Debate about Cambrian lophotrochozoans


Since 1990 there has been intense debate among paleontologists about the evolution in the Early Cambrian period of the "super-phylum" Lophotrochozoa, which is thought to include the modern molluscs, annelid worms and brachiopods, as well as their evolutionary "aunts" and "cousins".

Organisms featured in the debate

The name "halwaxiid" was formed by combining the names of two members of the proposed group, Halkieria and Wiwaxia. The group was defined as a set of Early to Mid Cambrian animals that had: a "chain mail" coat of three concentric bands of small armor plates that are called sclerites; in some cases, a small cap-like shell at the front end and in some cases both ends. Some scientists are unhappy with this loose definition, arguing that such traits may have arisen convergently rather being inherited from a common ancestor. This objection suggests that the group is not monophyletic, in other words that the group does not contain all and only the descendants of a single common ancestor. In fact the originators of the term "halwaxiid" prefer an evolutionary "family tree" in which the halwaxiids are not monophyletic.
Other fossil organisms became involved in the debate as there were plausible arguments for considering them closely related to the halkieriids or Wiwaxia or both. One was the more recently discovered Orthrozanclus, which looked an intermediate between the halkieriids and Wiwaxia as it had a shell like that of halkieriids, and unmineralized sclerites and long spines like those of Wiwaxia – in fact the article which first described Orthrozanclus introduced the term "halwaxiid". New finds of Odontogriphus put this animal into play as well since, despite its lack of sclerites or shells, its feeding apparatus looks very like Wiwaxia’s. The siphogonotuchids, a very Early Cambrian group known only from isolated sclerites among the small shelly fossils, also appear in analyses as their sclerites suggest this group may have been close to the ancestors of halkieriids.

Suggested phylogenies

Since 1995 several phylogenies, or evolutionary "family trees", have been suggested for these organisms. The position of Wiwaxia is highlighted, since the position of this organism has been central to the debate since 1990.

