ADVERTISEMENT
  About the SA Blog Network













Tetrapod Zoology

Tetrapod Zoology


Amphibians, reptiles, birds and mammals - living and extinct
Tetrapod Zoology Home

The Tet Zoo guide to mesosaurs

The views expressed are those of the author and are not necessarily those of Scientific American.


Email   PrintPrint



Speculative life reconstruction of Mesosaurus tenuidens, by Darren Naish.

A small group of long-snouted swimming reptiles from the Permian of Brazil, Uruguay, South Africa and Namibia – the mesosaurs – represent the oldest amniote group known to have taken to life in the marine realm. Do not confuse them with mosasaurs, a group of large to gigantic swimming lizards from the Cretaceous, usually considered close relatives of monitors and gila monsters.

Mesosaurs were small, with total lengths of 30-100 cm. The presence of these small swimming reptiles on both sides of the Atlantic was famously used to help support the existence of continental drift, since their distribution seemed to show that Africa and South America had once been in contact (Du Toit 1927).

A long, deep tail, elongate feet (evidence of webbing is preserved in some specimens (Rossmann & Maisch 1999)), thickened, usually dense-boned ribs, numerous fine, needle-like teeth set within elongate jaws and retracted nares indicate that mesosaurs were aquatic predators of small, krill-like crustaceans or similar prey. They probably used the long, flexible neck to sweep their jaws from side to side (Modesto 2006).

Famous skeleton reconstruction of Mesosaurus, produced by McGregor in 1908 but still essentially accurate.

It has been implicated on a few occasions that mesosaurs may have been suspension-feeders but in fact the tooth spacing seen in these reptiles does not support this interpretation. On this subject, Collin & Janis (1997) wrote “[interpretation of mesosaurs as suspension-feeders] depends on an erroneous interpretation of the mandibular teeth as small marginal upper teeth, and a more detailed examination of mesosaur functional morphology suggests that they probably captured individual prey selectively rather than processing large volumes of water nonselectively as to true suspension feeders (S. Modesto, personal communication)” (p. 453).

Excellent skeleton of a mesosaur from Michael Benton's 1990 book The Reign of the Reptiles. The specimen is identified therein as a Mesosaurus but I think it might actually be a Brazilosaurus.

Signore et al. (2002) reported new details on the skull morphology and possible lifestyle of the mesosaur Stereosternum. They noted that snout shape in this taxon recalled that of the extant crocodylians Gavialis and Tomistoma and they therefore proposed that, like these crocodylians, Stereosternum was piscivorous. They also noted that a juvenile Stereosternum had a shorter snout and limbs less suited for aquatic propulsion than adults and therefore suggested that juveniles may have been less aquatic than adults.

Within recent decades, just three mesosaur taxa have conventionally been recognised: Mesosaurus tenuidens Gervais, 1865, Stereosternum tumidum Cope, 1885 and Brazilosaurus sanpauloensis Shikama & Ozaki, 1966. However, several others have been named over the years. M. pleurogaster Seeley, 1892 has generally been regarded as a synonym of M. tenuidens but still requires evaluation (Modesto 1996). Noteosaurus africanus Broom, 1913 – named for an articulated hindlimb, pelvis and partial tail from the Dwyka Formation of South Africa – was argued by Modesto (1996) to be indistinguishable from either Mesosaurus or Stereosternum and hence a nomen dubium while Ditrochosaurus capensis Gürich, 1889 from Namibia – at times referred to as Mesosaurus capensis – was regarded by Rossmann & Maisch (1999) as another junior synonym of M. tenuidens. It seems that Stereosternum and Mesosaurus are sister-taxa.

Cranial reconstruction of Mesosaurus, from Piñeiro et al. (2012a). Note the narial obturator foramen and lower temporal fenestra.

Mesosaurs have usually been regarded as anapsid – that is, as lacking bony openings in the postorbital or temporal region of the skull. However, a lower temporal fenestra is present in Mesosaurus at least where it’s bordered by the jugal, quadratojugal and squamosal bones (Piñeiro et al. 2012a). Since this is similar to the condition present in synapsids it might show that lower temporal fenestration is primitive for Amniota (the more popular view is that the anapsid condition is primitive for Amniota). It’s also possible, however, that fenestration appeared and disappeared several times independently within this section of the cladogram. Brazilosaurus appears to lack temporal fenestration (Rossmann 2002).

Skull of Brazilosaurus sanpauloensis, from Rossman (2002). Note the relatively short teeth and robust rostrum compared to Mesosaurus. Also, no temporal fenestration. Click to enlarge.

Mesosaurs have usually been imagined as animals of coastal habitats but sedimentological evidence from Uruguay shows that at least some populations of Mesosaurus inhabited hypersaline, lagoon-like habitats where the only apparent prey items were pygocephalomorph crustaceans (Piñeiro et al. 2012b). A need to discharge excess salt may explain the presence of the so-called narial obturator foramen, a unique opening located just posterior to the naris. An oval region on the palate, adjacent to the choana, has also been identified as a possible salt gland. Small, spike-like teeth are present across part of the palate. Possible direct evidence for diet comes from a Brazilosaurus specimen that has two large, ovoid structures preserved within the stomach region (Rossmann 2002), though exactly what they are is unknown. Brazilosaurus has far shorter, stouter teeth than Mesosaurus and was evidently doing something very different.

The phylogenetic position of mesosaurs has proved difficult to resolve: recent proposals place them within Reptilia and near the origin of Parareptilia (and, variously, as the sister-group to remaining parareptiles, or the sister-group to the whole of Parareptilia). There’s no good evidence linking them with the superficially similar ichthyosaurs, hupehsuchians or thalattosaurs: members of all of those groups possess skull bone characters nesting them deep within Diapsida, and mesosaurs emphatically lack such characters.

Reconstructed life appearance of embryonic mesosaur from Mangrullo Formation of Uruguay, preserved in isolation on a slab. From Piñeiro et al. (2012c).

We have very little direct data on mesosaur palaeobiology, though one particularly exciting discovery was made very recently and got the international news coverage that it deserved. I’m referring to the discovery of mesosaur babies. One of these (discovered in the Brazilian Iratí Formation) was found inside the body of an adult and hence should be referred to as an embryo; many others (26 in total, all from the Mangrullo Formation of Uruguay) were found outside the bodies of adults, but often preserved in close association with them (Piñeiro et al. 2012c). These might be embryos that became aborted or otherwise disassociated with their mother’s bodies post-mortem, or they might be new babies that died in close association with their parents. There’s no evidence for cannibalism on the juveniles, and a statistical test showed that the association between the babies and adults is not random (Piñeiro et al. 2012c).

