• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Reductio Ad Absurdum: Testing The Evolutionary Relationships Of Ediacaran And Paleozoic Problematic Fossils Using Molecular Divergence Dates

Posted on: Saturday, 31 January 2004, 06:00 CST

ABSTRACT-Many of the late Neoproterozoic "Ediacaran fossils" have been referred to the Cnidaria, often on the basis of vague or poorly known features. However, representatives of the living Chondrophorina (=Porpitidae, Hydrozoa), Pennatulacea (Anthozoa), and Coronatae and/or Stauromedusae (Scyphozoa) have all been identified in Ediacaran biotas, based on specific morphological features preserved in a number of specimens. These three cnidarian groups have plausible Paleozoic representatives as well, but many of their Paleozoic fossils are also somewhat problematic. We test these systematic hypotheses by using them to calibrate divergence dates across the Cnidaria, based on an extensive molecular phylogeny of extant cnidarians. In this reductio ad absurdum approach, if a calibration based on one interpretation of a problematic fossil yields a glaringly inconsistent age for a better-known clade, that interpretation is likely to be mistaken. We find that assuming the existence of Pennatulacea and Scyphozoa in the "Ediacara biota" places the root of the Cnidaria between 800 and 1,000 Ma, a figure which is, at least, not out of line with other molecular clock estimates. However, assuming the existence of the Chondrophorina in the Neoproterozoic, or anywhere in the Paleozoic, pushes the root of the Cnidaria back to between 1,500 and 2,000 Ga, which is considerably older than the oldest previous estimates for the origin of the Cnidaria. We suggest that the likeliest explanation is that chondrophorines were not present in the late Precambrian or Paleozoic. The Ediacaran and Paleozoic fossils previously interpreted as chondrophorines probably represent other taxa.

INTRODUCTION

THE LATE Precambrian and earliest Cambrian soft-bodied fossils of the "Ediacara biota" are well-known paleontological enigmas. Consisting of a variety of structural types ranging from disc- shaped "medusoids" to leaflike "fronds," they have been interpreted as everything from invertebrate animals in or near extant phyla (e.g., Glaessner, 1984; Fedonkin, 1990) to members of other kingdoms, such as protists or lichens (e.g., Zhuravlev, 1993; Retallack, 1994) to members of independently multicellular taxa, not directly related to anything alive (e.g., Seilacher, 1992, 1994). Probably no single explanation will fit all the fossils. While some Ediacaran forms are quite unlike any known animal, others are still most plausibly interpreted as animals in or near extant phyla. These include representatives of the Porifera (Gehling and Rigby, 1996), Mollusca (Fedonkin and Waggoner, 1997), and possibly Arthropoda (Waggoner, 1999) and Annelida or Pogonophora (Sokolov, 1968).

Most Ediacaran fossils have historically been interpreted as cnidarians. The "medusoids," which are the most abundant fossils at most Ediacaran sites, were generally described as jellyfish (e.g., Wade, 1972; Glaessner, 1984). A few "medusoids" show tetraradial symmetry, such as Conomedusites and Stauridinium (e.g., Fedonkin, 1990), which is typical for non-anthozoan cnidarians: the Medusozoa. However, many Ediacaran "medusoids," if not all, are now thought to be benthic (e.g., Bruton, 1991; Gehling, 1991; Jenkins, 1992; Gehling et al., 2000). Some "medusoids" could be cnidarian polyps, and others are probably holdfasts for "frondlike" organisms that could be colonial cnidarians, but working out just what each specific type of "medusoid" was is difficult at this time. Other "medusoids," such as Tribrachidium and Albumares, are triradially symmetrical fossils with relatively complex structure. These triradial fossils cannot easily be placed in any known metazoan group, although they may be of a cnidarian grade of complexity. However, most of the Ediacaran "medusoids" show no defined degree of symmetry. "Medusoid" taxonomy is currently highly problematic due to the extreme taphonomic variability and morphological simplicity of many of the fossils (Gehling el al., 2000).

However, there are three extant cnidarian groups which have been claimed, on well-defined anatomical grounds, to be represented in the Ediacara biotas. (Fig. 1) The first is the Chondrophorina or "sailors-by-the-wind," a subclade of the Hydrozoa, whose extant genera Porpita and Velella are usually classified together in the family Porpitidae. Chondrophorines are disc-shaped pelagic colonial organisms that secrete a gas-filled float composed of multiple concentric chambers. The Ediacaran genera Ovatoscutum, Chondroplon, Kaisalia, Kullingia, and Eoporpita have been referred to the Chondrophorina (Gehling, 1991; Jenkins, 1992). Some of these are poorly known, taxonomically suspect, and/or interpreted as non- chondrophorines (e.g., Hoffman, 1988; Jenkins, 1992; Jensen et al., 2001). However, the Ediacaran genus Ovatoscutum looks very similar to the float of the living chondrophorine Velella. The genus Eoporpita shows extensive tentacles radiating from a central boss, which in some specimens appears chambered.

The second extant clade that may be represented in Ediacaran assemblages is the Pennatulacea or the "sea pens," a clade of colonial anthozoans within the Octocorallia (Anthozoa). Pennatulaceans are composed of one large primary zooid which forms a central stalk or rachis, with multiple secondary feeding zooids branching from the primary zooid. In the most derived pennatulaceans, the secondary polyps are organized into leaflike structures branching from the rachis. Exemplified by living genera such as Ptilosarcus, Pteroeides, Virgularia and Pennatula, this arrangement produces the featherlike appearance that is "typical" of pennatulaceans; however, this shape actually is found in a minority of extant genera. In most pennatulaceans, the secondary polyps branch directly from the rachis, and the colony form may be club- shaped or fan-shaped (Williams, 1992, 1995). A number of frondlike Ediacaran fossils have been linked with the Pennatulacea, notably Charnia, Charniodiscus, Rangea, and Glaessneria; these have a leaflike shape with a stalk and round bulbous holdfast, and a few have structures interpreted as spicules, zooids, and polyp leaves (Ivantsov, 1996; Jenkins, 1985, 1996). These fossil taxa have been linked specifically with derived living pennatulaceans. Waggoner (1998) suggested a similarity between certain Ediacaran "fronds" and the extant pennatulacean Renilla, which has a broad frondlike expansion of the rachis with variable arrangements of the secondary polyps, but which lacks polyp leaves. The fossil Ausia from Namibia has been interpreted as similar to a more basal pennatulacean, resembling living pennatulaceans in the Veretillidae (Hahn and Pflug, 1985). Taken at face value, all these reports would suggest that the Pennatulacea was not only present in the Vendian, but had diversified into several extant subclades by the Vendian.

