Lower Cambrian Vendobionts from China and Early Diploblast Evolution

See allHide authors and affiliations

Science  05 May 2006:
Vol. 312, Issue 5774, pp. 731-734
DOI: 10.1126/science.1124565


Ediacaran assemblages immediately predate the Cambrian explosion of metazoans and should have played a crucial role in this radiation. Their wider relationships, however, have remained refractory and difficult to integrate with early metazoan phylogeny. Here, we describe a frondlike fossil, Stromatoveris (S. psygmoglena sp. nov.), from the Lower Cambrian Chengjiang Lagerstätte (Yunnan, China) that is strikingly similar to Ediacaran vendobionts. The exquisite preservation reveals closely spaced branches, probably ciliated, that appear to represent precursors of the diagnostic comb rows of ctenophores. Therefore, this finding has important implications for the early evolution of this phylum and related diploblasts, some of which independently evolved a frondose habit.

Ediacaran assemblages represent Earth's earliest complex macroscopic organisms in the history of life (1). They lack skeletal hard parts but are relatively diverse (1, 2), have a defined community structure (3), and show a global distribution (2). Yet their phylogenetic interpretation remains highly controversial. Radical reassignments to lichens (4) and fungi (5) have won little support (1, 6), but the traditional assignments to animals (7) remain problematic. A few taxa can be compared, with varying degrees of reliability, to known animal groups, including sponges (8), cnidarians (9), mollusks (10), and arthropods (11). Most, however, remain in phylogenetic limbo because comparisons to either extant phyla or putative stem groups are frustratingly imprecise. The provocative vendobiont hypothesis (1214) seeks to unify disparate taxa, including forms otherwise assigned to groups as separate as cnidarians and arthropods, on the basis of a distinctive body plan with a modular quilted construction and possibly syncytial tissue. If correct, this hypothesis has two major implications. First, did at least some vendobionts derive independently from protistans, or are they a sister group of either animals (15) or even diploblasts? Crucial in this respect are various frondlike fossils, some of which have been compared to the cnidarians, specifically the pennatulaceans (7, 16), whereas others are clearly akin to other vendobionts (12, 17).

Here, we describe eight specimens of Stromatoveris psygmoglena gen. sp. nov., (18) a frondose fossil from the Lower Cambrian Chengjiang Lagerstätte near Kunming, Yunnan, and interpret them as a new vendobiont (Figs. 1 and 2). Five of them, including the holotype, were collected near Meishucun, whereas the other three are from Jianshan, near Haikou. Specimens show preservation typical of other soft-bodied fossils from Chengjiang and therefore provide exceptional morphological detail. Presumably they were rapidly buried by storm events, and most are oriented at a shallow angle to the bedding plane. The split between part and counterpart is therefore oblique, necessitating composite reconstructions of each specimen (Figs. 1, C to E, and 2, B and C).

Fig. 1.

The Cambrian vendobiont S. psygmoglena, gen.sp.nov. (A, B, and G) Holotype, ELI-Vend-05-001. (A) Upper surface, (B) fragment (counterpart) of lower surface, and (G) enlargement of the boxed area in (A). (C and D) ELI-Vend-05-002. (C) Composite photograph of upper and lower surfaces and (D) distal part of upper surface removed to reveal lower surface. (E) ELI-Vend-05-003, composite photo of part and counterpart to show both upper and lower surfaces. (F) ELI-Vend-05-004, composite photograph of part and counterpart to show upper surface and axial rod located between upper and lower surfaces.

Fig. 2.

Camera lucida drawings of S. psygmoglena, gen. sp. nov. (compare with Fig. 1). The interpretative drawings are composite, with the counterpart reversed and superimposed on the part. Correspondences are as follows: (A) and (E) correspond to Fig. 1, A, B, and G [(E) is enlargement of boxed area in (A)]; (B), to Fig. 1, C and D; (C), to Fig. 1E; and (D), to Fig. 1F.