These finds show that mesosaurs retained a low number of babies (one or two) in the body for an extended period of time; however, it’s not possible to say from these finds whether live babies or shelled eggs were produced (Piñeiro et al. 2012c). If eggs were laid, we don’t know whether they were laid on land, or in shallow water. The latter possibility exists because there are extant pleurodire turtles that do it. But if the juveniles found alongside the adults in the Mangrullo Formation represent free-swimming juveniles rather than embryos, do we have evidence for some sort of parental care? Such a possibility is certainly consistent with the low number of offspring seemingly produced by these animals; at the moment, however, saying anything further would be speculation.

And… with mesosaurs out of the way, all I have to do now is get through all the other fossil marine reptile groups. Don’t hold your breath.

For other Tet Zoo articles relevant to some of the subjects covered here, see…

Refs – -

Collin, R. & Janis, C. M. 1997. Morphological constraints on tetrapod feeding mechanisms: why were there no suspension-feeding marine reptiles? In Callaway, J. & Massare, J. (eds) Ancient Marine Reptiles. Academic Press (London), pp. 451-466.

Du Toit, A. L. 1927. A geological comparison of South America with South Africa. Carnegie Institute Publications, Washington 381, 1-158.

Modesto, S. P. 1996. Noteosaurus africanus Broom is a nomen dubium. Journal of Vertebrate Paleontology 16, 172-174.

- . 2006. The cranial skeleton of the Early Permian aquatic reptile Mesosaurus tenuidens: implications for relationships and palaeobiology. Zoological Journal of the Linnean Society 146, 345-368.

Piñeiro, G., Ramos, A., Goso, C., Scarabino, F. & Laurin, M. 2012b. Unusual environmental conditions preserve a Permian mesosaur-bearing Konservat-Lagerstätte from Uruguay. Acta Palaeontologica Polonica 57, 299-318.

- ., Ferigolo, J., Ramos, A. & Laurin, M. 2012a. Cranial morphology of the Early Permian mesosaurid Mesosaurus tenuidens and the evolution of the lower temporal fenestration reassessed. Comptes Rendus Palevol 11, 379-391.

- ., Ferigolo, J., Meneghel, M. & Laurin, M. 2012c. The oldest known amniotic embryos suggest viviparity in mesosaurs. Historical Biology 24, 620-630.

Rossmann, T. 2002. Studien an Mesosauriern (Amniota inc. sed., Mesosauridae): 3. Neue Aspekte zur Anatmie, Erhaltung und Paläoökologie aufgrund der Exemplare im Paläontologischen Institut der Universität Zurich. Neues Jahrbuch fur Geologie und Paläontologie, Abhandlungen 224, 197-221.

- . & Maisch, M. W. 1999. Das Mesosaurier-Material in der Bayerischen Staatsammlung für Paläontologie und Historische Geologie: Überischt und neue Erkenntnisse. Mitteilungen der Bayerischen Staatssammlung für Paläontologie und Historische Geologie 39, 69-83.

Signore, M., Raya, P. & Barbera, C. 2002. New specimens of Stereosternum tumidum (Mesosauria) and notes on its palaeoecology. In Norman, D. & Upchurch, P. (eds) SVPCA 50, Cambridge 2002, Abstract Volume. University of Cambridge, p. 38.

Darren Naish About the Author: Darren Naish is a science writer, technical editor and palaeozoologist (affiliated with the University of Southampton, UK). He mostly works on Cretaceous dinosaurs and pterosaurs but has an avid interest in all things tetrapod. His publications can be downloaded at darrennaish.wordpress.com. He has been blogging at Tetrapod Zoology since 2006. Check out the Tet Zoo podcast at tetzoo.com! Follow on Twitter @TetZoo.

The views expressed are those of the author and are not necessarily those of Scientific American.





Rights & Permissions

Comments 44 Comments

Add Comment
  1. 1. ezequielvera 1:17 pm 07/3/2013

    Great post, as always Darren!
    Do you know if there are other characters that link (or may link) mesosaurs with synapsids? or otherwise with diapsids? I’m asking this because I feel that mesosaurs look more like diapsids or parareptilian than synapsids… (that’s a deduction based on all my knowledge of fossil plants… thus not very strict nor academical). Is there any possibility that mesosaurs are diapsids, but upper temporal fenestra is lost (such as the antorbital fenestra in Crocodylus)?

    Link to this
  2. 2. SWestfall 2:56 pm 07/3/2013

    Somehow I knew you were going to have to put in the disclaimer about mosasaurs, but I didn’t know how soon you’d put it in the post.

    Link to this
  3. 3. naishd 6:18 pm 07/3/2013

    ezequielvera: thanks for the comment. Mesosaurs lack the derived characters of synapsids and indeed of eureptiles (the great clade that includes diapsids and some other reptile lineages). However, they do have characters that ally them with (or place them within) Reptilia, including a single coronoid (as opposed to two: the condition present in Synapsida, diadectomorphs and other non-amniotes), a slender femoral shaft (that of synapsids and non-amniotes is comparatively short and broad), a supinator process that’s parallel to the shaft of the humerus (rather than angled to it), and so on.

    The question is… are mesosaurs outside of the parareptile + eureptile clade (in which case they’re stem-reptiles, or – if you want to restrict the name Reptilia to the parareptile + eureptile clade – non-reptile sauropsids), or are they members of Parareptilia? They resemble parareptiles more than eureptiles in having an especially short postorbital region and in lacking a caniniform region in the jaws: on the basis of these characters and others, some workers find mesosaurs to be within Parareptilia, perhaps being the sister-group to the rest of the clade (and the presence of temporal fenestration doesn’t matter, since we now know that this is all over the place within Parareptilia).

    TLDR version: no, they lack diapsid characters. They’re either stem-reptiles or parareptiles.

    Darren

    Link to this
  4. 4. AlHazen 6:42 pm 07/3/2013

    There’s a dark patch near the crown (above the temporal fenestra in the view shown) on the Mesosaurus skull restoration shown: what is it? Just, as an amateur, looking at the picture, I would have guessed it was an upper temporal fenestra (the u.t.f. is near the top of the skull, more easily seen from above than from the side, in many archosaur skulls — crocodilian etc– that I’ve seen) and thought “Aha! a diapsid!”, but from what you say that can’t be right!