Wiwaxia

The most intense part of the debate centers round Wiwaxia. Conway Morris originally dismissed the earliest classification of Wiwaxia as a polychaete worm, because he thought there was little structural similarity between a polychaete's scale-like elytra and Wiwaxia’s sclerites, and because the arrangement of the sclerites, with quite different numbers in each band, showed no sign of the regular segmentation that is a feature of polychaetes. Instead he thought Wiwaxia was very similar to shell-less aplacophoran molluscs, that it must have moved on a mollusc-like muscular foot, and that its feeding apparatus looked like a primitive form of the molluscan radula, a tooth-bearing chitinous "tongue". Hence he classified it as a "sister" of the molluscs. When he briefly described the first articulated specimens of Halkieria in 1990, Conway Morris wrote of "the halkieriid-wiwaxiid body plan" and that the halkieriids might be close relatives of molluscs.
Shortly after this in 1990 Butterfield published his first paper on Wiwaxia. He argued that, since Wiwaxia’s sclerites appeared to be solid, they were not similar to the hollow sclerites of halkieriids. In fact he thought they were more similar to the chitinous bristles that project from the bodies of modern annelids and in some genera form leaf-like scales that cover the back like roof tiles - in composition, in detailed structure, in how they were attached to the body via follicle-like pockets in the skin, and in overall appearance. He also contended that Wiwaxia’s feeding apparatus, instead of being mounted in the middle of its "head", was just as likely to be mounted in two parts on the sides of the "head", an arrangement that is common in polychaetes. He therefore classified Wiwaxia as a polychaete.
Conway Morris and Peel largely accepted Butterfield's arguments and treated Wiwaxia as an ancestor or "aunt" of the polychaetes. However they maintained that Wiwaxia's feeding apparatus was much more like a molluscan radula. They also argued that Wiwaxia was fairly closely related to and in fact descended from the halkieriids, as the sclerites are divided into similar groups, although those of halkieriids were much smaller, more numerous and hollow. They wrote that in 1994 Butterfield had found Wiwaxia sclerites that were clearly hollow. The cladogram they presented showed the halkieriids split into three groups: one as "aunts" of brachiopods, animals whose modern forms have bivalve shells but differ from molluscs in having muscular stalks and a distinctive feeding apparatus, the lophophore; the second group of halkieriids as "aunts" of both Wiwaxia and annelid worms; and the earliest halkieriids as "great aunts" of all of these.
Marine biologist Amélie H. Scheltema et al. argued that Wiwaxia’s feeding apparatus is very similar to the radulas of some modern shell-less aplacophoran molluscs, and that the sclerites of the two groups are very similar. They concluded that Wiwaxia was a member of a clade that includes molluscs.
Danish zoologist Danny Eibye-Jacobsen regarded bristles as a feature shared by molluscs, annelids and brachiopods. Hence even if Wiwaxia’s sclerites closely resembled bristles, which he doubted, this would not prove that Wiwaxia’s closest relative were annelids. He also pointed out that the very different numbers of sclerites in the various zones of Wiwaxia’s body do not correspond to any reasonable pattern of segmentation. Since in his opinion Wiwaxia lacked other clearly polychaete features, he regarded this as an argument against classifying Wiwaxia as a polychaete. In his opinion there were no strong grounds for classifying Wiwaxia as a proto-annelid or a proto-mollusc.
Caron, Scheltema et al., in a paper about newly discovered and good specimens of Odontogriphus, thought Wiwaxia bore little resemblance to polychaetes as it showed no signs of the segmentation, appendages in front of the mouth, or "legs" – all of which are typical polychaete features. On the other hand they thought the teeth on the feeding apparatus of both Wiwaxia and Odontogriphus strongly resembled those of a modern group of molluscs, Neomeniomorpha. Hence they classified Wiwaxia as an "aunt" and Odontogriphus as a "great aunt" of molluscs.
Butterfield returned to the debate in 2006, repeating the arguments he presented in 1990 for regarding Wiwaxia as an early polychaete and adding that, while bristles are a feature of several groups, they appear as a covering over the back only in polychaetes. He also questioned whether the two or three rows of teeth in Wiwaxia and Odontogriphus could function in the same way as the belt-like molluscan radula with its many tooth-rows; on the other hand one group of polychaetes uses a single two-part "jaw" to scrape food off a substrate. While he agreed that Wiwaxia and Odontogriphus were very similar, he wrote, "… it certainly does not require that they be shoehorned into the same lineage. … the seemingly trivial distinction between these two taxa is exactly what is expected at the divergence points leading from a last common ancestor to extant phyla."
