4 pakicetids enter the LRT starting with a new one: Phiomicetus (Gohar et al. 2021)

Gohar et al. 2021 described
a new “protocetid,” Phiomicetus anubis (Fig.1).

Figure 1. Skull of Phiomicetus from the Middle Eocene, nesting basal to pakicetids.

From the Gohar et al. abstract
“Over about 10 million years, the ancestors of whales transformed from herbivorous, deer-like, terrestrial mammals into carnivorous and fully aquatic cetaceans.

Two problems in one opening sentence here:
1. Whales (traditional order Cetacea) are not monophyletic.
2. No deer-like mammals are ancestral to any of the whales according to the large reptile tree (LRT, 1915 taxa) which minimizes taxon exclusion. Overlooked Rhynchocyon (Figs. 5, 6) comes pretty close in limb morphology, but it is a leptictid, not an artiodactyl.

Figure 2. Andrewsiphius skull, humerus and femur with missing parts added based on Indohyus (small skeleton at left). This taxon moves out of whales with legs and nests now as a giant sister to Tenrec (upper left corner).

From the abstract continued
Protocetids are Eocene whales that represent a unique semiaquatic stage in that dramatic evolutionary transformation. Here, we report on a new medium-sized protocetid, Phiomicetus anubis gen. et sp. nov., consisting of a partial skeleton from the middle Eocene (Lutetian) of the Fayum Depression in Egypt.”

Always good to see new taxa!

Figure 3. Skeleton of Tenrec alongside restored skeleton model of Pakicetus.
Figure 3. Skeleton of Tenrec alongside restored skeleton model of Pakicetus. Someone should have recognized the similarity long before the LRT, yet years later, this still goes unnoticed by Gohar et al. 2021.

From the abstract continued
The new species differs from other protocetids in having large, elongated temporal fossae, anteriorly placed pterygoids, elongated parietals, an unfused mandibular symphysis that terminates at the level of P3, and a relatively enlarged I3.

Unique features of the skull and mandible suggest a capacity for more efficient oral mechanical processing than the typical protocetid condition, thereby allowing for a strong raptorial feeding style.

All good information, but now it is time for a phylogenetic analysis… The cladogram from Gohar et al. 2021 does not include any taxa basal to Pakicetus, so Phiomicetus can’t nest basal to Pakicetus and Andrewsiphius (Fig. 2) can’t nest with Tenrec (Fig. 2). You have to tell PAUP which taxon is the outgroup (= the basalmost taxon). If you tell PAUP Pakicetus is the outgroup taxon, then the other taxa that would have or might have nested basal to Pakicetus have to find another node to nest in. That’s why the LRT goes back to Ediacaran worms as outgroup taxa.

Figure 4. Cladogram from Gohar et al. 2021 nesting Phiomicetus within Pakicetus because no taxa basal to Pakicetus are offered. This is taxon exclusion based on assumption, rather than testing. Mammalodon is mistakenly included. It actually nests between hippos and desmostylians in the LRT. Andrewsiphius likewise should nest in the outgroup of these taxa.

From the abstract continued
Phylogenetic analysis nests Phiomicetus within the paraphyletic Protocetidae, as the most basal protocetid known from Africa.

In the LRT (subset Fig. 5) Phiomicetus also nests basal to protocetidae and pakicetidae. An overlooked clade of anagalids, leptictids and tenrecs are outgroups to Pakicetus in the LRT along with three pakicetids. The tenrecs, anagalids and leptictids are traditionally omitted from cladograms. Sometimes, hippos, cows and pigs are inserted as outgroup taxa. If only whale expert Phillip Gingerich had let The Triple Origin of Whales get published several years ago Gohar et al. could have used this validated set of outgroup taxa.

Figure 5. Subset of the LRT focusing on the tenrec-odontocete clade.

When Phiomicetus is added to the LRT
(Large Reptile Tree, 1915 taxa; subset Fig. 5) Protocetidae remains monophyletic, and Phiomicetus nests outside the Protocetidae in the proximal outgroup, the one including Indohyus (Fig. 2), Tenrec (Fig. 3) and relatives. Remingtonocetus (Fig. 8) and Georgiacetus were also added… and they nested with Phiomicetus. Protocetus (Fig. 9) was also added… and it nested with Zygorhiza in the archaeocetes. That tells us useful legs were absent, perhaps vestiges, in Protocetus.