The body is foliate, with a bluntly terminating stalk that lacks obvious attachment structures (Figs. 1; A, C, D, and F; and 2; A, B, and D). Body length ranges between 2.5 and 7.5 cm. Orientations were adopted for convenience and imply no direct homologies with other organisms. Upper and lower surfaces are markedly different. The former bears prominent branches with quite pronounced relief. Most specimens are too incomplete to count the precise number of branches [typical width was circa (ca.) 1 mm], but in the holotype about 15 are visible on either side of the midline. The latter is defined by a relatively narrow groove (Figs. 1A and 2, A and E). The most proximal branches arise along an adaxially inclined line, but all branches are parallel and fan distally at a shallow angle to the midline. Branches are mostly single, but occasionally they bifurcate or fuse in an irregular fashion (Figs. 1A and 2, A and E). In some specimens, branches were filled with sediment, suggesting that during life they were hollow tubes. The lateral margins of the branches show occasional short prongs or spurs. These may represent interconnections either between the branches or possibly into the interior of the frond. Branch numbers appear to have shown only a modest increase with body size, and differences in relative width in larger specimens suggest that the individual branches continued to grow. Distally, however, new branches can be seen to differentiate from the surface of the blade (Figs. 1A and 2, A and E).

The branches were evidently firmly attached to the body, but some branches show slight imbrication. The branches bear transverse, closely spaced striations (Figs. 1, A and G, and 2, A and B, and fig. S1, A and D), but there is no evidence for zooids. The regions between the branches were narrow and recessed, and in some specimens there are associated dark strands (Fig. 1; A, C, and E to G; and fig. S1; A, B, F, and G). These may have been discrete structures, possibly canal-like and presumably located in the body wall. They have irregular margins (Fig. 1E), which may have been extensions connecting to other regions of the body wall. Distal to the branches the surface of the frond is smooth, although a continuation of the midline occurs as a narrow meandering groove (Figs. 1A and 2A). This region was quite elongate (Figs. 1F and 2D) and in the living animal probably acutely tapered. Basal to the branches the midline area rapidly widens to define a more or less smooth stalk.

The lower region broadly appears to be divisible into two regions. More distally an ovoid central smooth area is strongly concave (Figs. 1D and 2B), but adjacent there are subdued ridges that run slightly oblique to the body margin and in the opposite direction to the branches on the upper surface. In addition, there is a diffuse ornamentation, roughly ovoid. More basally there is subdued ribbing and a prominent leopard-skin ornamentation, at least near the margin (Figs. 1B and 2A). The lower surface becomes smooth toward the basal region. One specimen (fig. S1, C and E) shows striking arcuate structures on the stem. Although these might be another type of surface ornamentation, they are interpreted as body-wall support, possibly collagenous.

The interior contains a substantial sediment fill (Figs. 1, C to E, and 2, C and D, and fig. S1B), more pronounced in the stalk, suggesting its cross section was approximately circular, whereas the blade was somewhat compressed. As with the branches, this sediment infill suggests the interior was largely hollow and possibly fluid-filled during the organism's life. An axial structure (ca. 0.4 mm across), with apparently ferruginous preservation, is located nearer to the upper surface. In one specimen (Figs. 1F and 2D), it has been partially excavated, and toward the base a similar structure occurs, but it runs transversely and possibly was rotated to its present position as a result of decay.

The organism was benthic and embedded in the seafloor by the stalk. Whether it lived upright or recumbent is less clear. The apparent rigidity of the axial structure would suggest the former orientation, whereas the subdued morphology of the lower surface, particularly the ovoid strongly concave smooth area (Figs. 1D and 2B), could be consistent with a recumbent mode. Mode of feeding is conjectural and depends in part on phylogenetic comparisons. Given the absence of definitive zooids, one possibility is that the branches were ciliated and served to transport food particles via the narrow grooves between the branches. Assuming a density of cilia comparable to typical suspension feeders, this would provide a highly effective trap. What appear to be interconnections between the branches (Figs. 1; A, E, and G; and 2; A and D) suggest an exchange system that presumably also connected to the interior. During the organism's life, the interior was presumably filled with fluid or gelatinous tissue. The central axis (Figs. 1F and 2D), analogous to the axial rod of pennatulaceans, presumably provided additional support, but no other internal organs are discernible.