FIGURE 1-Ediacaran taxa that may be members of the Cnidaria. 1, a possible pennatulacean, Charnia, from the Winter Coast of the White Sea, north Russia (University of California Museum of Paleontology collection). 2, field photograph of another possible pennatulacean referred tentatively to Charniodiscus, from Mistaken Point, Newfoundland. 3, a possible chondrophorine, Eoporpita, from the Winter Coast of the White Sea, north Russia (University of California Museum of Paleontology collection). 4, a possible coronate scyphozoan polyp tube, Corumbella (cross section), from the Mojave Desert of southern Nevada (Los Angeles County Museum of Natural History collection).

Finally, the third extant cnidarian group that may be represented in the Ediacara biota is the Scyphozoa (which may not be monophyletic; see below). As mentioned above, a few tetraradial "medusoids" could be scyphozoan jellyfish, but tetraradial symmetry is also typical of the Hydrozoa and Cubozoa. However, annulated tubes with definite tetraradial symmetry have been found in Ediacaran-age biotas in both South America and southwestern North America, and referred to the genus Corumbella (Hahn et al., 1982; Hagadorn and Waggoner, 2000). Corumbella is similar in many respects to the tubes secreted by the polyp stage of living coronate scyphozoans such as Stephanoscyphus (Leslie et al., 2001) and to the benthic cnidarians in the Stauromedusae (Collins et al., 2000). If the first hypothesis is true, Corumbella would be the earliest representative of Scyphozoa; if the second is true, Corumbella would belong to an outgroup of the extant Scyphozoa.

These hypotheses are difficult to test using Phanerozoic fossils. Pennatulaceans have a very sparse Phanerozoic record, although the Middle Cambrian fossil Thaumaptilon from the Burgess Shale is almost certainly a pennatulacean (Conway Morris, 1993). There is a sparse record of other octocorals in the early Paleozoic, including probable octocoral spicules from the early Cambrian of Australia (Microcoryne; see Bengtson et al., 1990); Cambrian polyps with eightfold symmetry and fringed tentacles, which are octocoral synapomorphies (Xianguangia; see Chen and Erdtmann, 1991 and Chen and Zhou, 1997); and a few plausible alcyonacean and gorgonian fossils from the Ordovician, Silurian, and Devonian (e.g., Glinski, 1956; Lindstrom, 1978; Bengtson, 1981). Paleozoic fossil scyphozoans are probably represe\nted by the Conulariida, a fairly well-known Paleozoic taxon of ribbed, conical, tetraradial fossils referred to either the coronate scyphozoans (Werner, 1966; Van Iten, 1991) or the stauromedusae (Jerre, 1994; Collins et al., 2000). The oldest conulariids are from the Late Cambrian (Hughes et al., 2000). Other tetraradial forms such as the Middle Ordovician Conchopeltis are often grouped together with the Conulariida to form a subgroup of Scyphozoa called Conulata (Moore and Harrington, 1956; but see Oliver, 1984). Scyphozoans may also be represented by the Early Cambrian Chinese genus Olivooides, whose fossils show ontogenetic development from a spherical embryo into an annulated tube, although these animals appear to have pentameral symmetry (Zhao and Bengston, 1999).

Chondrophorines ostensibly have a fossil record throughout the Phanerozoic (Stanley, 1986), but many of the Paleozoic fossils attributed to chondrophorines are controversial at best. For example, Plectodiscus has been described as a well-preserved chondrophorine from the Lower Devonian-age Hunsruckschiefer of Germany, but Otto (2000) has argued that the known specimens of Plectodiscus are not chambered, lacked zooids, and are in fact large, thin-shelled brachiopods. A Plectodiscus specimen from the Hunsruckschiefer that supposedly preserves the zooids of a chondrophorine does not show the diagnostic float, and Otto (2000) states that it is a pyritized trace fossil. The Cambrian genus Scenella, known from the Burgess Shale and similar Lagerstatte, has been the subject of debate; at least some specimens assigned to Scenella are more likely to be cap-shaped shells (Landing and Narbonne, 1992). Other proposed Paleozoic "chondrophorines" have also been reinterpreted as mollusc or brachiopod shells, trace fossils, or rotational sweep marks (e.g., Horny, 1985; Kase, 1988; Jensen et al., 2001). Still others may belong to poorly known metazoan taxa that may or may not belong in any known phyla (e.g., Dzik, 1991; Conway Morris et al., 1991; Geyer, 1994), and others are probably too poorly preserved to be informative (e.g., Waggoner and Collins, 1995).

Only careful morphological study of well-preserved fossils will ultimately resolve these controversies. However, other sources of information can be brought to bear on the problem of interpreting these fossils. A recent molecular phylogenetic analysis of the Cnidaria (Fig. 2) allows us to use "molecular clock" approaches to examine the possibility of chondrophorines, coronate or stauromedusan scyphozoans, and pennatulaceans in the Ediacara biota and afterwards. There are two possible approaches to this problem:

* Calibrating the clock using well-established fossils, and using that calibration to derive the likely age of the basal nodes. This is the usual way of using the molecular clock, and it has been used extensively in estimating divergence dates for the Metazoa and for various animal phyla and superphyla (see Table 1). However, it is not the approach we have used in this study. The fossil record of extant cnidarians is not rich. Scleractinian corals are the only major exception, but in view of their sudden appearance in the Triassic, it has been considered very likely that they arose from a Paleozoic softbodied anthozoan clade with no clear fossil record. Using the Phanerozoic cnidarian fossil record to calibrate the molecular clock for the Cnidaria as a whole would introduce a great deal of bias, since most, if not all, of the first fossil occurrences are likely to be significantly younger than the actual origins of the clades. This is potentially a problem in all clock analyses, since the only way that a first fossil occurrence could not be younger than the origin of the clade is if the fossil is the ancestor of the clade itself. However, it is likely to be a particularly serious problem in the Cnidaria.