    Link to this
  5. 5. ectodysplasin 6:49 pm 07/3/2013

    @darren,

    Be aware that there are ‘parareptiles’ with as many as three coronoids, and that reduction of coronoids is also seen in various ‘microsaur’ taxa.

    In addition, basalmost diapsids lack a caniniform region as well, but some ‘microsaurs’ have one. Some parareptiles may have something approaching a caniniform region. Diadectamorphs have canines.

    Ultimately, though, I think there’s a lot of instability in that part of the tree and it’s going to get progressively worse as we better-characterize a lot of these animals.

    Link to this
  6. 6. ectodysplasin 7:47 pm 07/3/2013

    Quoting myself….

    Ultimately, though, I think there’s a lot of instability in that part of the tree and it’s going to get progressively worse as we better-characterize a lot of these animals.

    Not to imply that the phylogenetic problem won’t be ultimately solvable (I think it will be) but rather that there’s a lot more work ahead of us before it is, and that there may be surprises in store for us all.

    Link to this
  7. 7. Halbred 7:50 pm 07/3/2013

    Oooh, hupehsuchids! One of my favorite bizarre marine reptile groups. They’re not especially close to ichthyosaurs, are they, despite being superficially similar?

    (actually not all that similar)

    Link to this
  8. 8. naishd 7:59 pm 07/3/2013

    AlHazen (comment # 4): yes, it looks like there’s a small, dorsally sited foramen in that diagram. I’m not sure what it’s meant to be – there’s no corresponding structure in the fossils. I mean, there’s a big pineal foramen, but nothing in between that and the lower temporal fenestra. Actually, I wonder if it’s meant to be the most lateral part of the pineal foramen?

    ectodysplasin (comment # 5): yes, good call… character distribution is horribly complex when we look at enough of it. The characters we’re talking about here – number of coronoids, presence/absence of a caniniform region – are obviously subject to a fair bit of convergence (cf, your mention of microsaurs); what’s happening when these characters are used as synapomorphies is that people are only looking at the ‘most basal’ taxa in their phylogenies (examples: parareptiles with three coronoids [did you have Belebey in mind?] are deeply nested within Parareptilia; and, sure, early diapsids might lack a caniniform region, but the non-diapsid eureptile Paleothyris does not). Plus, I suppose stuff becomes out of date pretty quickly…

    Darren

    Link to this
  9. 9. naishd 8:23 pm 07/3/2013

    Halbred (comment # 7): no no no, hupehsuchians are another group entirely; superficially similar, but very different in detail and not close to mesosaurs. I’ll cover hupehsuchians some other time…

    Darren

    Link to this
  10. 10. darkgabi 5:10 am 07/4/2013

    oh, *that* part of the amniote tree.. no one cares, it’s a mess, outdated. i have some plans, i just need the funding :p

    ok, my plans start slightly higher in the tree, but that area, including also early synapsids, is very interesting!

    Link to this
  11. 11. naishd 5:15 am 07/4/2013

    darkgabi: “it’s a mess”? Aww, come on, that will upset quite a few people who have worked hard on this part of the tree :) Have you seen (for example)…

    DeBraga, M. & Rieppel, O. 1997. Reptile phylogeny and the interrelationships of turtles. Zoological Journal of the Linnean Society 120, 281-354.

    Laurin, M. & Reisz, R. R. 1995. A reevaluation of early amniote phylogeny. Zoological Journal of the Linnean Society 113, 165-223.

    Tsuji, L. A. & Müller, J. 2009. Assembling the history of the Parareptilia: phylogeny, diversification, and a new definition of the clade. Fossil Record 12, 71-81.

    Darren

    Link to this
  12. 12. David Marjanović 6:55 am 07/4/2013

    Brazilosaurus appears to lack temporal fenestration (Rossmann 2002).

    That reconstruction looks odd, though. There’s no trace of the quadratojugal, the shape of the squamosal is majorly weird… Is there a photo anywhere?

    If eggs were laid, we don’t know whether they were laid on land, or in shallow water. The latter possibility exists because there are extant pleurodire turtles that do it.

    what

    what is this I don’t even

    Be aware that there are ‘parareptiles’ with as many as three coronoids

    Which ones?

    But in any case, what about the “supradentary” bone of various theropods that lies exactly where you’d expect the coronoid II?

    In addition, basalmost diapsids lack a caniniform region as well, but some ‘microsaurs’ have one. Some parareptiles may have something approaching a caniniform region. Diadect[o]morphs have canines.

    Yeah, that’s one of the more labile characters in general. The varanopids Archaeovenator and Pyozia have no trace of a caniniform region…

    It still carries phylogenetic signal, though.

    AlHazen (comment # 4): yes, it looks like there’s a small, dorsally sited foramen in that diagram. I’m not sure what it’s meant to be – there’s no corresponding structure in the fossils. I mean, there’s a big pineal foramen, but nothing in between that and the lower temporal fenestra. Actually, I wonder if it’s meant to be the most lateral part of the pineal foramen?

    I’m sure it’s supposed to be the pineal foramen, and the skull isn’t in perfect lateral view.

    darkgabi: “it’s a mess”? Aww, come on, that will upset quite a few people who have worked hard on this part of the tree :) Have you seen (for example)…

    She has, I know who she is. She has very good reasons for saying it’s a mess.

    Link to this
  13. 13. naishd 7:12 am 07/4/2013

    Thanks, David. On the “it’s a mess” thing – - I know it wasn’t made as a fully serious comment, and I know that there’s still a lot to do, that we’re bound to be surprised by emerging topologies etc, but saying stuff like this is often disingenuous, for two reasons. One is that it’s not really true: there are, as always, disagreements/conflicts over the positions of given taxa and given lineages, but the general shape of the tree doesn’t differ >that much< between studies. Nobody really finds, for example, a Peteresque topology where we need to throw our arms up and start from scratch. Two is that it implies that we know less than we do, and recalls statement made by nay-sayers like Feduccia and Martin that we've gotten everything wrong and can't have any confidence about published results.

    (And, by the way, I know who darkgabi is too. Encouraged her on twitter [and, I think, facebook] to come comment here.]