Butterfield accepted that Wiwaxia and Odontogriphus were closely related, but argued that they were stem-group polychaetes rather than stem-group molluscs. In his opinion the feeding apparatus of these organisms, with consisted of two or at most four rows of teeth, could not perform the functions of the "belt-like" molluscan radula with their numerous tooth-rows; the different tooth-rows in both Wiwaxia and Odontogriphus tooth-rows also have noticeably different shapes, while those of molluscan radulae are produced one after the other by the same group of "factory" cells and therefore are almost identical. He also regarded lines running across the middle region of Odontogriphus fossils as evidence of external segmentation, since the lines are evenly spaced and run exactly at right angles to the long axis of the body. As in his earlier papers, Butterfield emphasized the similarities of internal structure between Wiwaxia’s sclerites and the bristles of polychaetes, and the fact that polychaetes are the only modern organisms in which some of the bristles form a covering over the back.
In 2007 Conway Morris and Caron described a new fossil, Orthrozanclus, which had a mineralized shell like that of halkieriids, and unmineralized sclerites and long spines like those of Wiwaxia – and, like both of these, a soft underside which they interpreted as a muscular foot, and a similar arrangement of the sclerites into three concentric bands. Some of Orthrozanclus’s sclerites appear to have been hollow, as the specimen includes what look like internal castings. They took this find as evidence that the halkieriids, Wiwaxia and Orthrozanclus were very closely related and formed the group "halwaxiids". However the simplest "family tree" faces an obstacle: the siphogonuchitids appear in earlier rocks and had mineralized sclerites. Hence Conway Morris and Caron found it necessary to consider two more complex family trees, concluding that "Hypothesis 1" fitted the available data better, but fell apart if there were minor changes in the characteristics used:
  1. Odontogriphus, which had no sclerites or shells, was an "aunt" of the molluscs and the halwaxiids were a "sister" group to molluscs. Wiwaxia, with only unmineralized armor, was a surviving member of the most primitive halwaxiids. The siphogonuchitids developed mineralized sclerites but grouped into a simpler "chain mail" arrangement. Halkieriids then regained the more complex three-band "chain mail" arrangement and also developed shells. Orthrozanclus lost one of the shells and its sclerites lost the mineralization. Because the siphogonuchitids appear in this family tree between Wiwaxia and the other halwaxiids, the halwaxiids are not monophyletic in this tree.
  2. In "Hypothesis 2" the halwaxiids are monophyletic and the siphogonuchitids are their nearest non-halwaxiid relatives. The halwaxiids plus siphogonuchitids form a "sister" group to both the annelids and brachiopods, and the molluscs are the "aunts" of all of them.
Conway Morris and Caron published the first description of Orthrozanclus reburrus. This resembled the halkieriids in having concentric bands of sclerites, although only two and not mineralized; and one shell at what was presumed to be the front and which was similar in shape to Halkieria’s front shell. It also had long spines rather like those of Wiwaxia. Conway Morris and Caron regarded this creature as evidence that the "halwaxiids" were a valid taxon and were monophyletic, in other words shared a common ancestor with each other and with no other organism. They published two cladograms, representing alternative hypotheses about the evolution of the lophotrochozoa, the lineage that includes molluscs, annelids and brachiopods:
  1. This is the more likely, although it falls apart if the organisms' characteristics are changed even slightly:
  2. *Kimberella and Odontogriphus are early, primitive molluscs, without sclerites or any kind of mineralized armor.
  3. *Wiwaxia, the siphogonotuchids, Orthrozanclus and Halkeria form a side-branch of the mollusc family tree, which diverged in that order. This would mean that: Wiwaxia was the first of them to have sclerites, which were unmineralized; the siphogonotuchids were the first to have mineralized sclerites, although the scleritome was simpler; halkieriids then develop more complex scleritomes, while in Orthrozanclus the scleritome became unmineralized again and the rear shell vanished or became so small that it has not been seen in fossils. This hypothesis faces the difficulty that siphogonotuchids appear in earlier rocks and have simpler scleritomes than the other three groups.
  4. *The annelids and brachiopods evolved from the other main branch of the family tree, which did not include the molluscs.
  5. The alternative view is:
  6. *Kimberella and Odontogriphus are early, primitive lophotrochozoans.
  7. *The siphogonotuchids, Halkeria, Orthrozanclus and Wiwaxia form a group that is closer to the shared ancestor of annelids and brachiopods than it is to the molluscs. The siphogonotuchids are the first of the group to become distinctive, with two types of mineralized sclerites and a "shell" made of fused sclerites. Halkieriids had three types of sclerites and two one-piece shells. In Orthrozanclus the sclerites became unmineralized and in Wiwaxia the shells were lost.