* Calibrating the clock using the controversial fossils, and checking whether that calibration yields dates of origin for other clades that are consistent with other lines of evidence. This could be called the reductio ad absurdum approach, after a well-known technique in mathematical logic, in which a proposition is disproved if the chain of consequences deduced from it leads to a contradiction. If a calibration based on one interpretation of a controversial fossil yields a highly improbable age for several better-known clades, that interpretation is likely to be incorrect. This is the approach that we use here. We have carried out multiple "molecular clock" estimates of branching dates within a hypothesized phylogeny of the Cnidaria, all of which can be compared with the fossil record and with molecular divergence estimates of other taxa. Each of these clock analyses has been calibrated based on different interpretations of the Ediacaran fossil record.

Other researchers have used a similar tactic of comparing the estimated ages of clades using different calibration points. In a study of primate molecular phylogeny, Nei and Glazko (2002) showed that two different calibrations derived from the fossil record generally gave mutually consistent results within the range of statistical error. The primate-artiodactyl divergence, estimated from the fossil record at 90 Ma, calibrated the human-orangutan divergence at 11.3-13.7 Ma depending on the method of analysis, overlapping with the fossil-based estimate of 13 Ma. The fossil date of 13 Ma for the human-orangutan divergence in turn yielded molecular clock dates of 85.6-143.7 Ma for the primate-artiodactyl divergence. Huchon et al. (2000) used eight different calibration points from the fossil record in their phylogenetic and molecular clock analysis of rodents. They noted that some pairs of calibration points gave mutually consistent results, while others gave mutually inconsistent results. They suggested the use of several "cross- validated" calibration points would reduce dating errors. However, this cross-validation tactic has to our knowledge not previously been used in estimating deep divergence dates within the Metazoa; nor has it been used previously to test proposals of the taxonomic affinity of problematic fossils.

FIGURE 2-Maximum likelihood tree of 64 medusozoans plus 10 anthozoans as outgroups, assuming a general-time-reversible model of nucleotide evolution with rate heterogeneity. Score = -15,595.976. Scale bar denotes 0.1 nucleotide substitutions per site.

TABLE 1-Molecular clock estimates from various sources giving upper or lower bounds or direct estimates of the divergence data for the basal Cnidaria. For clarity, confidence intervals are usually not shown, and many dates are approximations summarized from somewhat heterogeneous data. Dates in parentheses were not corrected for relative rates of change (see Feng et al., 1997).

PROCEDURES

We used a molecular phylogeny for extant Cnidaria (Fig. 2) based on 18S rRNA sequences (1,768 nucleotides in length) for 74 taxa, aligned by hand. (See Table 2 for accession information; the full data set is publicly available at http://www.ucmp.berkeley.edu/ archdata/Collins-Medusozoa/, as well as on request from the authors.) The optimal tree was found using 10 replicate searches using the ML (maximum likelihood) algorithm of the program PAUP* (Swofford, 2000). We used the program Model-test (Posada and Crandall, 1998) to determine the model of nucleotide evolution that best fits the sequence data, using likelihood ratio tests. This gave us a general-time-reversible model with rate heterogeneity (assumed nucleotide frequencies: A = 0.2599; C = 0.1986; G = 0.2674; T = 0.2741; substitution types = 6, proportion of invariable sites = 0.5007, [gamma] for variable sites = 0.5646, substitution rates equal to 1.000 except between A and G = 2.481 and C and T = 4.069). Further information on the phylogeny, along with a full discussion of its implications for cnidarian systematics and evolution, has been published in Collins (2002).

Ideally, all taxa would be simultaneously analyzed in a molecular clock analysis. However, a likelihood ratio test rejects the hypothesis that the 18S data has evolved in a clocklike manner, indicating that there is significant heterogeneity in rates across the 74 cnidarian taxa examined here. Deletion of the longest branches, using the relative rates test, did not yield clocklike evolution in the remaining sequences; repeated deletions ultimately led to the excision of entire clades of interest from the phylogeny. Therefore, we repeatedly selected quartets of taxa that span the origins of three major taxa: the Cnidaria itself, the Medusozoa (non- anthozoan Cnidaria), and the unnamed clade consisting of the Cubozoa plus the Hydrozoa. Each quartet consisted of two pairs of sister taxa, and each pair was assigned an estimated date of divergence based on the fossil record (Tables 3, 4). We used the program QDate 1.1.1 (Rambaut and Bromham, 1998; available at http:// evolve.zoo.ox.ac.uk/) to estimate the date of divergence of the two pairs from each other, given divergence dates of each taxon pair (Fig. 3). QDate uses a maximum-likelihood approach to reconstruct the date of divergence of the pairs under the assumption that each pair evolves at its own substitution rate. Confidence intervals on the date are calculated using a log-likelihood approach. The program also runs a log-likelihood test for rate heterogeneity between the two pairs of taxa; quartets of taxa with heterogeneous substitution rates, in which all branches of the tree have their own substitution rates, can be identified and rejected. QDate allows dates to be reconstructed under a range of substitution models. We used the most flexible, the general time reversible model (GTR), in which each possible substitution is assigned its own probability, and the nucleotide frequencies and rate heterogeneity parameter ([Gamma], gamma) are specified. For each quartet, these parameters were estima\ted using maximum likelihood with PAUP* and subsequently provided to the QDate program.

Table 2 contains a number of relevant molecular clock estimates, including Lynch's 1999 estimate of abut 900 Ma for the origin of the diploblasts. Estimates of the date of the split between total-group Metazoa and total-group Fungi or Plantae place lower bounds on this branch point. Estimates of the date of the protostome-deuterostome divergence place rough upper bounds on the branch point, although these can only be tentative upper bounds because of the possibility that all extant cnidarian lineages diverged after the protostome- deuterostome split. Despite this potential problem, and despite the sizable margins for error of many of these estimates, it seems safe to say that a divergence date between 800 and 1,000 million years for the origin of the crown-group Cnidaria is, at the very least, not grossly contradicted by other analyses.