    Darren

    Link to this
  14. 14. David Marjanović 1:23 pm 07/4/2013

    “Nobody really finds, for example, a Peteresque topology where we need to throw our arms up and start from scratch.”

    Define “really”… ectodysplasin hasn’t published (beyond an SVP meeting abstract) but has found that kind of tetrapod phylogeny. Personally, I’m highly skeptical and expect that convergence between bottom-dwelling ambush hunters will be to blame, but…

    Link to this
  15. 15. ectodysplasin 1:51 pm 07/4/2013

    @Darren,

    >ectodysplasin (comment # 5): yes, good call… character distribution is horribly complex when we look at enough of it. The characters we’re talking about here – number of coronoids, presence/absence of a caniniform region – are obviously subject to a fair bit of convergence (cf, your mention of microsaurs); what’s happening when these characters are used as synapomorphies is that people are only looking at the ‘most basal’ taxa in their phylogenies (examples: parareptiles with three coronoids [did you have Belebey in mind?] are deeply nested within Parareptilia; and, sure, early diapsids might lack a caniniform region, but the non-diapsid eureptile Paleothyris does not). Plus, I suppose stuff becomes out of date pretty quickly…

    Well, the question is whether this indicates a problem of rampant convergence or whether this indicates a problem with the phylogeny more broadly. You mentioned ‘microsaurs’ as a source of potential convergence, but we’re not actually 100% certain that isn’t actually a problem with our phylogenies instead. So the presence of three coronoids in some ‘parareptiles’ (the taxon I’m thinking about has been presented in talks but I don’t think it has been published yet) may be a bizarre reversal….or it may actually mean that some ‘parareptiles’ aren’t amniotes at all, and instead belong further down the tree. The absence of a caniniform region in various amniotes could mean that the amniote LCA lacked a caniniform region…or it could mean that multiple lineages lost the caniniform region, and that a heterodont dentition with an incisiform, caniniform, and postcanine region is ancestral for a group including animals like Limnoscelis and ‘microsaurs’ (or at least ‘recumbirostrans’). How we interpret the evolution of any of these characteristics depends inherently on the phylogeny of near-amniotes.

    At the moment, ‘parareptile’ phylogenies seem to be pretty stable, but from talking with some of the folks currently working on the group, some key areas of the phylogeny are not trustworthy. It’s the same sort of problem we’ve got with ‘lepospondyls’…the trees all seemed to support monophyly (or at least paraphyly) but various pre-phylogenetic workers have repeatedly suggested that these groups are broadly polyphyletic. New data favors the latter among ‘lepospondyls’ (will you be at ICVM? I’ll be talking about lysorophians) and it will be interesting to see how this plays out with ‘parareptiles’ and other early amniotes. There’s some extensive work forthcoming on stem-reptiles and stem-amniotes including anatomical revisions of Protoclepsydrops, Coelostegus, Brouffia, and Hylonomus as well as some new phylogenetic work on the group, so I expect there will be some implications further down the tree as well.

    Link to this
  16. 16. ectodysplasin 2:15 pm 07/4/2013

    @david

    Define “really”… ectodysplasin hasn’t published (beyond an SVP meeting abstract) but has found that kind of tetrapod phylogeny. Personally, I’m highly skeptical and expect that convergence between bottom-dwelling ambush hunters will be to blame, but…

    :)

    The Diploceraspis stuff is currently moving slowly (other priorities at the moment), but aistopods and lysorophians are on schedule to enter review this calendar year, and we’ll be talking about both at SVP this year.

    I wouldn’t call our results “Peters-esque” because we’re finding relationships that have previously been suggested by workers like Smithson and Carroll (albeit without the aid of cladistics), but the results we’re finding are a radical departure from the current state of early tetrapod phylogeny in the literature. In addition, whereas the Peters trees are the result of swamping out phylogenetic signal by dropping in hundreds of distantly-related taxa and hoping that a handful of generalized gestalt characters can piece things out, we’re adding a lot of robust endocranial data to existing datasets with reasonable taxonomic sampling, and our new results seem to hinge directly on the strength of those new characters.

    Link to this
  17. 17. ectodysplasin 2:28 pm 07/4/2013

    @david,

    Yeah, that’s one of the more labile characters in general. The varanopids Archaeovenator and Pyozia have no trace of a caniniform region…

    Yes. Although Archaeovenator may be a juvenile and Pyozia is considered by some to be a parareptile, possibly a millerettid.

    Link to this
  18. 18. ectodysplasin 3:23 pm 07/4/2013

    @darren,

    Thanks, David. On the “it’s a mess” thing – – I know it wasn’t made as a fully serious comment, and I know that there’s still a lot to do, that we’re bound to be surprised by emerging topologies etc, but saying stuff like this is often disingenuous

    This depends on which portion of the tetrapod tree you’re working in. Phylogenetic work tends to undergo a maturation process; early phylogenetic work on a group recovers bizarre topologies that fly in the face of previous comparative research, and then over time, addition of new data, better treatment of data, and better methods eventually leads to a stable and robust result that makes some degree of both morphological and molecular sense. This is the case for the progression of molecular phylogenetics (see, for example, the early allozyme trees, holy crap were those bad), and also seems to be the case in well-studied morphological sets such as theropods.

    With early tetrapods, we’re not looking at a mature dataset. We really have two basic phylogenetic datasets; we have a matrix that was established by Clack and Panchen and has been developed by Clack, Ahlberg, and Klembara, and that matrix emphasizes stem tetrapods, anthracosaurs, baphetids, and seymouriamorphs, with very limited sampling of more diverse groups. We also have a matrix that was developed by Laurin and Reisz that has subsequently been developed in diverging ways by several working groups, with diverging results, but largely emphasizing lepospondyls with limited sampling of greater early tetrapod diversity. Ruta has done some work synthesizing these datasets, but the results are still sort of jumbled. All of these matrices recover somewhat different results, and these results typically have very weak support and very, very, very high rates of convergence (CI < 20% in some cases). These datasets also typically exclude more robust types of anatomical data (e.g. braincase) and there's a lot of character nonindependence (which David will agree with).

