Recalibrating clade origins, part 2

Marjanovic 2019 reports on
the origin of several clades based on fossils and molecules. Yesterday we looked at part 1, which focused on the abstract. Today: the origin of several more listed clades.

Gnathostomata (Chondrichthyes + Osteichichthyes)
Marjanovic cautiously proposes the mid-Florian (Early Ordovician, 475 mya) for the origin using traditional taxa and cladograms.

By contrast, the LRT splits off quasi-jawless sturgeons before the appearance of jawed sharks + other bony fish. It also splits off the jawed Loganellia + Rhincodon + Manta clade before the Polyodon + ratfish + sharks + skates clade and the Pachycormus + Hybodus clade before the dichotomy that resulted in the rest of the bony fish (the now polyphyletic ‘Osteichthyes‘)… so direct comparisons are not apples and apples here. Sturgeons first appear much later in the fossil record. Loganellia appears in the Early Silurian with an earlier genesis. So Marjanovic’s estimate may be a little early.

Osteichthyes (Actnopterygii + Sarcopterygii)
Marjanovic reports, “The oldest known uncontroversial crown-group osteichthyan is the oldest known dipnomorph, Youngolepis.” He suggests, “the minimum age for this calibration is the same as that for the next node,” the Silurian/Devonian boundary, 420 mya.

The LRT includes placoderms within one branch of the bony fish, so Entelognathus along with the stem-lungfish Guiyu, both in the Late Silurian are older than Marjanovic suggests with an earlier genesis. Sturgeons, which traditional workers consider a member of the Osteichyes, phylogenetically preceded Longanellia, which is known from Early Silurian strata. So, again we’re not comparing similar cladograms here. The LRT tests a wider gamut of taxa, which is an advantage in that it opens further possibilities than tradition dictates.

Dipnomorpha + Tetrapodomorpha (lungfish + lobe fin ancestors of tetrapods)
Marjanovic reports, “I suggest a hard minimum age of 420mya.” (See above).

The LRT includes Late Siluirian Guiyu within the stem-lungfish clade. so the split occurred earlier.

Tetrapoda (Amphibia + total group of Amniota)
Marjanovic reports, “the richer and better studied Famennian (end-Devonian) record, which has not so far yielded tetrapods close to the crown-group but has yielded more stemward tetrapods and other tetrapodomorphs (Marjanović and Laurin, 2019), should be used to place a soft maximum age around very roughly 365 Ma.”

Figure 3. Tersomius texensis, an amphibamid lepospondyl close to Dendrerpeton.

Figure 1. Tersomius texensis, an amphibamid lepospondyl close to Dendrerpeton.

In the LRT the last common ancestor of Amphibia + Amniota is Tersomius (Fig. 1), a late survivor in the Early Permian of an earlier genesis and radiation. The oldest taxa from this clade in the LRT are the basal amniotes / amphibian-like reptiles, Silvanerpeton and Eldeceeon from the Viséan (335 mya), with a long list of late surviving taxa between them and Tersomius, some eight nodes beyond the Late Devonian Acanthostega and Ichthyostega (365 mya). So the Tournaisian (355 mya) split suggested by Marjanovic seems about right.

Amniota (Theropsida + Sauropsida)
Marjanovic reports, “I refrain from recommending a maximum age other than that of the preceding Node, even though such an early age would imply very slow rates of morphological evolution in the earliest thero- and sauropsids.”

The LRT recovers a different basal dichotomy (Archosauromorpha + Lepidosauromorpha) and a different last common ancestor for all amniotes (Silvanerpeton) than Marjanovic is working with. Silvanerpeton is Viséan in age (~335 mya). In the LRT ‘Amniota’ is a junior synonym for Reptilia.

Crown group of Diapsida (Lepidosauromorpha + Archosauromorpha)
Marjanovic reports, “I cannot express confidence in a maximum age other than that of  Node 106, which I cannot distinguish from the maximum age of Node 105 as explained above. This leaves Node 107 without independent calibrations in the current taxon sample.”

The LRT finds two origins for reptiles with a diapsid skull architecture. So the tradtional clade ‘Diapsida’ is also a junior synonym for Reptilia and Marjanovic is using an outdated and under represented cladogram. Lepidosauromorph diapsids first appear with Paliguana in the earliest Triassic. Archosauromorph diapsids first appear with Erpetonyx and Petrolacosaurus in the Late Carboniferous with an earlier genesis. These taxa are not mentioned by Marjanovic.

Archosauria (Crocodile total group + Bird total group)
Marjanovic reports, “I accept the Permian-Triassic boundary (251.902 ± 0.024 Ma: ICS; rounded to 252) as the soft maximum age on the grounds that a major radiation of archosauromorphs at the beginning of the Triassic seems likely for ecological reasons.”

The LRT restricts membership within the Archosauria to just Crocodylomorpha + Dinosauria. So the maximum age for this dichotomy is younger and the last common ancestor is the PVL 4597 specimen (late Middle Triassic, 230mya) traditionally assigned to Gracilisuchus, but nesting apart from the holotype.

