Reptile Tree CI: too low? …and some thoughts from Dr. E.O. Wilson

A reader brought up the subject (he perceived it as a problem) of the very low consistency index (CI) in the large reptile tree. 340 taxa were tested against 228 characters. The reader thought too few characters were employed and that was the cause for the low CI.

The CI for the entire tree is .099. Very low, but not necessarily bad… as we shall see…

I mentioned in my reply to the reader that this was due to the high level of homoplasy in the reptile tree due to the large number of taxa sharing so many traits. I suggested the CI would rise with fewer taxa and all other variables held constant. That’s easy to do by chopping the large reptile tree apart.

Here are a few CIs focused on clades, using the same character list (228), but far fewer taxa in each case.

Watch those CI numbers rise!

All non-amniote tetrapods: .429

Captorhinids: .875

Permian/Triassic/Cretaceous gliders: .836

Rhynchocephalia (includes trilophosaurs and rhynchosaurs): .659

Tritosaurs (includes fenestrasaurs, pterosaurs): .735

Squamata: .439

Synapsida: .632

Enaliosauria (Claudiosaurus, mesosaurs, ichthyosaurs, thalattosaurs, plesiosaurs, placodonts and kin): .360

Archosauriforms: .267

Euarchosauriforms (no choristoderes, parasuchians, chanaresuchids or Lagerpeton) : .322

Archosauria (crocs and dinos): .417

Crocs: .705

Dinosauria (includes poposaurs): .523

Hope these numbers help the cause and provide understanding for the very low CI for the entire tree. The beauty of having a large tree is having the ability to easily chop it into many small trees to test various hypotheses. Unfortunately, there’s much more work for small focused trees to test larger gamuts of taxa or characters. Maybe that’s why it isn’t done.

We could always use better characters. I freely admit that some of my characters could use a little help. Some of them weight various traits, but they do it across the board. We could also use more characters, but if we add in the trait of a “carapace present” we would still end up with several convergent instances of that trait, which does nothing to increase the CI number.

It is what it is, dependent on the number of taxa employed and their rampant convergence in the large reptile tree. The data determines the results, not the other way around.

On an editorial note
A recent post engendered several irate responses from readers, most of which included negative attributions and insults. Those had to be trashed. Via private email I encouraged those readers to focus on the taxa, not the author, and make specific suggestions as to how to improve the taxon reconstructions. As you know, these then become the data that ultimately produce the family tree. Changes are made weekly. When new valid data come in, they are welcome. Readers who have no time or inclination to provide fresh data, yet find plenty of time to rant without substance, are not helping my cause or theirs. They end up frustrated. I end up wondering why they can’t provide even a little evidence. I’m led to believe those who do not provide evidence have none.

Hopefully every time I have dismissed a claim or promoted one in these 750 posts, I have provided pictorial and other evidence with references. Disregarding the evidence is your right. Choosing to support bad evidence is also your right. But those who do so run the risk of ultimately looking foolish to villainous when the tide turns. I run the risk of being wrong, but am always willing to right the wrongs. But I need fresh valid data.

I hope you all understand. I’m trying to keep this site and the data it is built upon on a professional level.

The large reptile tree has grown to where no paleontologist has gone before, conquering a larger gamut of reptiles and in doing so demystifying many relationships and former enigmas simply by being ever more inclusive. That’s a good thing, guys. Embrace it! Use it!

Some pertinent thoughts from ant expert, E. O. Wilson:
While generally lauded today, earlier in his career Dr. E. O. Wilson was reviled for some of his novel ideas*. The following Wilson quotes seem pertinent:

“Without a trace of irony I can say I have been blessed with brilliant enemies. I owe them a great debt, because they redoubled my energies and drove me in new directions.”

“Between scientists, you can have high competitiveness and jealousy and petty nit-picking, because we are human. But once something is nailed, the person who did it usually gets the credit, and we move on.” 

“But the best way to do it is – to make discoveries – is to make short imperfect experiments.” 

*(from Wikipedia)
In the water incident, Wilson’s lecture was attacked by the International Committee Against Racism, a front group of the Progressive Labor Party, where one member poured a pitcher of water on Wilson’s head and chanted “Wilson, you’re all wet” at an AAAS conference in November 1978.[38] Wilson later spoke of the incident as a source of pride: “I believe…I was the only scientist in modern times to be physically attacked for an idea.”[39]

27 thoughts on “Reptile Tree CI: too low? …and some thoughts from Dr. E.O. Wilson

  1. Okies The CI…

    Yes indeed it is probably due to the high levels of homplasy. And that’s the problem. The whole point of phylogenetic analysis, the very basis of the method, is to find the topology that minimises homoplasy!!!!! A homoplasy is unparsimonious for that particular character. And that is why the CI is considered an important measure of how strong the support for your tree is. A character list that contains a lot of homoplasy contains a lot of characters that are not useful for defining a clade (clades can only be defined by synapomorphies). You can’t just brush of a low CI like that. It is an important measure of how well supported your tree is.