Figure 7. Rhynchocyon, a living elephant shrew, is a living leptictid.
Figure 6. Rhynchocyon, a living elephant shrew, is a living leptictid.
Figure 4. Skeleton of the elephant shrew, Rhynchocyon. Note the digitigrade manus and pes, like those of basal artiodactyls.
Figure 7. Skeleton of the elephant shrew, Rhynchocyon. Note the digitigrade manus and pes, like those of basal artiodactyls, tenrecs and pakicetids.

From the abstract continued
Recovery of Phiomicetus from the same bed that yielded the remingtonocetid Rayanistes afer provides the first clear evidence for the co-occurrence of the basal cetacean families Remingtonocetidae and Protocetidae in Africa. The discovery of Phiomicetus further augments our understanding of the biogeography and feeding ecology of early whales.”

Andrewsiphius
(Fig. 2) is considered by Gohar et al. a member of the Remingtonocetidae (Fig. 8). The LRT nests Andrewsiphius as a tenrec, and Remingtonocetus basal to Pakicetus. The exclusion of valid more basal taxa is the problem in the Gohar et al. cladogram. Workers still consider the deer, cow and pig basal to whales, another embarrassing tradition borrowed and borrowed again by whale workers.

So where do odontocetes come from?
Anagalids, leptictids and tenrecs produced a bush of aquatic taxa with a large head and long snout: the pakicetids, archaeocetids and finally odontocetes. The extant leptictid, Rhynchocyon (Figs. 6,7), already has that long pointy snout and deer-like, whale-like ankles, but it has been traditionally omitted by whale experts. Same with the tenrecs.

Figure 8. Remingtonocetus skull in diagram and photo views. The photo shows the lateral process of the premaxilla contacting the anteriorly extended frontals, distinct from the diagram, which also gets rid of the left angle of the rostrum. This asymmetry improves echolocation sensing, as in tenrecs and odontocetes.
Figure 9. Protocetus skull. This first protocetid described (Fraas 1904) and recent addition to the LRT nests with Zygorhiza.

Protocetus was the first protocetid described
(Fraas 1904; Fig. 9) and nests with the archaeocete, Zygorhiza, in the LRT, far from the base of walking whales.

The traditionally omitted outgroups to whales in the LRT
(Fig. 5) appear to represent a novel hypothesis of interrelations. If there is a prior citation, please share it here so I can promote it. The LRT minimizes taxon exclusion, which makes it an ideal tool in cases like this where you want to know the outgroup to any vertebrate clade. That includes turtles, snakes and pterosaurs.

Sadly,
some paleontologists still hate it whenever the LRT recovers overlooked and omitted outgroups. Borrowing old and invalid cladograms, and trusting them, is a habit common among academic authors. That extends to authors of textbooks. According to one popular paleontologist who feels it necessary to occasionally harangue workers who do put in the effort to create wide-gamut cladograms, “It’s too much work.” So don’t expect things to change soon in paleontology.

References
Gohar et al. (7 co-authors) 2021. A new protocetid whale offers clues to biogeography and feeding ecology in early cetacean evolution. Proceedings of the Royal Society B https://doi.org/10.1098/rspb.2021.1368

18 thoughts on “4 pakicetids enter the LRT starting with a new one: Phiomicetus (Gohar et al. 2021)

  1. “According to one popular paleontologist who feels it necessary to occasionally harangue workers who do put in the effort to create wide-gamut cladograms, “It’s too much work.” So don’t expect things to change soon in paleontology.”

    This is disingenuous in the face of your refusal to expand your datasets to incorporate character evidence contrary to your findings. There’s no point in giving specifics, as I and many others have criticized this ad nauseum. I am simply taking this moment to comment so that passers by don’t assume you’ve actually taken the effort to examine these firsthand or considered your character evidence against what has been published. I will not be surprised if you refuse to post this comment as you have a track record now of leaving mine in the moderation queue.

    • Dear “N”… can you _please_ be specific? This is disingenuous in the face of your refusal to do so. You want me to “incorporate character evidence contrary to [my] findings.” Why? You want me to incorporate a styliform in Yi qi when that’s a misinterpretation? See? I’m being specific. You need to do this. Otherwise I have no idea what you’re whining about. That’s troll behavior. It is beneath the standards of this blog and those of major and minor academic publications. I am giving you a break here. If you have a track record of having your comments left in the moderation queue you must be using a new pseudonym today. I have never seen a comment from “N” until now. Don’t be whiner. Don’t be like Darren ‘N”aish. I still have hope for both of you. Put down your outrage and be specific. Without whining, please discuss cladograms, characters and taxa like the others who get their comments posted. Those who don’t get posted usually get an email explaining ‘the rules.’ So if you have a track record, you probably know better. Seek medical or psychiatric attention if this doesn’t clear up soon. Get back to science ASAP. I want to hear valid and specific criticism so I can make repairs as needed.

      • I was specific when you posted your crocodyliform trees. I indicated several characters that might change your results. Have you added them?

      • Hi Chris,

        The answer is no. When I said, please be specific, I wanted the reader to tell me where a mistake was made on character scoring or taxon nesting. Blackwashing and whining about the whole project doesn’t get us anywhere. On the same dime, if you offer a list of characters that MIGHT change results, that tells me you are guessing, too. Instead: tell me why one taxon should not nest with another. Perhaps a derived character shows up too early in a cladogram. Perhaps I overlooked some trait that moves a taxon out of node A and into node B. Show me a cladogram that shares all the pertinent taxa that recovers a different tree topology, then let’s check the scores for accuracy. One of us, or both of us, could be wrong. Show me a quadrate that curls left instead of out. That sort of specific thing. In other words, you weren’t being specific. You were being hopeful and general. Finally, I’ve been told to add characters since this cladogram had only 360 taxa. Don’t you think that argument has lost a little steam by now as it moves past 1900 taxa?

  2. First of all – that your data set now has 1900 taxa makes the addition of characters far more important, not less. No steam lost. In fact, the steam is building.

    Second – Gavialis and Gryposuchus should not be nesting with any of the thalattosuchians or tethysuchians. It’s an absurd result. There are many characters that would support this, such as the presence of more than two rows of dorsal osteoderms, complete encirclement of the internal choana by the pterygoids, procoelous vertebrae, the orientation of the median and lateral eustachian canals, pneumatization of the quadrate, and the presence of a mandibular fenestra. I may have listed more at the time of the discussion – you’ll have to check.

    When it comes to the mixture of marsupials and placentals in your trees, one would ask whether features shared by marsupials in the dentition (number of incisors, premolars, molars), the reflected lamina of the angular, the peculiar construction of the “petrosal,” and the epipubic bone are included. Not to mention the pouch, which is at least known in living forms. Likewise, placentals have their own dental characters, their own petrosal configuration, and additional details of the bones surrounding the ear. Have these been included? If not, why should we take your results seriously?

    You keep trying to put the ball in our court, but it’s never there. You’re presenting results that, to put it mildly, contrast with what the rest of us have been recovering. We keep pointing out that you have way, way too few characters to accomplish what you’re trying, and some of the characters you’re trying to include (e.g. for dermal bones in sharks, even though sharks have no such structures, nor do they have precursors) are positively misleading and should be left uncoded. We additionally point you to publications that include more characters.

    The ball is still in your court. It’s never left. You’ve been informed of the problems with your data set, and you refuse to accept anyone’s advice. So please let us know when you’ve added characters.

    • re: “makes the addition of characters far more important” Perhaps this is the overlooked difference in what you were taught vs. the factual results of the LRT, still able to lump and separate taxa, which is the one and only goal of a phylogenetic analysis. You are dealing with subtle differences between many dozen closely related taxa, where the number of rows of scutes is important. The LRT is dealing with much larger differences between fewer and less closely related taxa.

    • re: Gavialis and Gryposuchus… I only have cranial data for Gryposuchus. In my data both have a mandibular fenestra. So does Pelagosaurus. The rest is detail not avialable in the data. Question for you? Is Dyoplax in your taxon list. Traits are not what makes a clade. They come and go. Only a last common ancestor makes a clade. This whole ten-year experiment has been pointing out taxon exclusion issues from fish to mammals.

    • re: marsupials and placentals. You’re thinking supragenerically, not phylogenetically and generically. Be specific with genera. Try not to use broader taxa, like Marupialia. You’re also cherry-picking a few traits hoping to pull one Jenga block out and have the whole tower crumble. “If not, why should we take your results seriously?” You didn’t even wait for the answer before telling the jury not to take my results seriously. Statistically (which is the way PAUP works) that’s not going to happen when over 200 traits are scored. Remember, the last common ancestor of all placentals is a marsupial in the LRT. So those two are going to share virtually all traits at that node. Monodelphis lacks a pouch, yet is still considered a marsupial. In the LRT it is the last taxon prior to placentals. I don’t care what Monodelphis is considered. In the LRT it is part of an evolutionary continuum that leads to other taxa that also don’t have a marsupium. Also remember that reversals do happen. Some humans have a double vagina. Odontocetes have simple, single cusp molars. Etc. etc. Yes, epipubes are traits in the LRT, where both pterosaurs and marsupials have them by convergence. The petrosal is not tested as it is not visible on enough fossil taxa, which are typically crushed.

      • As far as ‘more characters’ goes, what do paleontologists say? “Beyond the scope of this study.”

        On the other hand, “More taxa” is the entire scope of this study.

        Your results may differ. And that’s okay.

    • re: “contrast with what the rest of us have been recovering.”

      I have suggested the addition of a few taxa, and workers refuse or are unaware. Peters 2000 showed that simply adding a few taxa pulled pterosaurs out of the Archosauria, where they were lost without any similar sisters. Twenty years later workers are still refusing to add those taxa. What’s up with that? No matter how many extra traits anyone uses unrelated taxa will never nest together IF given a more parsimonious alternative. The rest of you need to add a few taxa, not trash the guy who added a few taxa.

      re: “way too few characters to accomplish what you’re trying”

      This is false. I am trying to build a family tree of vertebrates and am doing so with the building materials I have. You can see this with your own eyes, yet refuse to believe what you are seeing. No more characters are necessary. I can’t help you if you don’t want to see what is.

      re: “you refuse to accept anyone’s advice”

      Advice = opinion. I need evidence. That’s all. Just evidence. The shark advice you offered is no advice. I need skull traits for sharks to be included in the LRT, whether ossified or not. The traits I have chosen to use are topographical. You have provided no alternatives. Referring to the above topic on taxa being attracted to one another, I can’t be attracted to your hypothesis because you provide no alternative competing hypothesis for homologous shark skull topography. Offer a working alternative. I’ve made 150,000 corrections to the LRT and imagery in ReptileEvolution.com, all based on evidence and understanding (= fresh insight). This is all I do: examine evidence and place it into the LRT.

  3. What’s so difficult to understand about the statement “there are no dermal bones and no structures homologous with them in the skulls of sharks?”
    It is as straightforward and scientifically accurate as saying analogously “humans do not have uropygial glands.”
    Or are you suggesting that all structures seen in all vertebrates are homologous with all structures present in their common ancestor?
    You do accept the hypothesis that novel structures can arise during the evolution of a lineage, no?
    I would like to think that you do, as it’s what the fossil and anatomical evidence conclusively show, but at times when you talk about your tree and “gradual accumulation of traits” (which is not what parsimony analyses construct trees on the basis of) and the asserted lack of need for any additional, more detailed characters, it implies that synapomorphies either don’t exist or should be ignored. As such, you would be removing or excluding the exact data that parsimony analyses do use as the basis for generating trees.

    • re: What’s so difficult to understand about the statement “there are no dermal bones and no structures homologous with them in the skulls of sharks?”

      It is difficult to understand, given that something in the pre-gill area of sharks is supporting nostrils, eyes, jaws, brain, etc. That area, (can we call that a “head”?) has pockets, processes and flat areas. That needs to be described using tetrapod homologs if the cladogram is going to produce useful phylogenetic results that include sharks. [I can’t believe I’m having to defend this. Makes me want to study something real, like quantum physics]

      Question for you: What is the difference between “a gradual accumulation of derived traits” and “novel structures can arise during the eovlution [sic] of a lineage?” The first you reject. The second you promote.

      The need for any additional, more detailed characters is demonstrably not needed because the LRT topology already lumps and separates well enough. Did you notice that all the lepidosaurs nest with lepidosaurs? And yet scalation traits are not scored. That all the placentals nest with other placentals, yet mammary glands are not scored. Neither is diphydonty. Shall I add “occipital condyle axis to jawline axis more than 30º”? Shall I add “femoral head angle to shaft greater than 60º? Neither is necessary though both would enrich the cladogram. Also see how no matter how many traits are added, someone will say “not enough traits.” You can describe an elephant down to each wrinkle, but if you don’t add a hyrax, you’ll never know that a hyrax is not too far from an elephant.

      I say, add taxa. Take your time. Do one a day, it’s not that hard.

      What else can I do for you?

  4. A couple of additional points:

    1. Yes, Pelagosaurus has a mandibular fenestra. It excludes the dentary. That’s the sort of detail that needs to be in your matrix if we’re to take your results seriously.

    2. I’ve also seen Dyoplax. It’s so poorly preserved that almost nothing can be said about its morphology in detail. (Published pictures are insufficient to work with it.)

    3. There’s actually lots of postcranial information for Gryposuchus out there.

    4. You dismiss osteoderm patterns as “details,” but I think you’re confusing them with scale patterns. Scale patterns are variable within species and probably don’t hold much signal beyond smaller clades, but osteoderm patterns are highly conserved. They’re widely used in phylogenetic analyses. It isn’t clear why you’re reluctant to use them.