The LRT finds the Archosauriformes first appeared in the Late Permian (260mya), arising from a sister to Youngoides romeri (FMNH UC1528) thereafter splitting into clades arising from the larger Proterosuchus and the smaller Euparkeria.

Alligatoridae (Alligatorinae + Caimaninae)
Marjanovic reports, “Given this uncertainty, I have used a hard minimum age of 65 Ma for present purposes, but generally recommend against using this cladogenesis as a calibration for time trees.”

The LRT does not include pertinent taxa surrounding this split.

Figure 1. Megapodius is the extant bird nesting at the base of all neognathae (all living birds except ratites).

Figure 2. Megapodius is the extant bird nesting at the base of all neognathae (all living birds except ratites). And it looks like a basal bird. It also looks a bit like the Solnhofen bird, Jurapteryx. It is easy to imagine diverse forms arising from this bauplan and the LRT indicates that is exactly what happened.

Crown group of Neognathae (Gallanseres + Neoaves)
Marjanovic further defines this clade as, “The last common ancestor of Anas, Gallus and Meleagris on one side and Taeniopygia.” More commonly Marjanovic nests a duck, a chicken and a turkey on one side and a zebra finch on the other as the basal dichotomy of all living birds, sans ostriches, kiwis and kin. Marjanovic reports, “As the soft maximum age I tentatively suggest 115 Ma, an estimate of the mid-Aptian age of the (likewise terrestrial) Xiagou Fm of northwestern China, which has yielded a diversity of stem-birds but no particularly close relatives of the crown.”

Taxa listed by Marjanovic are all highly derived taxa in the LRT where the scrub fowl, Megapodius (Fig. 2) and the tinamou, Crypturus, are basal neognaths. These would have had their genesis in the Earllest Cretaceous given that Early Cretaceous clades that redevelop or retain teeth are more derived.

More tomorrow…


References
Marjanovic D 2019. Recalibrating the transcriptomic timetree of jawed vertebrates.
bioRxiv 2019.12.19.882829 (preprint)
doi: https://doi.org/10.1101/2019.12.19.882829
https://www.biorxiv.org/content/10.1101/2019.12.19.882829v1

6 thoughts on “Recalibrating clade origins, part 2

  1. In the LRT the last common ancestor of Amphibia + Amniota is Tersomius (Fig. 1), a late survivor in the Early Permian of an earlier genesis and radiation.

    No.

    A phylogenetic analysis cannot find anything as an ancestor of anything else. They’re simply incapable of doing that. Even if you put an actual known ancestor into your data matrix, no phylogenetic analysis can tell you that that’s what it is.

    If you find a zero-length branch in a trichotomy with its two possible descendants, and if it’s older than all known fossils of those two branches, then that zero-length branch enters into consideration for being the last common ancestor of the other two. But that’s as far as phylogenetic analysis can go.

    Part of the reason is that no data matrix is exhaustive. For example, if your fossil is a complete articulated skeleton, you still cannot know if its absence of skeletal autapomorphies was matched by an absence of soft-tissue or molecular autapomorphies.

    • Disagree. When you have five taxa in a cladogram, the resolution is low. When you have over a thousand taxa in a cladogram, the resolution is higher. The LRT delivers a gradual accumulation of derived traits and a last common ancestor for every clade. That is why Silvanerpeton (at present) must be an amniote because all derived taxa are amniotes.

      • Uh, no. The resolution of a cladogram has no relationship to its size. A matrix of five taxa can lead to a fully resolved tree, and a matrix of over a thousand can lead to tens or hundreds of thousands of equally optimal trees.

        But I don’t understand how that relates to any part of my comment.

        That is why Silvanerpeton (at present) must be an amniote because all derived taxa are amniotes.

        …That is why all colorless green ideas must sleep furiously? I honestly don’t understand what you mean. First of all, what do you mean by “derived”?

      • Derived: ‘phylogenetic descendants of.’
        I don’t trust a matrix of over a thousand that leads to tens or hundreds of thousands of equally optimal trees… unless several of the taxa are skull only and others are skull-less taxa, for instance. In such a case it is better to reduce the taxon list to just well represented taxa. Then tentatively add others, as I have done with the red taxa in the LRT, where they most closely nest.

      • Derived: ‘phylogenetic descendants of.’

        Of what, then? You just said “all derived taxa are amniotes”.

        (This is also not how people apply “derived” to character states; they mean “apomorphic.)

        In such a case it is better to reduce the taxon list to just well represented taxa.

        Absolutely not, no. That way you lose valuable information. I cite a few studies on this in my preprint that you’ve blogged on.

        Never be afraid of missing data!

      • No fear here. Based on the large number of skull only, mandible only and skull-less taxa in the LRT, I’m not afraid of missing data. David, I suggest if there are issues, that you retreat to the better known taxa, establish your tree topology, then venture out by adding less well known taxa. Just be smart about it.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.