Although it cannot be placed in any known genus, in overall form Stromatoveris is similar to a number of Ediacaran frondlike fossils. It is most similar to the otherwise poorly documented Khatyspytia (19), but the latter is more slender and has a larger number of shorter branches. General resemblances also exist with the fronds Vaizitsinia (19), Charniodiscus (16, 20), Glaessnerina (16), and, more remotely, such forms as Charnia (19, 20). The phylogenetic position of these Ediacaran fronds (and by implication Stromatoveris) is controversial (21), with opposing views favoring cnidarians, especially pennatulaceans (7, 16), or vendobionts (12, 14). Key points in these differing interpretations include attachment of the branches to the frond and absence of unequivocal evidence for zooids, both of which are inconsistent with the pennatulacean hypothesis. The axial structure has an obvious counterpart in pennatulaceans, but on functional grounds this could be convergent. Stromatoveris also has some similarities to the mid-Cambrian Thaumaptilon (22), a possible Ediacaran survivor. This taxon was provisionally identified as a pennatulacean, in part on the basis of zooids. The discovery of a more convincing pennatulacean from the Chengjiang Lagerstätte (fig. S2) suggests that if Thaumaptilon is a cnidarian (assuming the zooids are correctly identified), then it is more primitive than the anthozoans (Fig. 3). In any event, Thaumaptilon and Stromatoveris are unlikely to be closely related. The latter taxon lacks obvious zooids and has a markedly different branching pattern. Branching in Stromatoveris also shows various irregularities [perhaps consistent with a less constrained morphogenetic program (21)] and has possible interbranch connections, and most importantly the distal branches differentiate from within the upper surface. The latter arrangement is unlike growth in pennatulaceans (or other colonial metazoans). Such a style of growth, although inferred in many frondose vendobionts, may bear reexamination, especially because at least some taxa show other distinctive characteristics, including a striking fractal growth (17).

Fig. 3.

Outline of metazoan phylogeny, showing proposed position of Stromatoveris and Thaumaptilon (22) as primitive ctenophores and cnidarians, respectively, so implying convergent evolution of a frondlike habit. Metazoan phylogeny is still in a state of flux, but here sponges are taken to be basal, with the calcareans possibly a sister group of all other metazoans (28). The position of the placozoans is controversial, but here they are treated as primitive diploblasts (29), evolving before the invention of nerve cells (30). Ctenophores are monophyletic (31) and are taken to be the sister group of cnidarians plus triploblasts (28). As argued in the text, ctenophores were primitively frondlike (vendobionts) before acquiring a globular body with separate comb rows that eventually were used in a pelagic existence. Although ctenophores have a biradial symmetry, this has a unique rotational element and may be derived and effectively unrelated to the biradial symmetry that may be primitive to cnidarians. Cnidarians are also monophyletic and are divided into anthozoans and medusozoans (32). Although previously ctenophores have been argued to be the sister group of all bilaterians, it is now widely accepted that cnidarians are the sister group (33). The triploblasts are composed of deuterostomes and protostomes.

The level of organization seen in Stromatoveris (and equivalent Ediacaran fossils) seems to transcend protistan complexity. It seems likely, therefore, that the vendobionts as currently recognized (12, 14) are not monophyletic. Taxa such as Ernietta and Pteridinium, built on simple modular units and apparently with an infaunal mode of life, may well be giant protistans (14). The frondlike fossils, however, are interpreted as metazoans, specifically diploblasts (Fig. 3). The pronounced disparity within the diploblasts, notably between cnidarians and ctenophores, has made their early evolution highly speculative. Dzik (23), however, has hypothesized a link between Ediacaran fronds and Cambrian ctenophores. Although it is difficult to accommodate, for instance, taxa such as Dickinsonia and Thaumaptilon in this scheme, the fine transverse structures seen on the branches of Stromatoveris are similar to those seen in Cambrian ctenophores (Fig. 3) despite their otherwise disparate body plans. In Stromatoveris, the ciliated branches are closely spaced and attached to the frond. In Cambrian ctenophores, the branches became separated and the body more globular. Both were probably benthic, using the ciliated rows for suspension feeding, whereas in contrast extant ctenophores are highly derived. This evolutionary transition is marked by a shift to a pelagic existence, acquisition of a gelatinous body plan, and co-opting of the ciliary rows from feeding to locomotion.

Stromatoveris joins a select group of Ediacaran survivors (2225). In comparison to those Cambrian survivors showing typical Ediacaran-like preservation (24, 25), the material shows new features unobservable in the coarser host matrix. Quality of preservation matches that of Thaumaptilon (22), but as noted the similarities between the two taxa are evidently convergent (Fig. 3). The possible example of an Ediacaran survivor, reported earlier from Chengjiang (26), has no similarity to the example described here, and its wider relationships are uncertain.

Supporting Online Material

Figs. S1 and S2

References and Notes

View Abstract

Navigate This Article