    Halkieriids and siphogonotuchids

Most of the debate has been about halkieriids' relationship to molluscs, annelids and brachiopods. However a paper published in 2008 argued that halkieriids were closely related to chancelloriids, which have been generally regarded as sponges.

Relationship to Molluscs, Annelids and Brachiopods

Halkieriid sclerites were known for a long time as elements of the small shelly fossils. The first articulated specimens, with all their hard parts together, were collected in 1989 and were described briefly in 1990 by Conway Morris and Peel. The same authors produced a more detailed analysis in 1995, which divided the halkieriids into three groups: one as "aunts" of brachiopods, animals whose modern forms have bivalve shells but differ from molluscs in having muscular stalks and a distinctive feeding apparatus, the lophophore; the second group of halkieriids as "aunts" of both Wiwaxia and annelid worms; and the earliest halkieriids as "great aunts" of all of these.
In 2003 Cohen, Holmer and Luter supported the halkieriid-brachiopod relationship, suggesting that brachiopods may have arisen from a halkieriid lineage that developed a shorter body and larger shells, and then folded itself and finally grew a stalk out of what used to be the back.
Vinther and Nielsen proposed instead that Halkieria was a crown group mollusc, in other words more similar to modern molluscs than to annelids, brachiopods or any intermediate groups. They argued that: Halkieria’s sclerites resembled those of the modern solenogaster aplacophoran shell-less molluscs, of some modern polyplacophoran molluscs, which have several shell plates, and of the Ordovician polyplacophoran Echinochiton; Halkieria’s shells are more similar to the shells of conchiferan molluscs, since shells of both of these groups show no trace of the canals and pores seen in polyplacophoran shell plates; the bristles of brachiopods and annelids are similar to each other but not to Halkieria’s sclerites.
However Conway Morris criticized Vinther and Nielsen's classification of Halkieria as a crown group mollusc, on the grounds that the growth of the spicules in the aplacophorans and polyplacophorans is not similar to the method of growth deduced for the complex halkieriid sclerites; in particular, he said, the hollow spines of various molluscs are not at all like the halkieriid sclerites with their complex internal channels. Conway Morris repeated his earlier conclusion that halkieriids were close to the ancestors of both molluscs and brachiopods.
In their description of the newly discovered Orthrozanclus, which has similarities to both Wiwaxia and the halkieriids, Conway Morris and Caron also took account of the siphogonuchitids, a group known only from isolated mineralized sclerites that resemble those of the halkieriids. They proposed the two "family trees" described above:
  1. The halkieriids formed part of a "sister" group to the molluscs. This hypothesis implies that the halwaxiids were not monophyletic, since the siphogonuchitids appear between Wiwaxia and the halkieriids. Nevertheless, Conway Morris and Caron found that this phylogeny fitted the available data better, although it was not robust.
  2. The siphogonuchitids plus the halwaxiids formed a "sister" group to the annelids and brachiopods, while the molluscs were "aunts" of all of these.

    Relationship to chancelloriids

Porter revived an early 1980s idea that the sclerites of Halkieria are extremely similar to those of chancelloriids. These were sessile, bag-like, radially symmetric organisms with an opening at the top.
Since their fossils show no signs of a gut or other organs, chancelloriids were originally classified as some kind of sponge. Butterfield and Nicholas argued that they were closely related to sponges on the grounds that the detailed structure of chancellorid sclerites is similar to that of fibers of spongin, a collagen protein, in modern keratose demosponges. However Janussen, Steiner and Zhu opposed this view on several grounds, including that: the silica-based spines of demosponges are solid, while chancellorid sclerites are hollow, probably made of aragonite, and filled with soft tissues connected to the rest of the animal at the bases of the sclerites; sponges have loosely bound-together skins called pinacoderms, which are only one cell thick, while the skins of chancellorids were much thicker and show signs of connective structures called belt desmosomes. In their opinion the presence of belt desmosomes made chancellorids members of the Epitheliazoa, the next higher taxon above the Porifera, to which sponges belong. They thought it was difficult to say whether chancellorids were members of the Eumetazoa, "true animals" whose tissues are organized into Germ layers: chancellorids' lack of internal organs would seem to exclude them from the Eumetazoa; but possibly chancellorids were descended from Eumetazoans that lost these features after becoming sessile filter-feeders.
The sclerites of halkieriids and chancelloriids resemble each other at all levels: both have an internal "pulp cavity" and a thin external organic layer; the walls are made of the same material, aragonite; the arrangement of the aragonite fibers in each is the same, running mainly from base to tip but with each being closer to the surface at the end nearest the tip. It is extremely improbable that totally unrelated organisms could have developed such similar sclerites independently, but the huge difference in the structures of their bodies makes it hard to see how they could be closely related. This dilemma may be resolved in various ways: