Secretary birds (genus: Sagittarius) are small, extant terror birds

Phorusrhachids, the giant terror birds of South America,
are the closest relatives of the secretary bird of Africa, Sagittarius, in the large reptile tree (LRT, 1869+ taxa). The similar, but more distantly related seriema (genus: Cariama) of South America is closer to flamingos (Phoenicopterus) than to terror birds, despite the geographic proximity and tradition.

Secretary birds must have appeared on the planet
prior to the rifting of South America from Africa in the Early Cretaceous, 100 mya (Fig. 1). Afterwards there was a growing ocean between secretary birds and terror birds, as in toucans and hornbills.

Figure 4. South America and Africa during the Albian, 100 mya. This is when toucans and hornbills must have separated.
Figure 1. South America and Africa during the Albian, 100 mya. This is when toucans and hornbills must have separated.

This secretary bird video and its narration is wonderful.
It sheds light on how early birds and bird-like theropds lived and preyed on a menu of Late Cretaceous and Tertiary snakes and mammals. Fascinating stuff, especially when the filmmakers propose a new use for feathers.

Figure 3. Skull of Phorusrhacos, a giant terror bird in three views.
Figure 2. Skull of Phorusrhacos, a giant terror bird in three views.
Figure 1. Phorusrhacos to scale with Dinornis, Struthio and Homo.
Figure 3. Phorusrhacos to scale with Dinornis, Struthio and Homo.
Figure 5. Cariama compared to Sagittarius. The former is closer to flamingos. The latter is closer to terror birds.
Figure 4 Cariama compared to Sagittarius. The former is closer to flamingos. The latter is closer to terror birds.
Figure 2. Sagittarius (secretary bird) and Cariama (seriema). While clearly related, these two nest at the base of two different major bird clades.
Figure 5. Sagittarius (secretary bird) and Cariama (seriema). While clearly related, these two nest at the base of two different major bird clades.

Plesiosaur phylogeny: revisiting O’Keefe 2001

O’Keefe 2001 analyzed members of the Plesiosauria
(Fig. 1) nesting long-necked Rhomaleosaurus (Fig. 3) with short-necked Kronosaurus (Fig. 3) and long-necked Thalassiodracon was the last common ancestor). O’Keefe also nested short-necked Dolichorhynchops with long-necked Styxosaurus and long-necked Plesiosaurus was a last common ancestor. Why the mix up? Let’s hear from the author himself.

O’Keefe reported,
“Characters from the entire skeleton support these relationships, although characters of the skull roof and palate are especially useful.”

Figure 1. Cladogram from O’Keefe 2001, colorized here with purples and blues for pliosaurs, reds and oranges for plesiosaurs.

By contrast
the large reptile tree (LRT, 1869+ taxa, subset Fig. 2) nests Rhomaleosaurus basal to the dichotomy that splits long-necked taxa from short-necked taxa. A clade of long-necked taxa with only two long hind flippers (Thalassiodracon + Yunguisaurus, Figs 3, 6) precedes Rhomaleosaurus (Fig. 3). Surprisingly a clade of four-long-flippered taxa (Hauffiosaurus and kin, Fig. 3) nest independent of other four-flippered taxa. The LRT is able to recover examples of convergence like this.

Figure 2. Subset of the LRT focusing on Sauropterygia. Colors applied based on figure 1. These results more or less randomize the results of O’Keefe 2001, but puts long-neck taxa with long-neck taxa.
Figure 3. Pliosaur origins according to the LRT, beginning with Pistosaurus.

Let’s compare and contrast the two results.
O’Keefe had firsthand examinations of his plesiosaur taxa. I worked only from the literature. On a side note, I remember seeing O’Keefe examining plesiosaurs in Germany while I was examining pterosaurs. He’s an interesting guy.

We both agree
that long-necked, no flipper Pistosaurus (Fig. 3) was basal to most later clades of four-flipper plesiosaurs however they divide thereafter.

O’Keefe splits his Plesiosauroidea and Pliosauroidea
at this points and nests long-necked Thalassiodracon at the base of his short-neck Pliosauroidea (= Kronosaurus + Rhomaleosaurus clade).

The LRT also nests Thalassiodracon basal to Rhomaleosaurus,
but both prior to the long-neck vs. short-neck dichotomy (Fig. 3). Hauffiosaurus nests 11 steps prior to the other four-long-flipper taxa, one node prior to Pistosaurus. So this appears to be an independent and convergent acquisition not recognized nor recovered by O’Keefe 2021.

O’Keefe 2001 did not have
the benefit of the LRT when he reported, “The Sauropterygia is a clade of basal diapsids, more closely related to lepidosaurs than archosaurs but near this basal dichotomy.”

In the LRT, the closer relationship is with marine younginiforms and archosauriforms, not lepidosaurs.

O’Keefe also reported,
“Some recent work has indicated that Testudines is the sister group of Sauropterygia, although this work is controversial.”

This phrasing is typical of workers who have no idea how clades are related to one another and either take the word of a fellow worker or not, depending on the general mood. I would have said the same thing at the time. The LRT (2010-2021) resolves all such issues, but it really is the job of paid professionals to do this. The LRT currently fills this vacuum, hopefully only temporarily.

Back in 2001, O’Keefe was also filling a vacuum
when he reported, “No comprehensive review of the plesiosaur skull has been attempted since the work of Andrews (1910, 1913).”

Setting an outgroup
O’Keefe reports, “The condition of the Permian plesiomorphic diapsid Araeoscelis, described in detail by Vaughn (1955), is an acceptable model from which to derive the sauropterygian skull roof.”

The LRT likewise nests Araeoscelis at the base of the Sauropterygia, but many overlooked transitional taxa (Fig. 4) fill the gap between them, several with a diapsid morphology.

Figure 3. Spinoaequalis and descendant marine younginiformes.
Figure 4. Spinoaequalis and descendant marine younginiformes. These give rise to pachypleurosaurs, plesiosaurs, placodonts, mesosaurs, ichthyosaurs and thalattosuchians. Araeoscelis and other more primitive diapsids are not shown. These are some of the transitional taxa overlooked by O’Keefe twenty years ago, which is forgivable. I made similar mistakes back then.

O’Keefe reports,
“The nasal is lost in all Plesiosauroidea, including Plesiosaurus.” There are lots of Plesiosaurus skulls. This one (Fig. 5), P. dolichodeirus appears to keep the frontal (blue) and nasal (pink) separate.

Figure x. Plesiosaurus skull (BMNH 39490) showing nasals.
Figure 5 Plesiosaurus skull (BMNH 39490) apparently showing nasals.

O’Keefe 2001 introduced Hauffiosaurus
(Fig. 3) to the world of paleontology, He wrote, “The skeleton is approximately 2.5 m long, and displays an interesting mix of plesiomorphic, derived, and apomorphic features.”

According to the LRT, which minimizes taxon exclusion, Hauffiosaurus developed four long flippers independently, and phylogenetically before and convergent with the last common ancestor of all pliosaurs and plesiosaurs… given the present taxon list (Fig 2).

Building the O’Keefe cladogram
O’Keefe did what lots of paleontologists still do. He chose his outgroup taxa. He wrote, “Three taxa were chosen as outgroups for this analysis.”

The LRT chooses valid outgroup taxa for you because it minimizes taxon exclusion.

O’Keefe wrote,
Thirty-one plesiosaur genera were coded for inclusion in the phylogenetic data matrix.”

By contrast,
only 19 sauropterygian taxa are more derived than Pistosaurus (Fig. 3) in the LRT (subset Fig. 2). Twenty are more primitive. Among them is Hauffiosaurus (Fig. 3). Nesting taxa in the outgroup that O’Keefe considered ingroup taxa affects the way the rest of the taxa relate to one another and we’re off to a rocky start.

O’Keefe wrote,
“All known clades are well-represented, however, and the omission of some ingroup taxa from some clades should not influence the results reported here.”

That’s a bold statement falsified by the LRT some twenty years later. Two Hauffiosaurus LRT sisters, Acostasaurus and Anningsaura, known from skulls only, were described after O’Keefe published in 2001, so he can’t be held responsible for those.

The character list
O’Keefe wrote, “The 34 taxa listed above were scored for 166 morphological characters. Of these characters, 107 concerned the skull and 59 were postcranial.”

238 multi-state characters nest taxa in the LRT.

O’Keefe wrote,
“A second analysis was performed with the ‘morphometric’ characters removed, based on the finding that the pliosauromorph body type may have evolved convergently (O’Keefe 2002).”

Note this citation is a year later than O’Keefe 2001. How did he cite a future publication? It was in ‘in press’. Sometimes workers produce more than one paper from a single study.

Figure 3. Albertonectes reconstructed. This 11 m elasmosaur is the longest thusfar recorded. This may be the breathing pose, swallowing air, then submerging the neck. When horizontal the air could be passed back to the lungs, as hypothesized for Dinocephalosaurus.
Figure 6. Albertonectes reconstructed. This 11 m elasmosaur is the longest thus far recorded. This may be the breathing pose, swallowing air, then submerging the neck. When horizontal the air could be passed back to the lungs, as hypothesized for Dinocephalosaurus. This could be a feeding pose, exhaling a bubble net from a long trachea to surround a school of fish overhead.

O’Keefe’s results
“The topology of clade Plesiosauria replicates many of the findings advanced by earlier workers. The basic dichotomy between the Plesiosauroidea and the Pliosauroidea is a well-supported finding.”

Not really. If memory serves, plesiosaurs traditionally had small heads and long necks, while pliosaurs traditionally had larger heads and shorter necks. O’Keefe nests long-necked Thalassiodracon (Fig. 3) at the base of the pliosaurs. The LRT nests it with another long-hind-flipper only taxon, Yunguisaurus, not described until 2006.

Figure 7. Muraenosaurus enters the LRT from these data.
Figure 7. Muraenosaurus data used in the LRT.

Muraenosaurus enters the LRT today alongside Rhomaelosaurus.
O’Keefe reported, “The elasmosaur-like long neck and small head evolved independently in Muraenosaurus.”

This is supported by the LRT. Let your cladogram do what you want it to do… nest taxa without bias or preconception.

O’Keefe FR 2001. A cladistic analysis and taxonomic revision of the Plesiosauria
(Reptilia: Sauropterygia). Acta Zool. Fennica 213: 1–63.
O’Keefe FR 2002. The evolution and functional morphology of plesiosaur and pliosaur morphotypes inthe Plesiosauria (Reptilia: Sauropterygia). — Paleobiology 28(1). [In press at the time in 2001].

Late Jurassic basal sea turtle, Thalassemys, enters the LRT

Joyce, Mäuser and Evers 2021
document two new specimens of a basal sea turtle from the Late Jurassic Solnhofen Formation, Thalasssemys bruntrutana (Fig. 1. NKMB Watt18/211; Rütimeyer 1873). One fossil preserves soft tissue around the hind pes. The shell is disc-like with lateral ribs and the skull is small with large orbits. Manual digit 5 is the longest and metacarpal 4 is the longest.

Other Solnhofen turtles include
Eurysternum, Solnhofia and Palaeomedusa.

Figure 1. Thalassemys bruntrutana in situ with diagram from Joyce, Mäuser and Evers 2021 who nested Thalassemys at the base of their monophyletic clade of sea turtles (Fig. 2).

The Joyce, Mäuser and Evans cladogram
nests Thalassemys at the base of several Late Jurassic fossil sea turtles (Fig. 2, red arrow).

Figure 2. Turtle cladogram from Joyce, Mäuser and Evans excludes taxa that would change this tree topology.
Figure 2. Turtle cladogram from Joyce, Mäuser and Evans excludes taxa that would change this tree topology.

in the large reptile tree (LRT, 1867+ taxa; Fig. 3) Thalassemys nests at the base of one sea turtle clade among several.

Figure 3. Subset of the LRT focusing on turtle evolution. The outgroup includes pareaiasaurs and Stephanospondylus..

Taxon exclusion in the Joyce, Mäuser and Evans cladogram
prevents these turtle experts from understanding 1) the dual origin of turtles from pareiasaurs, 2) softshell sea turtles, like Ocepecephalon, are not related to hardshell sea turtles; 3) the extant pig-nosed turtle, Carettochelys is related to the extant leatherback, Dermochelys; and 4) the giant Cretaceous sea turtle, Archelon, is derived from Macrochelys, the snapping turtle (video below) and thus is not related to the clade of other ‘hardshell’ sea turtles. Adding taxa solves all these problems.

Noticeable by its absence,
the basal softshell turtle, Odontochelys, is not mentioned in the text nor listed among the turtle taxa in Joyce, Mäuser and Evans 2021 (Fig. 2). Also absent from the Joyce, Mäuser and Evans cladogram are any turtle outgroup taxa. These guys are turtle experts and they don’t know where turtles came from? C’mon! Add taxa to find out (see Fig. 3 for the current solution).

The generic evolution of sea turtles
is more or less documented in the following three videos. Note the transition from typical asymmetrical left-right-left-right tetrapod locomotion to symmetrical flapping in river turtles and sea turtles. Turtles are essentially weightless in water, so they didn’t have to develop extended coracoids and mighty muscles like birds and pterosaurs, or extended clavicles like bats to power their flaps against the pull of gravity. Instead, sea turtles seem to have simply and slowly learned to swim by increasingly symmetrical flapping over generations as their forelimbs gradually became longer, flatter, more flexible and evolved into flippers.

Evidently turtles were in no hurry to learn how to flap while swimming.
Several clades did it (Fig. 3), so this must have come  about quite naturally, if slowly.

Joyce WG 2003. A new Late Jurassic turtle specimen and the taxonomy of Palaeomedusa testa and Eurysternum wagleri. PaleoBios. 23 (3): 1–8.
Joyce WG, Mäuser M and Evers SW 2021. Two turtles with soft tissue preservation from the platy limestones of Germany provide evidence for marine flipper adaptations in Late Jurassic thalassochelydians. PLoS ONE 16(6): e0252355.
Rütimeyer L 1873. Die fossilen Schildkröten von Solothurn und der übrigen Juraformation. Neue Denkschriften der allgemeinen Schweizerischen Gesellschaft für die gesamten Naturwissenschaften 25:1–185.

Shark skull evolution updated

Short one today,
updating an older graphic with the latest ReptileEvolution images in order (Fig. 1).

Figure 1. Shark skull evoluition updated today, June 2, 2021.

More taxa =
more parsimony, smoother transitions, greater detail, more support.

Loganellia is Early Silurian, so more primitive taxa are earlier, in the Ordovician.

start with Chondrosteus.

Multiple opercula (gill openings)
begin with Rhincodon.

Marginal teeth
start with Ginglymostoma.

Nasals extend beyond the jaws
(convergent with or reversal to sturgeons) with Ginglymostoma and Sphyrna, the hammerhead.

Nasals extend waaay beyond the jaws
in guitarfish (Rhinobatos) + sawfish (Pristis) on one branch and Squaloraja + the goblin shark (Mitsukurina) on another by convergence.

Nasals are much smaller
in Isistius and Ozarcus leading to Hybodus and bony fish, which also return to a single gill opening (as in Prohalecites Fig. 2, Amia, Fig. 1).

Ancestors of Gregorius also evolve to
the moray eel (Gymnothorax, not shown), Harpagofutator (Figs. 1, 2), Edestes, and Helicoprion (Fig. 1).

Ancestors of Gregorius also evolve to
spiny sharks, like Homalacanthus (Early Devonian, Fig. 2) so all more primitive taxa appeared earlier, in the Silurian.

Ratfish and Iniopterygians
arise from horn sharks like Heterodontus. These taxa also reduce the number of gill openings.

Torpedo rays
arise from hammerheads. Other skates, rays and manta rays have separate origins.

Paddlefish (Polyodon)
with a single gill opening arise from basking sharks (Cetorhinus) with multiple gill openings.

Figure 2. Representatives from the Early Devonian radiation that gave us bony fish, including Prohalecites and Homalcanthus.

This appears to be a novel hypothesis
of interrelationships (Fig. 3). If someone proposed this earlier, let me know so I can promote that citation.

Figure 3. Subset of the LRT focusing on elasmobranchs.

Don’t put all your ‘faith’ in one to a dozen traits
(like gill opeinings) in order to exclude taxa. That’s called “Pulling a Larry Martin.” Expand your taxon list and see where evolution takes your cladogram.

According to the LRT,
humans, bats, frogs, dinosaurs and pterosaurs all have shark ancestors, so we are all sharks. Something else to get used to.

Pteraichnus wuerhoensis, a new pterosaur ichnospecies

Li, Wang and Jiang 2021 report
on a new slab filled with small Early Cretaceous pterosaur tracks (Figs. 1–3).

Figure 1. Diagram from Li, Wang and Jiang 2021 shown about half life size. Colors added here to help visually divide left manus (red) from right manus (purple) and pedal tracks (black). Several pedal trackways are colorized and with arrows. Only a few manus impressions can be associated with each of these pedal trackways. Blue arrows point to the two specific imprints reproduced as photos in figure 2.

The authors report,
IVPP V 26281.2, which is a greyish green sandstone block (125 cm × 25 cm) with 114 natural casts (convex hyporelief) of small pterosaur tracks was collected from Huangyangquan Reservoir tracksite 1, Wuerho region, northwestern Junggar Basin, Xinjiang, China.”

Figure 2. From Li, Wang and Jiang 2021, two of the best preserved tracks presented at about 1.5x natural size if viewed on an 72dpi monitor.

Coeval (Early Cretaceous) pterosaur taxa known from bones
include two slender dsungaripterids, Noripterus (Fig. 3) and Dsungaripterus. These two were derived from sharp-snouted Late Jurassic Solnhofen germanodactylids in the large pterosaur tree (LPT, 258 taxa). The authors suggested the tracks were made by Noripterus (Fig. 3), which is chronologically and geographically reasonable, until one notices a mismatch in proportions and size.

Figure 3 Photographs and matching diagrams (central column) from Li, Wang and Jiang 2021. Colors and hypothetical bones in right column added to the photos. Comparative Late Jurassic pedes (right column) were added for comparison after phylogenetic analysis matched the track data to Solnhofen Rhamphorhynchus taxa in the LPT.

Li, Wang and Jiang report,
“There is no preservation of impressions formed by the fifth metatarsal and digit V.” …and that’s how these two ichnites were scored individually in the LPT.

Even so, after phylogenetic analysis
both ichnites surprisingly nested within the Late Jurassic genus Rhamphorhynchus (Figs. 4–6), though millions of years after Rhamphorhynchus fossils were buried in Solnhofen strata.

Figure 2. Rhamphorhynchus specimens to scale. The Lauer Collection specimen would precede the Limhoff specimen on the second row.
Figure 4. Rhamphorhynchus specimens to scale. The new WordPress tools no longer permit a ‘click’ to send to a new url. I miss that tool. See figures 5 and 6 for enlargements of taxa chosen by the LPT as the best matches among all 258 tested pterosaurs for the two Wuerho trackmakers.

It is worthwhile to take a closer look at the Wuerho pedal impressions
(Fig. 2). The medial metatarsals appear to be slightly elevated because their impressions disappear as if the sole of the foot was arched (see Peters 2011 for more details and examples). Such a pedal arch would put more weight (= creating deeper impressions) on the medial and lateral digits. The impression of the lateral digit appears to have an unexpected joint not found in pterodactylid or ctenochasmatid ichnites. That ‘bend’ was also drawn by Li, Wang and Jiang (Figs. 2, 3) producing an unexpected lateral bump. Given the present data, the ichnites appear to document metatarsal 4 resting upon pedal digit 5 in this trackmaker (Fig. 3), thus obscuring the appearance of a straight metatarsal 4 that would have been sitting on top of that bent digit 5.

If this is incorrect,
and the Wuerho trackmakers had no pedal digit 5, that would also be in accord with, and convergent with, other post-Jurassic pterosaur descendants of Middle to Late Jurassic dorynathids and scaphognathids in the LPT. These also greatly reduce pedal digit 5 and so produce four convergent clades of ‘pterodactyloid’-grade pterosaurs in the Latest Jurassic and Cretaceous. As mentioned above, scoring in the LPT assumed no long pedal digit 5.

Figure 1. Rhamphorhynchus intermedius (n28) reconstructed.
Figure 5. Rhamphorhynchus intermedius (n28) reconstructed. This basalmost member of the genus Rhamphorhynchus has a pes similar to the small trackmaker of the Wuerho tracksite.

I traced several pedal trackways in the slab
(Fig. 1), but only a few left and right manus impressions are associated with each pedal trackway. That indicates bipedal locomotion with an occasional manus touch to the substrate for these trackways. The authors counted 57 manus imprints and 57 pes imprints, but only a few of these are associated with pedal tracks in typical pterosaur fashion. A collection of manus only tracks appears at mid-slab. Thus the matching numbers (57 and 57) are coincidental.

Figure 2. Sometimes it helps to reconstruct roadkill fossils like the Washington University Rhamphorhynchus WU 970001.
Figure 6. The Washington University Rhamphorhynchus WU 970001 has a pes similar to that of the larger trackmaker at the Wuerho tracksite.

Other Solnhofen and pre-Solnhofen taxa gave rise to Cretaceous taxa in the LPT. Rhamphorhynchus was the sole exception until now. The present evidence suggests that Rhamphorhynchus-derived taxa also survived into the Early Cretaceous Wuerho refugium of China along with the last stegosaur, Wuerhosaurus.

According to Wikipedia,
Wuerhosaurus is a genus of stegosaurid dinosaur from the Early Cretaceous Period of China and Mongolia. As such, it was one of the last genera of stegosaurians known to have existed, since most others lived in the late Jurassic.”

So… there’s at least one more late surviving taxon
that somehow survived into the Junggar Basin of Wuerho and nowhere else (so far).

In an effort to make pterosaur tracks more easily identified,
Peters 2011 produced a catalog of dozens of pterosaur pedes and ichnospecies, but that work was not cited by Li, Wang and Jiang.

The LPT scores relatively few pedal traits
and ichnites can be difficult to trace and interpret. Given that proviso, surprisingly these two ichnite taxa would rather nest within the genus Rhamphorhynchus than closer to the genus Noripterus or any other of 258 taxa in the LPT. Even with a scoring handicap.

So, let’s keep an open mind
with regard to this wonderful new slab from the Early Cretaceous of China (Fig. 1). Let’s not exclude any taxa a priori, nor attempt to force fit the ‘pterosaur next door’. The slab may preserve evidence of an overlooked late survivor from the Late Jurassic, the only Rhamphorhynchus descendants that left clues to their existence in the Early Cretaceous refugium of Wuerho, China.

Li Y, Wang X and Jiang S 2021. A new pterosaur tracksite from the Lower Cretaceous of Wuerho, Junggar Basin, China: inferring the first putative pterosaur trackmaker. PeerJ 9:e11361 DOI 10.7717/peerj.11361
Peters D 2011. A Catalog of Pterosaur Pedes for Trackmaker Identification.Ichnos 18(2):114-141.

Why does a starfish resemble an octopus in ventral view?

Turns out the answer is both homology and convergence.
Those huge and fascinating starfish and octopus ‘arms’ are oversize ‘lips’ surrounding a central mouth. This trait is also found, much reduced, on the last common ancestors of both taxa, nematodes and lanclets (Figs. 1-3). On the other hand, the ‘suckers’ that line starfish and octopus ‘arms’ are convergent. Their last common ancestors do not have suckers, but one might have homologous precursor structures (Figs. 4, 5). Or not.

Figure 1. Starfish in oral view.

The starfish is a radial animal
with a mouth at the center of five radiating arms that help it grasp prey, travel across the sea floor and draw food to the mouth.

Figure 2. Octopus in oral view. The apparentlly missing arm where the label is located has been flipped under the octopus and exits the top of the frame of this picture.

The octopus is a radial animal
with a mouth at the center of eight radiating arms that help it grasp prey, travel across the sea floor and draw food to the mouth.

had a bilateral, worm-like last common ancestor that evolved in two different directions: echinoderm and mollusc. Then some members of both clades expanded their radial mouth parts, reducing and deleting other body parts.

Earlier we looked at the last common ancestor
of molluscs and echinoderms, enoplid nematodes (Fig. 3, genus Enoplus). We also looked at the most primitive chordates, hagfish (Fig. 3, genus Myxine) and the most primitive molluscs, garden slugs (crawling bilateral molluscs without shells, Fig. 3). These taxa all share the same worm-like, bilateral body plan. All three depend on external secretions of mucous either to slide upon (nematodes and mollucs) or to eject in vast quantities when disturbed (hagfish). Neither the starfish, nor the octopus depend on external mucous secretions. The ‘ink’ sac of cephalopods creates highly concentrated melanin, not mucous.

Figure 1. Nematodes, hagfish and slugs have so many traits in common, one wonders why they are not related to one another.
Figure 3. Nematodes, hagfish and garden slugs have so many traits in common, one wonders why they are not related to one another in college textbooks and lectures.

In the video below,
octopus expert Danna Staaf begins octopus evolution with a snail-like cephalopod with a small, chambered, conical shell and tiny ‘arms’. That’s the traditional story. And it’s probably true. But what came before such shelled molluscs? Answer: Bilateral worm-like molluscs without shells (Fig. 3). Slugs tie snails to mussels and cephalopods.

Both starfish and octopus ‘arms’ are large, fleshy, mobile projections
at the circumference of the oral cavity. Such structures are first found in nematodes, like Enoplus (Fig. 3), and basal chordates, like hagfish (Fig. 3) and lancelets (Figs. 4, 5) where they are soft, tiny and delicate at their genesis. On lancelets, these external sand filters are called oral or buccal cirri. In starfish and cephalopods, these grow to enormous sizes as they take on more duties. At the same time, the rest of the body shrinks, more so in starfish than in octopuses, which develop eyes, brains, siphons and other body parts not found in starfish. Even so, in lancelets, starfish, slugs and octopuses, the oral cavity remains ventral.

Figure 4. Extant lancelet (genus: Amphioxus) in cross section and lateral view. The oral cirri are here labeled ‘sand filters’.

Starfish and octopus arms are not simple structures.
Both sets of arms are complicated ventrally with flexible suckers. Are these homologous? Not exactly. A close view of buccal cirri in lancelets reveals pointed lateral projections that act like sand screens. At present it appears that suckers developed convergently in starfish and octopuses probably de novo between those tiny projections.

Figure 5. Lancelet cirri under magnification. From these many strands/ arms surrounding the lancelet oral cavity starfish and octopus arms evolve. Arrows point to lateral projections that are not suckers in any sense, but occupy similar positions at right angles to the strands/ arms.

These are not the only related taxa to expand the mouth
(= oral cavity) and reduce the rest of the body. Tunicates and crinoids also do so, but in different ways, like expanding the atrium. We looked at their strange body plans earlier here.

Abel and Werneburg 2021 discuss tetrapod skulls without a valid phylogeny

Summary for those in a hurry:
Please, please, please… BEFORE you start promoting any pet hypothesis, stop what you’re doing and start with a valid cladogram.

Abel and Werenburg 2021
are “Pulling a BIG Larry Martin” by focusing their classification efforts on just the temporal region without knowing how tetrapods are interrelated.

Example 1: The authors have no idea that diapsid-grade skulls arose twice within Tetrapoda.

Example 2: The authors recognize invalid clades (e.g. Parareptilia) still taught in textbooks and lectures.

The Abel and Werenburg 2021 published cladogram
(Fig. 1) promotes many other modern myths they willingly accept without testing.

Worse yet,
no phyogenetic patterns emerge from their cladogram. Their proposal produces results no better than random scattering.

Figure 1. Cladogram from Abel and Werneburg 2021, here split in two for legibility, promotes almost every taxonomic myth while displaying their demonstrably random skull nomenclature proposal.

From the abstract:
“Here, we introduce a novel comprehensive classification scheme for the various temporal morphotypes in all Tetrapoda that is independent of phylogeny and previous terminology and may facilitate morphological comparisons in future studies.”

I know they’re trying to help, but without a valid phylogeny the authors are promoting fiction.

From the introduction:
“Additionally, the temporal region can vary distinctly in morphology among closely related taxa (Gow, 1972 [Millerettidae]; Tsuji, Müller & Reisz, 2012 [Microleter]), specimens of the same species (Cisneros, 2008 [Procolophon]; Ezcurra, Butler & Benson, 2015 [Proterosuchia]), or throughout ontogeny (Gow, 1972 [Millerettidae]; Haridy et al., 2016 [Delorhynchus]).”

All the more reason to start with a specimen-based cladogram, like the large reptile tree (LRT, 1865+ taxa) and ‘let the chips fall where they may.‘ Don’t be caught focusing your efforts on just a few traits because convergence is rampant within the clade Tetrapoda.

From the introduction:
“Here, we provide a completely new classification scheme for temporal morphology in both tetrapod groups (amphibians and reptiliomorphs), enabling us to discuss the diversity of
the temporal skull region without adding confusion by expanding or modifying the vast number of previous perspectives.”

Without a valid cladogram, Abel and Werneburg actually add confusion to current taxonomy and systematics. When two unrelated taxa (e.g. mammals and turtles; cynodonts and ichthyosaurs) have similar skull morphologies that are given the same skull description (e.g. suprafossal, suprafenestral) confusion is the result (Fig. 1). There is no need to introduce new names for convergent morphologies. We want to get away from naming non-homologous structures.

Abel and Werneburg provide a history of temporal clade nomenclature.
None of the historical references include a valid cladogram due to massive taxon exclusion, as documented by the LRT, which minimizes taxon exclusion. So this is two authors’ look at decades of mistakes without providing a solution based on rock-solid phylogeny.

Abel and Werneburg display their need for a valid cladogram
when they unambiguously state, “The ancestral morphotype of Amniota is ambiguous (Piñeiro et al., 2012) and highly dependent on the nesting of certain key clades (see Section III.1e). The ancestral amniote could have possessed a scutal skull (13 in Fig. 5), as traditionally assumed, or an infrafenestral as seen in early Synapsida and Parareptilia (e.g. Romer & Price, 1940; Cisneros et al., 2004, 2021).”

At this point Abel and Werneburg should have looked at each other and said, “What are we doing? We have no idea how reptiles are organized! Before we type another word we need to create a valid phylogeny that minimizes taxon exclusion. We need to do the work!”

It only gets worse:
more nomenclature, less clarity with every new paragraph in this paper.

If you’re a paleontologist with nothing better to do,
try producing a wide gamut phylogenetic analysis based on hundreds of taxa and traits. Take a few years off to do this. Then report what your cladogram recovers. That will help your profession and your science. Don’t focus your attention on temporal fenestrae or any other trait (e.g. ankles, scutes, wings, dorsal fins, teeth, etc.). Don’t promote old myths. Test them.

establish a broad, overall view of tetrapod interrelationships. Then, with the authority your new cladogram provides, only then allow your focus to fall on smaller subsets. You’ll never have another taxonomic problem like the dozens Abel and Werneburg discuss throughout their paper. That’s essentially gossip. Step in and provide actual assistance.

You chose paleontology because you wanted to make discoveries.
Some are made in the field. Others are made in the cladogram.

Abel P and Werneburg I 2021. Morphology of the temporal skull region in tetrapods: research history, functional explanations, and a new comprehensive classification scheme. Biol. Rev. (2021), pp. 000–000. 1 doi: 10.1111/brv.12751

Cambroraster: documents how some flatworms became trilobites

Moysiuk J and Caron JB 2019
brought us a new semi-segmented ‘radiodont’ that bridges the gap between unsegmented flatworrms and segmented trilobites: Cambroraster (Fig. 1), a middle Cambrian late survivor of an Early Cambrian or earlier radiation.

Figure 1. Cambroraster falcatus, here compared to a trilobite, a flatworm and Anomalocaris. The flatworm is unsegmented. The trilobite is segmented. The other two are semi-segmented, but otherwise share a long list of traits. Note the extant swimming flatworm moves in much the same way as imagined in anomalocarids, but without semi-segmentation.

Maysiuk and Caron report,
“Cambroraster’s morphology is consistent with a nektobenthic lifestyle. The broad, vaulted H-element is convergent with the head armature of limulids [34], carcinized crustaceans and ‘filter chamber’ resuspension feeding trilobites, but also the head shields of some ostracoderm fish.”

the authors omit segmented trilobites and unsegmented flatworms from their cladogram.

the authors include several segmented, multi-legged, velvet worm-like taxa, including Hallucigenia and Opabinia. Their cladogram indicated these two Cambrian legged worms were ancestral to semi-segmented, legless anomalocarids. Thus, with regard to anomalocarids, the authors presented an upside-down cladogram, with primitive legless taxa at derived nodes. This comes from taxon exclusion: excluding flatworms, in this case.

The authors conclude,
“The inferred ecology of Cambroraster indicates that the evolution of large nektobenthic consumers, alongside smaller carnivores like trilobites, occurred in tandem with the radiation of these prey, in line with hypotheses emphasizing the catalysing effects of escalation during this radiation. As large and abundant nektobenthic carnivores, hurdiids like Cambroraster likely had a considerable impact on the local benthic community through both predation and bioturbation.”

The authors mistakenly assume
that Cambroraster (and other anomolacarids) used its oral cone as a predatory organ for biting large prey. Others have considered this unlikely given the largely immobile circular construction of the oral cone ventral to the cephalon, as in flatworms and trilobites (Fig. 1). Rather than a predator, Cambroraster likely fed like a trilobite on immobile, ubiquitous and defenseless algal mats.

What and how did trilobites eat?
According to “There has been a long history of speculation about the feeding habits of trilobites, ranging from predators, scavengers, filter-feeders, free-swimming planktivores, and even parasites or hosts of chemoautrophic symbionts. Using modern-day crustaceans as an analog, it is reasonable to suggest that the majority of trilobites may have been predator-scavengers, as the majority of marine crustaceans are today. Fortey and Owens suggest indicate a shift away from predation and into particle feeding, which includes scavenging for bits of benthic detritus (as the group of olenids below might be doing), or perhaps grazing on beds of algae.” On their well-researched webpage, provides other forms of feeding on tiny and/or buried prey or particles.

Carapace before limbs
Cambroraster displayed a carapace and began to produce segments before it had legs. That means Cambroraster was producing chitin, the material that covered the limbs of trilobites. That’s the next transitional taxon to look for: a Cambroraster with tiny segment buds arising near the ventral midline (Fig. 1).

We first looked at the origin of Anomalocaris
and trilobites via flatworms earlier here.

The traditional origin of trilobites according to
The Early Cambrian included several orders of trilobites, so their genesis among artrhopods must have been earlier, in the Ediacaran. “Probably the key distinguishing character, one that also allowed trilobites to be preserved so well (and which accounts for their sudden prominence in the Cambrian), is calcification of the exoskeleton.”

Two opposing questions arise:
1. Did trilobites arise from slender velvet worm-types, essentially outer tubes over internal tubes with one leg pair per segment and a terminal mouth, and thereafter widen and flatten and rotate the oral cone ventrally to become a trilobite? Or…

2. Did trilobites start off as wide flatworms with a ventral oral cone and thereafter segment their bodies and grow legs and gills arising from each segment?

We already know
that all segmented worms arose from unsegmented round worms (nematodes with a mouth on one end and an anus on the other), and all nematodes arose from flatworms (with a more primitive mouth = anus). The transition from flatworm to trilobite via anomalocarid skips the roundworm and segmented (annelid) worm steps.

So, is this true?
Was there a third flatworm radiation that went directly to segmentation without rotating the mouth anteriorly while developing a separate anal opening posteriorly. The present evidence indicates this is a strong possibility.

Unfortunately, all trilobite clade members are now extinct,
but that fact may be part of today’s solution. If the present hypothesis (that trilobites, anomalocarids and Cambroraster were all gentle algal mat grazers) is correct, then the extinction of this clade can be explained by the near extinction of the algal mat during a planet-wide extinction event, like the end Permian. Thereafter algae could have slowly returned from distant and remote refugia not inhabited by now extinct trilobites and their relatives. concludes:
“The likely scenario is that trilobites arose from Precambrian bilaterians, arguably arthropods, that gave rise to Cambrian arachnomorphs, among them trilobites. The evidence is neither clear nor unambiguous. The fossil record is spotty, but suggestive. Perhaps it is the simple, dorsally unsegmented Precambrian fossil, Parvancorina, that offers the most reasonable link to arachnomorphs.”

Figure 2. Bigotinella is an Early Cambrian trilobite lacking a post-cephalon. Parvancornina is from the Ediacaran, but that may be too soon for trilobite legs. Neither of these taxa preserve segments. These two may be alga-eating semi-segmented flatworm descendants with a ventral mouth and a dorsal shield, like Cambroraster.

Parvancorina bears a distinct resemblance to Cambroraster.
Moreover, Early Cambrian trilobites seem to be known better from head shields lacking ‘post-cranial’ thorax material (Figs. 2, 3). Note: This is what we see in Cambroraster (Fig. 1).

Figure 3. Early Cambrian trilobites are often, as shown here, known only by their hard parts, their cephalon. That means the post-cephalon was soft, as in Cambroraster and flatworms.

If the evidence points in a certain direction
maybe that’s where we should be looking for more data. Perhaps someone else can dive a little more deeply into this issue given this new direction.

The takeaway:
As usual, add a few taxa and see where it take you.

Hagadorn JW 2009. Taking a Bite out of Anomalocaris. In Smith MR, O’Brien LJ, Caron J (eds.). Abstract Volume. International Conference on the Cambrian Explosion (Walcott 2009). Toronto, Ontario, Canada: The Burgess Shale Consortium (published 31 July 2009).
Moysiuk J and Caron JB 2019. A new hurdiid radiodont from the Burgess Shale evinces the exploitation of Cambrian infaunal food sources. Proc. R. Soc. B 286:20191079.


Publicity for Cambroraster:
‘Millennium Falcon’ predator soared across ocean floor at dawn of animal life
By Joshua Sokol Jul. 30, 2019 , 7:01 PM

“The ‘ship’ was one of the largest known animals of its day to churn up the sea floor. It sailed in fleets over muddy ocean sediment, plying its unusual claws in the hunt for small prey.

“Cambroraster had a round mouth lined with toothlike plates, fronted with comblike claws it could hold out like a basket. Its eyes sat in deep notches that give the carapace its signature “spaceship” look.

“Hagadorn said the most likely diet of Anomalocaris was similar to that of modern arthropods such as crabs, lobsters and shrimps, which mostly eat soft items such as worms in the mud or microorganisms or plankton in the water. It could have eaten very small trilobites and recently molted trilobites whose new shells had not yet hardened, but the vast majority of trilobites would have broken Anomalocaris’ mouth parts.”

Bird and pterosaur palates compared

Short one today.
See figure 1 to compare primitive and derived bird and pterosaur palates. Lots of convergence! In each clade, one an archosaur, the other a lepidosaur, former tooth-bearing bones create a solid palate, meeting each other at the midline, as the rostrum comes to a sharp point, and the vomers gradually disappear.

these are cherry-picked taxa chosen for maximum disparity within their clade and maximum convergence between clades.

Earlier we looked at
the evolution of pterosaur palates in an 8-part series, ending here. Links there will lead you to the first seven parts. The evolution of bird palates is in process.

Housekeeping the bird subset of the LRT, part 3

Another several days and nights of more binge study trying to figure out
the topology of the crown birds recovers a slightly more parsimonious hypothesis of interrelationships in this subset of the large reptile tree (LRT, 1865+ taxa; Fig. 1). A few surprises were recovered (see below) and I’m starting to understand why paleontologists gave up studying phenomic traits in bird phylogeny and turned instead to genomic molecules for their cladograms. Earlier struggles with the clade of crown birds can be found here and here.

Figure 1. Revised bird subset of the large reptile tree. Colors indicate morphology and niche. Note the major division between largely land and tree birds vs largely water birds.

Problem number 1:
Most birds more or less fuse the premaxilla, maxilla and nasal bones. Cornified tissues (a keratinous beak) sometimes covers the anterior rostrum further obscuring underlying sutures. The extent of the premaxilla vs maxilla varies greatly among birds. Sometimes the two are laminated one atop the other. My bird palate data needed a closer look. Many corrections were made based on comparative anatomy.

Problem number 2.
Although the cladogram (Fig. 1) is fully resolved, the ‘backbone’ of the cladogram still needs to be further strengthened for higher Bootstrap/Jackknife scores. This is probably not the ‘most’ parsimonious tree possible, but a ‘more’ parsimonious tree than previous presentations.

Many surprises popped up,
resolving several issues:

Surprise number 1:
Mousebirds (Urocolius) + quetzals (Pharomachrus), neither of which have long legs, split off early from the rest of the Neognathae (non-ratites) along with tiny Cyrilavis and headless Palaeoglaux, both from the Eocene. Based on phylogenetic bracketing, I hope to find a long-legged mousebird ancestor someday.

Figure 2. The skulls of the ibis (Threskiornis), hoopoe (Upupa) and grackle (Quiscalus) compared. Also shown is a hatchling ibis showing the short, grackle-like rostrum. Paedomorphosis likely creates derived neognath birds because most are small and have short legs and short beaks, like the hatchlings of more primitive neognath birds.

Surprise number 2:
The long-legged wading ibis (Threskiornis) now nests with the short-legged grackle (Quiscalus) and the colorful hoopoe (Upupa, Fig. 2). Other workers nested the ibis with pelicans and the hoopoe with either hornbills or kingfishers.

Figure 3. At left and inset photo: the screamer, Chauna, compared to the extinct and flightless solitaire, Pezophaps, at right. Now these two nest together in the LRT.

Surprise number 3:
The extinct and flightless solitaire, Pezophaps, is related to the extant screamer (Chauna, Fig. 3), and both arise from more primitive pigeons (Columba). So these are giant pigeons with long legs, a reversal recalling traits from more primitive taxa.

Surprise number 4:
Pigeons now arise from the corn crake (Crex) via the sand grouse (Pterocles). Derived taxa include increasingly more plant matter in their diet. Screamers (Fig. 3) are herbivores, but will feed their young with small captured animals.

Surprise number 5:
Long-legged cranes (Cicconia) are basal to short-legged grebes (Gavia) and flightless penguins (Aptenodytes).

NOT a surprise number 1:
New World vultures (Vultur) now nest with the osprey (Pandion). These should have been nested together earlier.

Don’t be surprised
to see future changes in the bird clade of the LRT. Based on past experience, the present cladogram (Fig. 1) still needs a bit of polishing. The weakness has never been in the LRT, only in my ability to process the subtle and sometimes barely visible data found in birds. Having a catalog of data for rapid and thorough comparative anatomy in the form of constantly updated images at has been the key to bringing this ongoing study more and more into focus. Apologies for earlier oversights. I’m learning as I go without much guidance, because there just isn’t that much guidance out there. And there’s lots of convergence.

Added a few days later:
so charming!