RESULTS

The ML tree derived from the general-time-reversible model of nucleotide evolution is generally well supported, and is very similar to trees derived from the same dataset using the maximum parsimony criterion (see Collins, 2002 for full discussion). Our results are summarized graphically in Figures 4 through 7; the raw numerical data is available at http://www.ucmp.berkeley.edu/ archdata/Waggoner_Collins_2001/ or on request from the authors. There is a considerable range of estimates for the dates of origin for major cnidarian groups, depending on the quartet and the fossil calibration dates. For instance, the origin of Cnidaria is estimated to be 672 Ma if the Cubozoa had diverged by 300 Ma and the Scleractinia by 235 Ma. In contrast, the oldest inferred date for the origin of Cnidaria that does not violate the rate-homogeneity assumption is 1,811 Ma, based on the interpretation that both pennatulaceans and chondrophorines were present in the Late Neoproterozoic. Since these dates are not statistically independent of each other, it is not statistically meaningful to take their mean and standard deviation as a "consensus" estimate.

However, despite the scatter, the results do show general patterns. The distribution of estimates for the base of the Cnidaria is almost bimodal (Figs. 4, 5). Estimates that do not use the Chondrophorina lie between 800 and 1,200 Ma. Those that do not include either the Cubozoa or Chondrophorina place the base of the Cnidaria between 750 and 1,000 Ma, with upper confidence limits up to 1,200 Ma and lower limits down to 700 Ma. This range of values is still very broad (the confidence limits span an interval the size of the Phanerozoic!), but is at least consistent with previous molecular estimates for the age of the Cnidaria (Table 1). On the other hand, almost all dates for the base of the Cnidaria that are based on the assumption that the Chondrophorina diverged in the Neoproterozoic or Paleozoic lie between 1,600 and 2,000 Ma, with low limits at about 1,400 Ma and high limits up to 2,400 Ma. Estimates for the Cnidaria that were calibrated using either Cubozoa at 300 Ma, or Chondrophorina at 220 Ma, fall in between these two ranges.

TABLE 2-List of taxa used in this study. Accession numbers refer to the GenBank database.

Dates for internal branches within the Cnidaria follow the same pattern as dates for the base of the Cnidaria: Dates based on Paleozoic and Proterozoic chondrophorines are always significantly older than dates based on any other taxa (Figs. 6, 7). For example, calibrations using chondrophorines at 555 Ma and cubozoans at 300 Ma yield an age of 1,978 Ma (+641/-457 Ma) for the Cubozoa-Hydrozoa split within the Cnidaria-which is older than most of our estimates for the age of the base of the Cnidaria. Some authors have suggested that not only were chondrophorines present in the Ediacara biota, but the two living lineages of chondrophorines, represented by the living genera Porpita and Velella, had diverged from each other by the late Neoproterozoic (e.g., Stanley, 1986). We attempted to use this as a calibration, but this violated the rate homogeneity test. However, this calibration did yield multiple estimates for the base of the Cnidaria that were consistently greater than 2,000 Ma. Our results do not support a Porpita-Velella divergence in the Neoproterozoic or Paleozoic.

The best-documented and least controversial fossil date in the analysis is the first fossil appearance of Scleractinia in the Middle Triassic. The statistically valid dates for the base of the Cnidaria that are based on a Triassic origin of the Scleractinia are unusually recent (649 and 527 Ma), except when both chondrophorines and Triassic scleractinians are used to calibrate the same analysis. Using both yields dates between 850 and 950 Ma for the base of the Cnidaria; apparently the effects of these two calibrations cancel each other out. Using the Triassic appearance of the Scleractinia as a calibration point yields unusually young dates for the base of the Cnidaria. However, the Scleractinia almost certainly evolved from soft-bodied sea anemones or anemone-like ancestors, and the Scleractinia-Actiniaria divergence probably happened millions of years before the first appearance of fossil scleractinian corals (Oliver, 1996). This probably explains why dates for the Cnidaria calibrated using the Middle Triassic appearance of Scleractinia are unusually young: the first appearance of scleractinian corals postdates the origin of the clade.

One way around this problem would have been to calibrate the clock using fossils of sea anemones. However, the fossil record of sea anemones is sparse and hard to interpret. We elected not to use purported soft-bodied anemone-like fossils, such as the Cambrian Mackenzia and the Ediacaran fossil Inaria, to calibrate our clock; their taxonomic position is unclear. However, a small group of Middle Ordovician corals, the Kilbuchophyllida, is thought to be a short-lived biomineralized offshoot of the lineage that gave rise to both Scleractinia and modern Actiniaria (Oliver, 1996). Assuming that kilbuchophyllids form an outgroup to both Actiniaria and Scleractinia, as implied by the phylogeny in Oliver (1996), this gives us an upper bound for the date of the Scleractinia-Actiniaria divergence. Using the Kilbuchophyllida to calibrate this divergence yields estimates for the base of the Cnidaria of 700-1,000 Ma, which are much more consistent with other estimates in this and other studies (Fig. 5). The Lower Cambrian corallike Tabulaconariida and Middle Cambrian Cothoniida are thought to be basal Hexacorallia (Oliver, 1996); using their ages to calibrate the clock also yields consistent dates of 700-1,000 Ma for the base of the Cnidaria.

TABLE 3-Calibration points used in this study. Proterozoic and Cambrian dates from Benus (1988), Grotzinger et al. (1995) and Martin et al. (2000); others from Harland et al. (1990).

DISCUSSION

Molecular clock analyses suffer from a number of well-known limitations that involve differential rates of molecular evolution across taxa and genomes, and in particular over time within lineages (Goodman et al., 1987; Ayala, 1997). Nevertheless, we draw two general conclusions from our results. First, choices of both fossil calibrations and taxa analyzed have a profound impact on divergence estimates based on molecular sequence data. These choices introduce inaccuracies into molecular clock analyses that are typically ignored when errors are calculated. The scatter in our results suggests that this source of error in other published studies may be substantial (see also Huchon et al., 2000, 2002). Second, our results are consistent with those of other molecular clock studies in suggesting that there is a sizable history of Metazoa that is absent from the recovered fossil record. Taphonomic hypotheses that could explain such a gap in the fossil record should be explored in greater detail.

There is certainly additional work that should be done to confirm and expand on our results. We suspect that the dates in this study are overestimates of the real branching dates; molecular clocks suffer from an inherent bias towards overestimation (Rodriguez- Trelles et al., 2002). Furthermore, in a molecular clock analysis of the divergence of the Metazoa, Bromham et al. (1998) found that dates based on 18S rRNA were considerably older, as well as being less well constrained, than dates based on mitochondrial genes. They suggested that the evolution of paired sites in rRNA was not adequately modeled by their dating method, and recommended that mitochondrial genes should be used in clock studies. Recent work by Peterson and Takacs (2002) has shown that using amino acid sequences of "housekeeping" genes yields clock dates that are not only younger than 18S rRNA dates, but that agree much more closely with the fossil record. We agree that other sequences may yield more accurate dates, and we hope to use other data sets in the near future to test our results. Also, relatively new methods of deriving divergence dates involving Bayesian inference (Huelsenbeck et al., 2000; Kishino et al., 2001) should be useful in future studies. These methods are in principle able to handle rate heterogeneity on all branches of a phylogeny.