    • re: Pelagosaurus fenestra: That’s the sort of detail that croc experts should use. I have to apply every trait to 1917 taxa. It’s a different task. At present: “mandibular fenestra: present/absent/dual mandibular fenestrae” is about all I can do.

      re: Dyoplax: “almost nothing can be said” Then say nothing. I use excellent data, mediocre data and poor data in the LRT. When better data comes along, I change out data. This is an ongoing project designed from the start to include previously omitted taxa. Every new taxon is something I am meeting for the first time. Freshman mistakes have been made. That is par for the course. Corrections to date: upwards of 150,000. Who else confesses that?

      Because Dyoplax is Triassic it probably should be included in crocs studies. Score what can be scored and leave a “?” where necessary.

      re: osteoderm patterns: Do they apply to any other taxa other than crocs? General traits used in the LRT need to be scored for 1917 taxa. It’s a different task.

      I realize paid professionals have to maintain high peer-reviewed standards. I’ve peeked behind the curtain. You have pressures I will never know. I don’t worry about what others think. I don’t have to follow the textbook. I don’t worry about publish or perish. There’s no money, prestige, travel, etc. in this for me. I’m in this purely for the science, even though it is, and will always be, at a freshman level. You don’t have to accept my contributions or suggestions. Just pretend I don’t exist.

      As far as ‘more characters’ goes, what do paleontologists say? “Beyond the scope of this study.”

      On the other hand, “More taxa” is the entire scope of this study.

      • I think you’ve stated your own problem. Phylogenetic relationships in a particular group might be reflected in systems not found in other groups – dermal skull bones (not expressed in sharks, not even as precursors), osteoderms, digits, ear ossicles, and so on. By excluding these characters, you are denying your analysis the opportunity to recover the relationships they would otherwise support. Adding taxa will do absolutely nothing to correct for this problem.

        What was it you said about Dyoplax? Use question marks? THe same is true here. You absolutely cannot score sharks for dermal bone characters. To do otherwise is to add false “information” that can mislead your result.

        If the relationships among crocodyliforms are primarily supported by character systems not reflected in your analysis, then we are obliged to disregard your results. They’re relying on too coarse a sample of the morphological data that express systematic variation in that group.

      • Chris, you wrote:”By excluding these characters, you are denying your analysis the opportunity to recover the relationships they would otherwise support.”

        — We’re not talking about just a few characters. You are denying that sharks have any cranial characters that can be scored. We have to find a way to score sharks the way they need to be scored, like the other taxa, despite their shortcomings. They need a place ‘at the table’.

        You wrote: “If the relationships among crocodyliforms are primarily supported by character systems not reflected in your analysis, then we are obliged to disregard your results.”

        Fair enough and understood. The LRT is like a map in your glove compartment. You only have to pull it out if you get lost along the way. Thank you for your thoughts.

  5. “We’re not talking about just a few characters. You are denying that sharks have any cranial characters that can be scored. We have to find a way to score sharks the way they need to be scored, like the other taxa, despite their shortcomings. They need a place ‘at the table’.”

    No, I’m not saying sharks can’t be scored for cranial characters. I’m not even saying anything remotely similar to that. They can absolutely be scored for chondrocranial characters, for example. Bony vertebrates have chondrocrania. There are lots of characters held in common – sensory capsules, openings and pathways for nerves and blood vessels, the pituitary fossa, the labyrinth, the endocranial cavity itself, the occipital arches that form around the foramen magnum, and so on. The same is true for the splanchnocranium, as long as you exclude the dermal bones that form around it.

    You say things like “we have to use tetrapod terms for sharks.” Except that using homologous terms only works if the structures themselves are homologous. Since sharks don’t have nasal bones, using characters referring to the nasals is an error. You might as well be using characters expressing the number of phalanges in the thumb – sharks have neither thumbs nor phalanges, and so these characters can’t be scored for them. It’s like using characters expressing variation in the antlers while analyzing primate phylogeny. Trying to use terms like “phalanx” when referring to a shark’s ceratotrichia would be as meaningless as referring to a part of a shark’s chondrocranium as the premaxilla.

    Others have been able to include sharks and bony fishes at the same table. They do so without scoring dermatocranial characters for sharks. I would strongly recommend that you look at other studies that have analyzed these animals and look at the characters they used. Dismissing them on the grounds that they didn’t look at the right taxa would be entirely missing the point here – I’m asking you to look at the characters they used. You should then consider adding them for yourself.

    I have to ask if other readers understand what I’m saying, because I don’t seem to be getting through here.

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.