    So whereas someone saying that theropod phylogeny (which is relatively mature) is "a mess" would be disingenuous, describing early tetrapod phylogeny as such is a different story. In early tetrapod phylogeny, there's still room for legitimate debate on things like lepospondyl monophyly, lissamphibian origins, and even things like timing of the origin of the tetrapod crown and of the lissamphibian crown, and that's what we're seeing in the literature right now. This debate includes description of new data and new types of data (e.g. Hillary Maddin's extensive work on caecilians), description of new species in phylogenetically interesting places in the tree (e.g. Gerobatrachus, but also things like the new ‘microsaur’ we published a few months ago), and use of new methods (e.g. CT) to revise anatomy of known taxa, as well as cleaning up logically nonindependent characters from the matrix, accounting for ontogenetic variation in the specimens being studied, correcting for functionally nonindependent characters, and so on.

    So what I’d say is that saying “phylogeny of x is a mess” is not disingenuous…it’s often a way of saying that the phylogeny of a taxon is not yet mature and that topologies can and will change as the datasets mature. However, this language is sometimes hijacked by disingenuous people with ulterior motives when questioning portions of phylogeny that are significantly more mature (e.g. the BAND people) or by people who are more enamored with the idea of discovering new topologies than with putting in the legwork to actually contribute meaningfully to the discussion (e.g. Peters).

    Link to this
  19. 19. naishd 4:53 pm 07/4/2013

    Thanks for awesome comments. I should cover stem-taxa more often… not enough discussion of this sort happening online (right? RIGHT?).

    WRT “However, this language is sometimes hijacked by disingenuous people with ulterior motives” (comment # 18): yup, that’s one of the two points I was trying to make.

    Sadly, I’m a no-show for ICVM this year, but I will be at SVP and will try and make your talk (depends on those goddam concurrent sessions, of course… I have to prioritise the Mesozoic).

    The long comments are much appreciated, I’m sure the general readership feels the same.

    Darren

    Link to this
  20. 20. Therizinosaurus 5:05 pm 07/4/2013

    “There’s a dark patch near the crown (above the temporal fenestra in the view shown) on the Mesosaurus skull restoration shown: what is it?”

    It’s a depressed area on the lateral parietal. It’s only visible in some specimens, is never as obvious as in that illustration, and was not mentioned in the paper or Modesto’s (2006). Seems to be in the right spot for a supratemporal fenestra homolog though.

    “But in any case, what about the “supradentary” bone of various theropods that lies exactly where you’d expect the coronoid II?”

    In well preserved specimens, the supradentary is fused with the coronoid. Whether this means only one coronoid is present (albeit expanded) or that two ancestral coronoid elements are present and fused, I don’t know.

    “One is that it’s not really true: there are, as always, disagreements/conflicts over the positions of given taxa and given lineages, but the general shape of the tree doesn’t differ that much between studies. Nobody really finds, for example, a Peteresque topology where we need to throw our arms up and start from scratch.”

    I think this could VERY easily be due to the fact everybody is using basically the same set of highly flawed characters (from Gauthier> Laurin and Reisz> deBraga and Rieppel) and few taxa. See this article at The Theropoda Database for example. Everything ectodysplasin says about basal tetrapods is true as well for basal amniotes. Not that I expect Peters’ topology to be valid at all, as he refuses to learn how to perform phylogenetic analyses, has a tiny matrix, uses fictional morphologies, etc., but he has a point that analyses of this area up till now have been pretty terrible. Have any even divided up Mesosauridae into genera?

    Link to this
  21. 21. ectodysplasin 5:13 pm 07/4/2013

    @darren,

    Thanks for awesome comments. I should cover stem-taxa more often… not enough discussion of this sort happening online (right? RIGHT?).

    Absolutely! Always happy to oblige.

    The problem with early tetrapods in general (not just the stem) is that there’s little material and few workers, and most of the current work is all foundational anatomical research and doesn’t really lend itself to more accessible discussions about ecology, behavior, and neat morphological specializations. There are a few really interesting discussions and debates (e.g. what sort of locomotion did the earliest terrestrial tetrapods have, how did early tetrapods specialize on diet, how did fossoriality evolve in early tetrapods, how did metamorphosis evolve, where do modern lissamphibians fall out in the tree, what were weirdos like diplocaulids doing with their bizarre anatomy) but by and large we’re looking at relatively conservative animals with relatively conservative morphology doing relatively conservative things. The differences between, say, baphetids, colosteids, and edopoids might be of interest to a specialist, but ecologically and functionally these are all essentially interchangeable animals (see the recent paper by Phil Anderson & colleagues in Integrative and Comparative Biology, for example).

    Sadly, I’m a no-show for ICVM this year, but I will be at SVP and will try and make your talk (depends on those goddam concurrent sessions, of course… I have to prioritise the Mesozoic).

    The stuff I’m covering is mainly anatomical data for specialists (with a little functional morphology discussion) but you should make a point of catching the aistopod talk if you can.

    Link to this
  22. 22. Yodelling Cyclist 6:24 pm 07/4/2013

    Just like to second Darren’s comment and say thank you for the detailed comments. Fascinating to read.

    Link to this
  23. 23. Yodelling Cyclist 12:23 pm 07/5/2013

    Aquatic amniotes….pictures of skulls……memory being awakened of a picture quiz from the dawn of time….

    Link to this
  24. 24. David Marjanović 12:43 pm 07/5/2013

    There’s some extensive work forthcoming on stem-reptiles and stem-amniotes including anatomical revisions of Protoclepsydrops, Coelostegus, Brouffia, and Hylonomus as well as some new phylogenetic work on the group

    *Homeric drool*

    This is the case for the progression of molecular phylogenetics (see, for example, the early allozyme trees, holy crap were those bad)

    Or indeed, much later, the “the guinea pig is not a rodent” trees that survived till 2001, where mice and rats (often separately!) were pulled out of Placentalia by long-branch attraction and sometimes pulled everything after them *coughkumarandhedgesninetyeightcough*.

    With early tetrapods, we’re not looking at a mature dataset. We really have two basic phylogenetic datasets;

    Actually, we have five off the top of my head:

    – Laurin 1994 (thesis) through Vallin & Laurin 2004; partly evaluated by Sigurdsen & Green 2011; Vallin & Laurin 2004 was used as the morphological part of the combined matrix of Pyron 2011;
    – McGowan 2002, which is ridiculously tiny, but nonetheless contains characters that the others all ignore; evaluated by checking every cell by me & Laurin 2008b;
    – Anderson 2001, expanded by Anderson 2007 and Anderson et al. 2008b; partly evaluated by me & Laurin 2009 and Sigurdsen & Green 2011; independent derivations that don’t take these into account by Maddin and coauthors 2011/2012 and Huttenlocker et al. 2013, and I have to read Olori’s thesis in more detail;
    – Ruta et al. 2003, independent derivations by Pawley 2006 (thesis), Ruta & Coates 2007 and Danto et al. 2013, derivations with eliminations and additions of taxa by Ruta, Clack, Klembara, Bolt and/or coauthors since 2006; evaluation of Ruta & Coates 2007 ongoing (Germain thesis 2008, me thesis 2010);
    – Carroll 2007 with its overwhelming number of correlated characters and even several OTUs that Carroll himself explicitly considers paraphyletic in the same paper.

    Table 1 of my 2013 Geodiversitas paper with Michel Laurin provides an overview over these analyses (before 2013) and their sizes, though we ignored all those without lissamphibians.

    We also have a matrix that was developed by Laurin and Reisz that has subsequently been developed in diverging ways by several working groups, with diverging results

    Really? Like what?

    and very, very, very high rates of convergence (CI < 20% in some cases).

    This is not a bad thing. It’s a sign of honesty*, and indeed a sign that the correlated characters aren’t all correlated with each other.

    The CI is well known to be size-dependent (more characters + more taxa = more opportunities for homoplasy); large theropod matrices always have a CI < 30 %. That’s why the RI and the RC have been developed.

    * There’s, uh, one dinosaur worker whose analyses from the 90s and early 00s had CIs < 80 %. That’s because he cherry-picked the characters. He has stopped doing that… though he has also largely stopped publishing phylogenetic analyses.

    These datasets also typically exclude more robust types of anatomical data (e.g. braincase)

    I’d even state it the other way around: they neglect everything except the dermal skull to an astonishing degree. Pawley’s thesis (2006) was an exception, but Pawley still hasn’t published.

    In well preserved specimens, the supradentary is fused with the coronoid.

    Always?

    “One is that it’s not really true: there are, as always, disagreements/conflicts over the positions of given taxa and given lineages, but the general shape of the tree doesn’t differ >that much Laurin and Reisz > deBraga and Rieppel) and few taxa.

    Something is missing here because of misparsed < and > signs. Use their HTML entities: &lt;, &gt;.

    Have any even divided up Mesosauridae into genera?

    Nope. The monophyly of Mesosauridae is obvious enough – and Brazilosaurus is so poorly known and described that the benefits of considering it separately are, alas, limited. :-(

    how did fossoriality evolve in early tetrapods

    And how many times… :-)

    The differences between, say, baphetids, colosteids, and edopoids might be of interest to a specialist, but ecologically and functionally these are all essentially interchangeable animals

    Heh. Their lower jaws perhaps! Colosteids were elongate and lacked, uh, “temporal embayments”, baphetids had those huge holes in the skull for, presumably, some reason, and how close Edops really is to the cochleosaurids isn’t quite obvious. (I need to go to Harvard and see Edops – the skull, the whole skull, and nothing but the skull, alas.)

    but you should make a point of catching the aistopod talk if you can

    I definitely will!

    Link to this
  25. 25. David Marjanović 12:45 pm 07/5/2013

    had CIs < 80 %

    Aaaargh, wrong HTML entity: they had CIs > 80 %, I think up to 89.

    Link to this
  26. 26. Andreas Johansson 3:06 pm 07/5/2013

    I should cover stem-taxa more often… not enough discussion of this sort happening online (right? RIGHT?).

    On the subject of stem anything, one frequently exasperating thing in popsci is that we’re told that a fossil critter is “shark” or “croc” or whatever and not whether those terms are being understood as crown groups or total groups or what.

    Oh, and also, I’m muchly in favour of as many posts on early tetrapods as Darren can muster.

    Link to this
  27. 27. Therizinosaurus 4:27 pm 07/5/2013

    “they had CIs > 80 %, I think up to 89.”

    UP to 89?! Hahaha. His 1999 Science analyses vary from 80 to 97, with an average of 89.

    “”In well preserved specimens, the supradentary is fused with the coronoid.”

    Always?”

    Seemingly, as in Monolophosaurus, Ornitholestes, tyrannosaurines and Dromaeosaurus. Brochu (2003- pg. 44) describes the situation quite well.

    “Something is missing here because of misparsed signs.”

    Yup, now fixed by Darren.

    Link to this
  28. 28. ectodysplasin 8:41 pm 07/5/2013

    @David,

    – Laurin 1994 (thesis) through Vallin & Laurin 2004; partly evaluated by Sigurdsen & Green 2011; Vallin & Laurin 2004 was used as the morphological part of the combined matrix of Pyron 2011;
    – McGowan 2002, which is ridiculously tiny, but nonetheless contains characters that the others all ignore; evaluated by checking every cell by me & Laurin 2008b;
    – Anderson 2001, expanded by Anderson 2007 and Anderson et al. 2008b; partly evaluated by me & Laurin 2009 and Sigurdsen & Green 2011; independent derivations that don’t take these into account by Maddin and coauthors 2011/2012 and Huttenlocker et al. 2013, and I have to read Olori’s thesis in more detail;
    – Ruta et al. 2003, independent derivations by Pawley 2006 (thesis), Ruta & Coates 2007 and Danto et al. 2013, derivations with eliminations and additions of taxa by Ruta, Clack, Klembara, Bolt and/or coauthors since 2006; evaluation of Ruta & Coates 2007 ongoing (Germain thesis 2008, me thesis 2010);
    – Carroll 2007 with its overwhelming number of correlated characters and even several OTUs that Carroll himself explicitly considers paraphyletic in the same paper.

    Laurin (1994) is the one I’m thinking of, but it has been developed further by Reisz, etc. The matrix of Anderson (2001) is, at its core, this same matrix with some differences in sampling, character philosophy, and scope, although there have been additional characters added as well. The differences in topology between the Anderson (2001) matrix and the matrix you and Michel have been developing are relatively trivial, with some key differences (specifically, the position of lissamphibians). In some ways, the disagreements between your advisor and mine are kind of ironic, given the wholesale agreement throughout most of their trees. Unfortunately, both of these matrices are pretty susceptible to perturbation when we add braincase; both seem to collapse pretty badly when we add these new characters, and the topologies recovered get really ugly really quickly.

    McGowan (2002) has not been widely used, although a few characters have been adapted into other matrices.

    Pawley (2006) remains unpublished.

    Ruta et al. (2003) is largely a supermatrix, and combines some of the taxon-specific matrices into one large matrix.