TABLE 4-Taxon pairs used to calibrate the divergence dates of various higher taxa in this study.

However, what is most important about our results is not so much that they give the exact ages of basal cnidarian nodes. What is important is this: some hypotheses of what the Ediacaran fossils are lead to sets of deduced molecular dates that are both internally consistent with each other, and consistent with independent estimates based on other sequences, taxa, and calibrations. Other hypotheses of the affinity of certain Ediacaran fossils yield estimates that are neither internally nor externally consistent. We predict that molecular dates for the Cnidaria based on other sequences and \algorithms will show the same pattern as our results from 18S rRNA: calibrations based on the assumption of chondrophorines in the Precambrian will be consistently older than others.

FIGURE 3-Model used by the program QDate (after Rambaul and Bromham, 1998). The divergence dates of each pair of taxa, t^sub x^ and t^sub y^, are estimated from the fossil record. QDate then estimates t^sub z^, the date for the basal node of the quartet, using a maximum-likelihood algorithm, under the assumption that each pair of taxa has its own mutation rate ([mu]^sub x^ and [mu]^sub y^).

FIGURE 4-Molecular estimates for the age of the base of the Cnidaria, calibrated using Ediacaran and Cambrian ages for the Pennatulacea. One of the species pair dates for each quartet is indicated along the X-axis; the other is indicated with symbols as shown in the key. Estimates with no error bars failed the rate- heterogeneity test.

Our results do not contradict the hypotheses that pennatulaceans and scyphozoans were present in the "Ediacara biota" and/or in the Cambrian. In fact, pennatulaceans may not have been present as early as 565 Ma. The sea penlike forms from Mistaken Point, Newfoundland, show some striking differences, not just from pennatulaceans, but from metazoans in general (BW, personal observations, 2001). Furthermore, Dzik (2002) has argued that many of the frondlike Ediacaran forms should be considered stem-group ctenophores, not colonial cnidarians. However, moving the Pennatulacea calibration date forward from 565 million years, even well into the Cambrian, has little effect on the estimated date of divergence for the Cnidaria. On the other hand, the hypothesis of chondrophorines in the late Precambrian and Paleozoic is never supported. Calibrations using Ediacaran, Cambrian, and Devonian dates for the origin of the chondrophorines lead to age estimates for the base of the Cnidaria that are centered about 500 million years older than the most extreme previous proposal. When Cambrian and Ediacaran dates are used for the chondrophorines, the confidence limits on the basal Cnidaria date overlap with the age of the oldest known eukaryote body fossils (Han and Runnegar, 1992). In this context, it is worth noting that Fedonkin (1998) has suggested that several ovoid, strongly ribbed Ediacaran fossils, including the "chondrophorines" Ovatoscutum and Chondroplon as well as dickinsoniids and an undescribed taxon (informally named "Andiva"), might be bilaterians with a relatively firm, segmented dorsal shield or carapace. Eoporpita, on the other hand, might be a benthic polyplike form. We suggest that alternative hypotheses such as these deserve closer attention.

FIGURE 5-Estimates for the age of the base of the Cnidaria, calibrated using Cambrian ages for the Hexacorallia, and Devonian and Triassic ages for the Scleractinia. See Figure 4 caption for explanation of symbols.

FIGURE 6-Estimates for the age of the base of the Medusozoa (i.e., Cnidaria minus Anthozoa). See Figure 4 caption for explanation of symbols.

FIGURE 7-Estimates for the age of the Cubozoa-Hydrozoa split. See Figure 4 caption for explanation of symbols.

We reiterate that molecular analyses of extant taxa, by themselves, cannot confirm or disprove hypotheses of the affinities of a problematic fossil. Only careful morphological and comparative analyses of the fossils themselves can do that. However, we believe that molecular phylogenetic dating, using the reductio ad absurdum approach, can provide useful information that can be brought to bear on testing the affinities of problematic fossils.

ACKNOWLEDGMENTS

An earlier version of this work was presented at the 2001 North American Paleontological Convention in Berkeley, California, in a special session titled "When Clocks Collide." We thank C. Brochti, C. Sumrall, and J. M. Theodor for organizing the session and for inviting us to contribute our work, and two anonymous reviewers for their contributions. AGC acknowledges the support of NSF grant EAR- 9814845 to J. Valentine and J. Lipps, and the muses B&S. BW acknowledges the help of T. de'Clouet, B. Ellman, R. Mercurio, S. Moore, J. Raines, and R. Vogel.

REFERENCES

AYALA, F. J. 1997. Vagaries of the molecular clock. Proceedings of the National Academy of Sciences of the United States of America, 94:7776-7783.

AYALA, F. J. A. RZHETSKY, AND F. J. AYALA. 1998. Origin of the metazoan phyla: molecular clocks confirm paleontological estimates. Proceedings of the National Academy of Sciences of the USA, 95:606- 611.

BABCOCK, L. E., AND R. A. ROBISON. 1988. Taxonomy and paleobiology of some Middle Cambrian Scenella (Cnidaria) and hyolithids (Mollusca) from western North America. University of Kansas Paleontological Contributions, Number 121, 22p.

BENGTSON, S. 1981. Atractosella, a Silurian alcyonacean octocoral. Journal of Paleontology, 55:281-294.

BENGTSON, S., S. CONWAY MORRIS, B. J. COOPER, P. A. JELL, AND B. N. RUNNEGAR. 1990. Early Cambrian fossils from south Australia. Memoirs of the Association of Australasian Palaeontologists, 9:1- 364.

BENUS, A. P. 1988. Sedimentological context of a deep-water Ediacaran fauna (Mistaken Point Formation, Avalon Zone, eastern Newfoundland), p. 8-9. In E. Landing, G. M. Narbonne, and P. Myrow (eds.), Trace Fossils, Small Shelly Fossils and the Precambrian- Cambrian Boundary. New York State Museum and Geological Survey Bulletin, Number 463.

