx and gut with the most rudimentary neuronal apparatus. Vertebrates evolved by developing the somatic part, originally as the organ of locomotion in the posterior part of the animal — a beginning can be seen in the locomotory tail of tunicate tadpole larvae. In vertebrates, the somatic part has grown forwards and dorsally, covering and finally encapsulating the visceral part.

Perhaps the most striking feature of the vetulicolians is the strong division of the animal into anterior and posterior halves, just as Romer speculated. To a mind attuned to prospective vertebrate ancestors, it is easy to look at a vetulicolian and see — for example — a tunicate tadpole larva.

References

1. Shu, D. -G. et al. Nature 414, 419-424 (2001). | Article | PubMed |
2. Halanych, K. M. Mol. Phylogenet. Evol. 4, 72-76 (1995). | Article | PubMed |
3. Romer, A. S. Evol. Biol. 6, 121-156 (1972).