    There are also some temnospondyl-specific matrices (e.g. Yates and Warren, 2000, and a recent contribution by Schoch (2013). Still waiting on Julia McHugh’s thesis to be published.

    The Carroll (2007) matrix has not really entered into the general debate in the literature in a significant way, to date.

    I’d even state it the other way around: they neglect everything except the dermal skull to an astonishing degree. Pawley’s thesis (2006) was an exception, but Pawley still hasn’t published.

    Yes. Huttenlocker et al. (2013) add a few braincase characters (e.g. presence of a supraoccipital sinus, presence of a ventral flange of the frontal articulating with the sphenethmoid, presence of an anterior accessory ossicle mediating the contact between the sphenethmoid and the frontal, etc), but only a handful.

    The Anderson (2001) matrix is about 50% postcranial, but there seems to be a lot of convergence and/or missing data there.

    Link to this
  29. 29. John Harshman 8:42 pm 07/5/2013

    “they had CIs > 80 %, I think up to 89.”

    I can do better than that. There’s a theropod neontologist who frequently managed CIs above 90%; 100% once or twice, if I recall. Unfortunately, he’s no longer with us.

    Link to this
  30. 30. ectodysplasin 8:49 pm 07/5/2013

    more @david,

    This is not a bad thing. It’s a sign of honesty*, and indeed a sign that the correlated characters aren’t all correlated with each other.

    The CI is well known to be size-dependent (more characters + more taxa = more opportunities for homoplasy); large theropod matrices always have a CI < 30 %. That’s why the RI and the RC have been developed.

    * There’s, uh, one dinosaur worker whose analyses from the 90s and early 00s had CIs < 80 %. That’s because he cherry-picked the characters. He has stopped doing that… though he has also largely stopped publishing phylogenetic analyses.

    Yes, it’s a sign of honesty, but it can also be a sign that there’s an unsustainable amount of convergence in a dataset. Also worth pointing out that RI is generally rather low in these trees as well. Comparative tests (e.g. S-H tests) generally don’t find that one group’s tree is significantly better than another at explaining any group’s data (if you have an afternoon to waste, try running a few of these for your own edification…it really underscores how little certainty we have in the tree in general, regardless of misleading measures like bootstrap). Generally, I don’t think that, mathematically, the lepospondyl and temnospondyl hypothesis are substantially different (although biologically they clearly are), but I think this is more a factor of how poor our matrices are with respect to endocranial data.

    Right now, though, there’s a lot of noise in our analyses, and not enough signal.

    Link to this
  31. 31. ectodysplasin 9:53 pm 07/5/2013

    one final response before I finish packing for ICVM:

    >>how did fossoriality evolve in early tetrapods

    And how many times…

    Indeed, as well as specialization within the fossorial realm. Not all fossorial early tetrapods approached fossoriality equally, and there are probably multiple digging modes even within fossorial clades (compare Carrolla to Pelodosotis for example). Some amphibamids may turn out to have been fossorial as well (though right now this is only a hunch) and there have been reports of estivation burrows for diplocaulids.

    >>The differences between, say, baphetids, colosteids,
    >>and edopoids might be of interest to a specialist, but
    >>ecologically and functionally these are all
    >>essentially interchangeable animals

    Heh. Their lower jaws perhaps! Colosteids were elongate and lacked, uh, “temporal embayments”, baphetids had those huge holes in the skull for, presumably, some reason, and how close Edops really is to the cochleosaurids isn’t quite obvious. (I need to go to Harvard and see Edops – the skull, the whole skull, and nothing but the skull, alas.)

    Yes, but in many ways these are all specialist concerns. Functionally, we’re still looking at largely aquatic animals with dorsally-situated eyes, and a generalist skull morphology that’s at that less-than-effective halfway point between effective aquatic raptorial predation and effective buccal pumping. There may be somewhat minor differences in anatomy that are potentially of some interest with extensive interpretation (the lack of an otic notch in colosteids as indicative of auditory function, the antorbital fenestra of baphetids as indicative of bizarre pterygoideus morphology, etc) but once again, this is mostly of interest to specialists. I can’t imagine most people would have an interest in the minor differences between a baphetid and a colosteid.

    Link to this
  32. 32. David Marjanović 7:45 am 07/6/2013

    UP to 89?! Hahaha. His 1999 Science analyses vary from 80 to 97, with an average of 89.

    Whoa. My memory has been merciful on me!

    Laurin (1994) is the one I’m thinking of, but it has been developed further by Reisz, etc.

    Yes; fortunately, however, this is a linear sequence, with each version using all of the corrections and additions from all previous versions.

    The matrix of Anderson (2001) is, at its core, this same matrix with some differences in sampling, character philosophy, and scope, although there have been additional characters added as well.

    Hm. Taxon sampling has always been very different – it started out as an attempt to sample all lepospondyls (at species level!) and Eocaecilia and pretty much nothing else, then a bunch of dissorophoids and batrachians were added, but between and outside those there’s still just Acanthostega, Greererpeton, Proterogyrinus, Seymouria, Limnoscelis, Eryops and what else? No wonder Greererpeton comes out as a temnospondyl.

    Ruta et al. (2003) is largely a supermatrix, and combines some of the taxon-specific matrices into one large matrix.

    Yep. That’s why it and its descendants contain so many duplicated characters and so many characters where state 1 is a very specific, carefully described condition that often combines several characters while state 0 is “everything else”, no matter how many distinguishable states lurk therein.

    > > And how many times…

    Indeed, as well as specialization within the fossorial realm. Not all fossorial early tetrapods approached fossoriality equally, and there are probably multiple digging modes even within fossorial clades (compare Carrolla to Pelodosotis for example).

    That’s part of my point: I don’t find them close together. :-)

    Some amphibamids may turn out to have been fossorial as well (though right now this is only a hunch) and there have been reports of estivation burrows for diplocaulids.

    Huh.

    On baphetids vs. colosteids: colosteids were rather eel-like; baphetid postcrania are very poorly known, but what’s known suggests a shorter trunk and longer limbs.

    On buccal pumping, two things: 1) the above had a postbranchial lamina on the cleithrum, so probably relied on internal gills; 2) caecilians are buccal pumpers – they simply inhale on average 16 times for every time they exhale. It is not true that buccal pumping only works if you have a head the size of your lungs.

    Link to this
  33. 33. ectodysplasin 8:52 am 07/6/2013

    Sitting down with my morning coffee before flying to Barcelona…

    Hm. Taxon sampling has always been very different – it started out as an attempt to sample all lepospondyls (at species level!) and Eocaecilia and pretty much nothing else, then a bunch of dissorophoids and batrachians were added, but between and outside those there’s still just Acanthostega, Greererpeton, Proterogyrinus, Seymouria, Limnoscelis, Eryops and what else? No wonder Greererpeton comes out as a temnospondyl.

    Yes, it’s a relatively limited sample. Periodically, I think that the solution is to scrap the matrix itself, scavenge the characters, and rebuild it from scratch. Ultimately, it was designed to do something very specific (investigate lepospondyl relationships) but at this point, it’s being used for very different purposes. I’ve more or less broken the matrix at this point, so it’s not really even particularly useful for its original purpose.

    >>Indeed, as well as specialization within the
    >>fossorial realm. Not all fossorial early tetrapods
    >>approached fossoriality equally, and there are
    >>probably multiple digging modes even within fossorial
    >>clades (compare Carrolla to Pelodosotis for example).

    That’s part of my point: I don’t find them close together.

    This is where the braincase data begins to become rather important. We have a new animal that ties brachystelechids in with other recumbirostrans really quite nicely, and most of the important morphology is in the braincase.

    Link to this
  34. 34. ectodysplasin 9:03 am 07/6/2013

    Anyways, that seems to be sufficiently off-topic from the original mesosaur conversation!

    Link to this
  35. 35. David Marjanović 8:47 am 07/7/2013

    We have a new animal that ties brachystelechids in with other recumbirostrans really quite nicely

    Interesting. Meanwhile, speaking of brachystelechids, the long-awaited redescription of Batropetes by Sabine Glienke came out yesterday; I’m looking forward to reading it tomorrow!

    Link to this
  36. 36. Hai~Ren 10:34 am 07/7/2013

    I find it interesting that besides the mesosaurs, Claudiosaurus, and “younginiforms”, it seems like there aren’t many other marine tetrapods from the Paleozoic (maybe the archegosauroid temnospondyls?). Yet in the Triassic, we see plenty of marine reptile groups appear – hupehsuchids, thalattosaurs, ichthyosaurs, protorosaurs, and sauropterygians (not sure if I’ve missed any). Makes me wonder whether there was something about the post-Permian recovery that facilitated adaptation to marine environments.

    Link to this
  37. 37. ectodysplasin 3:29 pm 07/7/2013

    Holocephalans were really really ecologically diverse and dominated estuarine ecosystems up into the early Triassic. That may have contributed.

    Link to this
  38. 38. David Marjanović 11:29 am 07/8/2013

    Reading the redescription of Batropetes now. :-) It promises two more species in a forthcoming paper.

    it seems like there aren’t many other marine tetrapods from the Paleozoic

    Well, plenty have been found in brackish to marine sediments. Iberospondylus comes from a beach, for instance. Many classic sites seem to represent some kind of mangrove forest.

    Tulerpeton must have been fully marine, because it was found 200 km away from the nearest land.

    Yet in the Triassic, we see plenty of marine reptile groups appear –

    As well as a marine temnospondyl clade, Trematosauridae.

    Link to this
  39. 39. ectodysplasin 10:30 pm 07/8/2013

    There’s also a bizarre “trimerorhachid” from a coastal site outside of Robinson, Kansas, that preserves bones of this animal and the lungfish Sagenodus in stromatolites. We could also perhaps make note of the aistopod Lethiscus, which comes from the marine Wardie Shale, and Phlegethontia from Mazon are primarily found in the marine Essex fauna (although Phlegethontia is also known from freshwater localities, as are most other aistopods).

    Link to this
  40. 40. actionshoz 1:10 pm 07/26/2013

    DeBraga and Rieppel (1997) have (likely unintentional) errors in their matrix construction that cause weighting issues to favor their preferred topology (not even starting to look at problems like taxon selection, miscoded OTUs, etc). Analyses which more or less copied the same characters out are guilty of the same. To know how far off the interrelationships of eureptiles or the sauropsid branch as a whole will be a lot of work. I think the broad relationships of major terrestrial groups are probably correct (the marine radiations are another story).

    Maisch (2010) found mesosaurs to be eureptiles, in a sister relationship with ichthyosaurs. Surprised you didn’t mention this. ;-) Though Maisch did not explore the factors leading to this, it is more complex than one might think (w4tp).

    Link to this
  41. 41. actionshoz 9:14 pm 07/26/2013

    re: miscellaneous characters mentioned above, especially those which could be associated with feeding, these could be subject to substantial convergence. Look at lizards, there are multiple convergent losses of limbs, digits (and supposed reevolution of digits), etc. Palatal dentition disappears and reevolves, etc. I think we need to be careful when looking at morphological data. It may not be as relevant as we think, especially when our data sampling is not as thorough as we’d like to believe.

    Link to this
  42. 42. naishd 6:03 am 07/27/2013

    actionshoz: yeah, as clear from the whole discussion upthread, there’s a lot of work to do and a lot that’s likely to change, plus there are major problems with the existing, published data sets and phylogenies.

    Re: Maisch (2010), which paper are you referring to? Maisch has, of course, argued on and off that ichthyosaurs are well outside of Diapsida, but I wasn’t aware that he’d supported a link with mesosaurs.

    Darren

    Link to this
  43. 43. actionshoz 11:38 am 07/27/2013

    Maisch, M.W. 2010. Phylogeny, systematics, and origin of the Ichthyosauria — the state of the art. Palaeodiversity 3: 151-214.

    I certainly wouldn’t regard the analyses as substantial, but I did think the results were interesting. There are some ways I’ve re-examined the data (characters, taxon sampling, scoring, etc) that others might find more so. Remember, Maisch only revised/added two OTUs to each data set examined (Rieppel and Reisz, 1999; Laurin and Reisz, 1995).

    Link to this
  44. 44. naishd 3:40 pm 06/16/2014

    I forgot to come back and check the comments here. Just want to say thanks to actionshoz… if that is your real name.

    Link to this

Add a Comment
You must sign in or register as a ScientificAmerican.com member to submit a comment.

More from Scientific American

Scientific American Special Universe

Get the latest Special Collector's edition

Secrets of the Universe: Past, Present, Future

Order Now >

X

Email this Article

X