BOSCHMA, H. 1956. Milleporina and Stylasterina, p. F89-F106. In R. C. Moore (ed.), Treatise on Invertebrate Paleontology, Pt. F, Coelenterata. Geological Society of America and University of Kansas Press, Lawrence.

BROMHAM, L., A. RAMBAUT, R. FORTEY, A. COOPER, AND D. PENNY. 1998. Testing the Cambrian explosion hypothesis by using a molecular dating technique. Proceedings of the National Academy of Sciences of the USA, 95:12386-12389.

BRUTON, D. L. 1991. Beach and laboratory experiments with the jellyfish Aurelia and remarks on some fossil "medusoid" traces, p. 125-129. In A. M. Simonetta and S. Conway Morris (eds.), The Early Evolution of Metazoa and the Significance of Problematic Taxa. Cambridge University Press, Cambridge.

CHEN, JUN-YUAN, AND B.-D. ERDTMANN. 1991. Lower Cambrian fossil Lagerstatte from Chengjiang, Yunnan, China: insights for reconstructing early metaozoan life, p. 57-76. In A. M. Simonetta and S. Conway Morris (eds.), The Early Evolution of Metazoa and the Significance of Problematic Taxa. Cambridge University Press, Cambridge.

CHEN, JUN-YUAN, AND G. ZHOU. 1997. Biology of the Chengjiang fauna. Bulletin of the National Museum of Natural Science (Taiwan), 10:11-106.

COLLINS, A. G. 2002. Phylogeny of Medusozoa and the evolution of cnidarian life cycles. Journal of Evolutionary Biology, 15:418-432.

COLLINS, A. G., A. C. MARQUES, AND M. G. SIMOES. 2000. The phylogenetic placement of Conulatae within Cnidaria. GSA Abstracts with Programs, 32.

CONWAY MORRIS, S. 1993. Ediacaran-like fossils in Cambrian Burgess Shale-type faunas of North America. Palaeontology, 36:593- 635.

CONWAY MORRIS, S., L. E. SAVOY, AND G. HARRIS. 1991. An enigmatic organism from the 'Exshaw' Formation (Devonian-Carboniferous), Alberta, Canada. Lethaia, 24:139-152.

DOOLITTLE, R. F., D. F. FENG, S. TSANG, G. CHAO, AND E. LITTLE. 1996. Determining divergence times of the major kingdoms of living organisms with a molecular clock. Science, 271:470-477.

DZIK, J. 1991. Is fossil evidence consistent with traditional views of the early metazoan phylogeny? p. 47-58. In A. M. Simonetta and S. Conway Morris (eds.), The Early Evolution of Metazoa and the Significance of Problematic Taxa. Cambridge University Press, Cambridge.

DZIK, J. 2002. Possible ctenophoran affinities of the Precambrian "seapen" Rangea. Journal of Morphology, 252:315-334.

FEDONKIN, M. A. 1990. Systematic description of Vendian metazoa, p. 71-120. In B. S. Sokolov and A. B. Iwanowski (eds.), The Vendian System. Volume 1, Paleontology. Springer-Verlag, Berlin.

FEDONKIN, M. A. 1992. Vendian faunas and the early evolution of Metazoa, p. 87-129. In J. H. Lipps and P. W. Signor (eds.), Origin and Early Evolution of the Metazoa. Plenum Press, New York.

FEDONKIN, M. A. 1998. Metameric features in the Vendian metazoans. Italian Journal of Zoology, 65:11-17.

FENG, D. F., G. CHO, AND R. F. DOOLITTLE. 1997. Determining divergence times with a protein clock: update and recalibration. Proceedings of the National Academy of Sciences of the USA, 94:13028- 13033.

FOSTER, M. W. 1979. Soft-bodied coelenterates in the Pennsylvanian of Illinois, p. 191-267. In M. H. Nitecki (ed.), Mazon Creek Fossils. Academic Press, New York.

GEHLING, J. G. 1991. The case for Ediacaran fossil roots to the metazoan tree. Memoirs of the Geological Society of India, 20:181- 223.

GEHLING, J. G., AND J. K. RIGBY. 1996. Long-expected sponges from the Neoproterozoic Ediacara fauna of South Australia. Journal of Paleontology, 70:185-195.

GEHLING, J. G., G. M. NARBONNE, AND M. A. ANDERSON. 2000. The first named Ediacaran body fossil, Aspidella terranovica. Palaeontology, 43:427-456.

GERSHWIN, L., J. H. LIPPS, R. H. Dorr JR., AND D. DAMROW. 2001. Late Cambrian jellyfish behavioral taphonomy in central Wisconsin. PaleoBios, 21(S2):56.

GEYER, G. 1994. An enigmatic bilateral fossil from the Lower Cambrian of Morocco. Journal of Paleontology, 68:710-716.

GLAESSNER, M. F. 1984. The Dawn of Animal Life: A Biohistorical Study. Cambridge University Press, Cambridge, xii + 244 p.

GLINSKI, A. 1956. Plumalina conservuta n. sp. (Gorgonaria) aus dem Mittel-Devon der Eifel. Senckenbergiana Lethaea, 37:53-57.

GOODMAN, M., M. M. MIYAMOTO, AND J. CZELUSNIAK. 1987. Pattern and process in vertebrate phylogeny revealed by coevolution of molecules and morphologies, p. 141-176. In C. Patterson (ed.), Molecules and Morphology in Evolution: Conflict or Compromise? Cambridge University Press, Cambridge.

GROTZINGER, J. P., S. A. BOWRING, B. Z. SAYLOR, AND A. J. KAUFMAN. 1995. Biostratigraphic a\nd geochronologic constraints on early animal evolution. Science, 270:598-604.

HAGADORN, J. W., AND B. M. WAGGONER. 2000. Ediacaran fossils from the southwestern Great Basin, United States. Journal of Paleontology, 74:349-359.

HAGADORN, J. W., R. H. DOTT, AND D. DAMROW. 2002. Stranded on an Upper Cambrian shoreline: Medusae from Central Wisconsin. Geology, 30:147-150.

HAHN, G., AND H.-D. PFLUG. 1985. Polypenartige Organismen aus dem Jung-Prakambrium (Nama-Gruppe) von Namibia. Geologica et Palaeontologica, 19:1-13.

HAHN, G., R. HAHN, O. H. LEONARDOS, H.-D. PFLUG, AND D. H. G. WALDE. 1982. Korperlich erhaltene Scyphozoen-Reste aus dem Jungprakambrium Brasiliens. Geologica et Palaeontologica, 16:1-18.

HAHN, G., R. HAHN, O. H. LEONARDOS, H.-D. PFLUG, AND D. H. G. WALDE. 1982. Korperlich erhaltene Scyphozoen-Reste aus dem Jungprakambrium Brasiliens. Geologica et Palaeontologica, 16:1-18.

HAN, T.-M., AND B. RUNNEGAR. 1992. Megascopic eukaryotic algae from the 2.1 billion-year-old Negaunee iron-Formation, Michigan. Science, 257:232-235.

HARLAND, W. B., R. L. ARMSTRONG, A. V. COX, L. E. CRAIG, A. G. SMITH, AND D. G. SMITH. 1990. A Geologic Time Scale, 1989 Edition. Cambridge University Press, Cambridge, 263 pp.

HASEGAWA, M., H. KISHINO, AND T. YANO. 1985. Dating of the human- ape splitting by a molecular clock of mitochondrial DNA. Journal of Molecular Evolution, 21:160-174.

HOFFMAN, H. J. 1988. An alternative interpretation of the Ediacaran (Precambrian) chondrophore Chondroplon Wade. Alcheringa, 12:315-318.

HORNY, R. 1985. Bellerophacmaea gen. n. (Mollusca) and Palaeolophacmaea Donaldson, 1962 (Coelenterata), an example of homeomorphy in invertebrate paleozoology. Casopis Narodniho Muzea v Praze, 154:119-124.

HUCHON, D., F. M. CATZEFLIS, AND E. J. P. DOUZERY. 2000. Variance of molecular datings, evolution of rodents and the phylogenetic affinities between Ctenodactylidae and Hystricognathi. Proceedings of the Royal Society of London, Series B: Biological Sciences, 267:393-402.

HUCHON, D., O. MADSEN, M. J. J. B. SIBBALD, K. AMENT, M. J. STANHOPE, F. CATZEFLIS, W. W. DE JONG, AND E. J. P. DOUZERY. 2002. Rodent phylogeny and a timescale for the evolution of Glires: evidence from an extensive taxon sampling using three nuclear genes. Molecular Biology and Evolution, 19:1053-1065.

HUELSENBECK, J. P., B. LARGET, AND D. L. SWOFFORD. 2000. A compound Poisson process for relaxing the molecular clock. Genetics, 154:1879-1892.

HUGHES, N. C., G. O. GUNDERSON, AND M. J. WEEDON. 2000. Late Cambrian conulariids from Wisconsin and Minnesota. Journal of Paleontology, 74:828-838.

IVANTSOV, A. 1996. Oldest skeletal remains in the terminal Neoproterozoic of the Russian platform. PaleoBios, 17:S7.

JENKINS, R. J. F. 1985. The enigmatic Ediacaran (late Precambrian) genus Rangea and related forms. Paleobiology, 11:336- 355.

JENKINS, R. J. F. 1992. Functional and ecological aspects of Ediacarian assemblages, p. 131-176. In J. H. Lipps and P. W. Signor (eds.), Origin and Early Evolution of the Metazoa. Plenum Press, New York.

JENKINS, R. J. F. 1996. Aspects of the geological setting and palaeobiology of the Ediacara assemblage, p. 33-45. In M. Davies, C. R. Twidale, and M. J. Tyler (eds.), Natural History of the Flinders Ranges. Royal Society of South Australia.

JENSEN, S., M. L. DROSER, AND J. G. GEHLING. 2001. A scratch circle origin for the medusoid fossil Kullingia. PaleoBios, 21:S74- S75.

JERRE, F. 1994. Anatomy and phylogenetic significance of Eoconularia loculata, a conulariid from Silurian of Gotland. Lethaia, 27:97-109.

KASE, T. 1988. Reinterpretation of Brunonia annulata (Yokoyama) as an early Cretaceous carinariid mesogastropod (Mollusca). Journal of Paleontology, 62:766-771.

KISHINO, H., J. E. THORNE, AND B. J. WILLIAM. 2001. Performance of a divergence time estimation method under a probabilistic model of rate evolution. Molecular Biology and Evolution, 18:352-361.

LANDING, E., AND G. M. NARBONNE. 1992. Scenella and "a chondrophorine (medusoid hydrozoan) from the basal Cambrian (Placentian) of Newfoundland." Journal of Paleontology, 66:338.

LEE, M. S. Y. 1999. Molecular clock calibrations and metazoan divergence dates. Journal of Molecular Evolution, 49:385-391.

LESLIE, S. A., L. E. BABCOCK, A. M. GRUNOW, AND G. R. SADOWSKI. 2001. Paleobiology and paleobiogeography of Corumbella, a Late Neoproterozoic Ediacaran-grade organism. PaleoBios, 21:S83-S84.

LINDSTROM, M. 1978. An octocoral from the Lower Ordovician of Sweden. Geologica et Palaeontologica, 12:41-47.

LYNCH, M. 1999. The age and relationships of the major animal phyla. Evolution, 53:319-325.

MARTIN, M. V., D. V. GRAZHDANKIN, S. A. BOWRING, D. A. D. EVANS, M. A. FEDONKIN, AND J. E. KIRSCHVINK. 2000. Age of Neoproterozoic bilaterian body and trace fossils, White Sea, Russia: implications for metazoan evolution. Science, 288:841-846.

MOORE, R. C., AND H. J. HARRINGTON. 1956. Scyphozoa, p. F27-F38. In R. C. Moore (ed.), Treatise on Invertebrate Paleontology, Pt. F, Coelenterata. Geological Society of America and University of Kansas Press, Lawrence.

NEI, M., AND G. GLAZKO. 2002. Estimation of divergence times for a few mammalian and several primate species. Journal of Heredity, 93: 157-164.

NIKOH, N., N. IWABE, K. KUMA, M. OHNO, T. SUGIYAMA, Y. WATANABE, K. YASUI, Z. SHI-CUI, K. HORI, Y. SHIMURA, AND T. MIYATA. 1997. An estimate of divergence time of Parazoa and Eumetazoa and that of Cephalochordata and Vertebrata by aldolase and triose phosphate isomerase clocks. Journal of Molecular Evolution, 45:97-106.

OLIVER, W. A. JR. 1984. Conchopeltis: its affinities and significance. Palaeontographica Americana, 54:141-147.

OLIVER, W. A. JR. 1996. Origins and relationships of Paleozoic coral groups and the origin of the Scleractinia. In G. D. Stanley Jr. (ed.), Paleobiology and Biology of Corals. Paleontological Society Papers 1, p. 107-134.

OTTO, M. 2000. Supposed soft tissue preservation in the Hunsruckschiefer (Lower Devonian, Rheinisches Schiefergebirge): the example of brachiopods. Palaontologische Zeitschrift, 74:79-89.

PETERSON, K., AND C. TAKACS. 2002. Fossils, molecular clocks and the Cambrian explosion. Geological Society of America Abstracts with Programs, 33(7):A440.

POSADA, D., AND K. A. CRANDALL. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics, 14:817-818.

RAMBAUT, A., AND L. BROMHAM. 1998. Estimating divergence dates from molecular sequences. Molecular Biology and Evolution, 15:442- 448.

RETALLACK, G. J. 1994. Were the Ediacaran fossils lichens? Paleobiology, 20:523-544.

RODRIGUEZ-TRELLES, F., R. TARRIO, AND F. J. AYALA. 2002. A methodological bias toward overestimation of molecular evolutionary time scales. Proceedings of the National Academy of Sciences of the USA, 99:8112-8115.

SEILACHER, A. 1992. Vendobionta and Psammocorallia: lost constructions of Precambrian evolution. Journal of the Geological Society, London, 149:607-613.

SEILACHER, A. 1994. Early life on Earth: Late Proterozoic fossils and the Cambrian explosion, p. 389-400. In S. Bengtson (ed.), Early Life on Earth. Nobel Symposium 84. Columbia University Press, New York.

SOKOLOV, B. S. 1968. Vendian and Early Cambrian sabelliditida (Pogonophora) of the USSR. Proceedings of the IPU, International Geological Congress, 23:79-86.

SORAUF, J. E., AND M. E. SAVARESE. 1995. Lower Cambrian coral from South Australia. Palaeontology, 38:757-770.

STANLEY, G. D. JR. 1986. Chondrophorine hydrozoans as problematic fossils, p. 68-86. In A. Hoffman and M. H. Nitecki (eds.), Problematic Fossil Taxa. Oxford University Press, New York.

STANLEY, G. D. JR., AND B. SENOWBARI-BARYAN. 1999. Upper Triassic reef fauna from the Quesnel terrane, Central British Columbia, Canada. Journal of Paleontology, 73:787-802.

SUN, WEI-GUO. 1986. Late Precambrian pennatulids (sea pens) from the eastern Yangtze Gorge, China: Paracharnia gen. nov. Precambrian Research, 31:361-375.

SWOFFORD, D. L. 2000. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Sinauer Associates, Sunderland, Massachusetts.

VAN ITEN, H. 1991. Evolutionary affinities of conulariids, p. 145- 155. In A. M. Simonetta and S. Conway Morris (eds.), The Early Evolution of Metazoa and the Significance of Problematic Taxa. Cambridge University Press, Cambridge.

WADE, M. 1972. Hydrozoa and Scyphozoa and other medusoids from the Precambrian Ediacara fauna, South Australia. Palaeontology, 15:197-225.

WAGGONER, B. M. 1998. New approaches to the systematics and paleobiology of the late Precambrian soft-bodied biotas. Unpublished Ph.D. thesis, University of California at Berkeley, xv + 266 pp.

WAGGONER, B. M. 1999. Los comienzos de la historia evolutive de los artropodos: que nos pueden contar los fosiles? p. 115-131. In A. Melic, J. J. de Haro, M. Mendez, and I. Ribera (eds.), Evolucion y Filogenia de Arthropoda. Sociedad Entomologica Aragonesa, Barcelona.

WAGGONER, B. M., AND A. G. COLLINS. 1995. A new chondrophorine (Cnidaria, Hydrozoa) from the Cadiz Formation (Middle Cambrian) of California. Palaontologische Zeitschrift, 69:7-16.

WERNER, B. 1966. Slephanoscyphus (Scyphozoa, Coronatae) und seine direkte abstammung von den fossilen Conulata. Helgolander Wissenschaftliche Meeresuntersuchungen, 13:317-347.

WILLIAMS, G. C. 1992. Biotic diversity, biogeography, and phylogeny of pennatulacean octocorals associated with coral reefs in the Indo-Pacific. Proceedings of the Seventh International Coral Reef Symposium, 2:729-735.

WILLIAMS, G. C. 1995. Living genera of sea pens (Coelenterata: Octocorallia: Pennatulacea): illustrated key and synopses. Zoological Journal of the Linnaean Society, 113:93-140.

WRAY, G. A., J. S. LEVINTON, AND L. H. SHAPIRO. 1996. Molecular evidence for deep preCambrian divergences among metazoan phyla. Science, 274:568-573.

YOCHELSON, E. L., W. STURMER, AND G. D. STANLEY JR. 1983. Plectodiscus discoideus (Rauff): a redescription of a chondrophorine from the Early Devonian Hunsruck Slate, West Ger\many. Palaontologische Zeitschrift, 57:39-68.

ZHAO, Y, AND S. BENGTSON. 1999. Embryonic and post-embryonic development of the Early Cambrian cnidarian Olivooides. Lethaia, 32: 181-195.

ZHURAVLEV, A. YU. 1993. Were Ediacaran Vendobionta multicellulars? Neue Jahrbuch fur Geologie und Palaontologie, 190:299-314.

ZHURAVLEV, A. YU. 1999. A new coral from the Lower Cambrian of Siberia. Paleontological Journal, 33:502-508.

ACCEPTED 15 AUGUST 2003

BEN WAGGONER AND ALLEN G. COLLINS1

Department of Biology, University of Central Arkansas, Conway 72035-5003 USA, and Museum of Paleontology, University of California, 1101 Valley Life Sciences Building, Berkeley 94720-4780 USA

1 Present address: TiHo Hannover, ITZ, Ecology & Evolution, Bunteweg 17d, 30559 Hannover, GERMANY,

Copyright Paleontological Society Jan 2004

More News in this Category