    Following on from this idea, it is true that adding more characters does not necessarily improve the CI; if you add characters which are homoplasies, it will obviously make things worse. But adding more characters that actually define clades WILL raise it no end. And those are the characters you want! This is why splitting up the tree, but keeping the same 230 odd characters raises the CI for each clade: the homoplasies will now be counted as synapomorphies, since they are present in all taxa in that lineage, and the taxa for which they are homoplaseous will not have been analysed. It will also include characters which are identical for all taxa in each tree, which will raise the CI because of how it is calculated, but which are not useful for defining clades (they should not be included in phylogenetic analyses, which is why they have an odd behaviour with the CI)

    And following from your reply to me in the earlier post: I would not say you needed 3000 characters if you added 1000. I would say you had enough characters once you have a decent CI and better Bremer supports. There is no hard and fast rule about how many characters you need. That is what support metrics are for: they tell you when you need more information.

    Its like any statistical test: if you’re comparing the correlation between two curves, say, you might get a correlation coefficient of 0.65. Is that good or significant? We don’t know just from that one numer, it depends on a variety of things like sample size. That’s what the p value is for: it tells you whether this value of correlation coefficient is good for THIS PARTICULAR DATASET with this sample size. Its the same with phyloegentic analysis: you get however many most parsimonious trees with length of whatever. Is the tree any good? We don’t know just from those numbers, as tree length is not independent of sample size: you need a measure of support. That’s what the CI is for.

      • A few to start with (sorry if I accidentally incluse one or two you already have)

        The following are considered to be characters which define Synapsids, not including them may be why you get a polyphyletic synapsida:

        Postparietals paired or not?

        JAw joint level with occiput or posterior to it?

        Septomaxilla (a bone you don’t seem to have any characters for) shape: multistate – pillar-like, sheet-like

        Septomaxillary foramen: full sized, reduced or absent

        Cervical vertebrae count: less than or more than 3

        Latissimuus Dorsi attachment on humerus: tranverse rudge or posterior directed tubercle

        And a couple which define Sauropsids (to the exclusion of synapsida); not using these might be why Synapsids plot next to archosaurs instead of next to Sauropsids:

        Tabuluars: longer than wide or wider than long

        Supraorbital foramen: absent or present

        Pocket in parasphenoid for cervical musculature: present or absent

        Wings on paraspehnoid, broadening it posteriorly: present or absent

        Medial centrale: present or absent

        That’ll do for now. See what adding these will do, then maybe we’ll look at some others

      • Neil,

        I’ve answered these with regard to Aerosaurus, a basal synapsid in my tree, as best as I could, having never dealt with these traits before. Considering the high level of convergence in the Reptilia, and the polyphyletic nature of the clade Sauropsida in the large reptile tree, I don’t know whether or not we’ll find or should expect to find any list of traits exclusive to one clade.

        The following are considered to be characters which define Synapsids, not including them may be why you get a polyphyletic synapsida:

        Postparietals paired or not? Answer: Aerosaurus not clear.

        JAw joint level with occiput or posterior to it? Answer: Aerosaurus posterior.

        Septomaxilla (a bone you don’t seem to have any characters for) shape: multistate – pillar-like, sheet-like Answer: Aerosaurus not revealed

        Septomaxillary foramen: full sized, reduced or absent Answer: Aerosaurus not revealed

        Cervical vertebrae count: less than or more than 3 Answer: Aerosaurus more

        Latissimuus Dorsi attachment on humerus: tranverse ridge or posterior directed tubercle. Answer: Aerosaurus – must be a ridge. I don’t see a tubercle unless it starts off as that little foramen.

        And a couple which define Sauropsids (to the exclusion of synapsida); not using these might be why Synapsids plot next to archosaurs instead of next to Sauropsids:

        Tabulars: longer than wide or wider than long Answer: Aerosaurus longer than wide.

        Supraorbital foramen: absent or present Answer: Aerosaurus not revealed

        Pocket in parasphenoid for cervical musculature: present or absent. Answer: Aerosaurus not revealed

        Wings on paraspehnoid, broadening it posteriorly: present or absent. Answer: Aerosaurus appear to be present, but not sure what is and what isn’t in this case.

        Medial centrale: present or absent Answer: Answer: Aerosaurus present. There is a long centrale (medial? lateral? both?) in Eunotosaurus, Diadectes, Eocaptorhinus, Labidosaurus, Procolophon

      • Okies I’ve had a check, the centrale in the taxa you mentioned; in all that I could find info on it seems to be the lateral, so absence of a medial centrale still holds.

        However, applying them to Aerosaurus (which both you and the general scientific population agrees to be a Synapsid) Isn’t getting to the nub of the issue. Ideally, you’d want to add these to your matrix, but failing that, look at them in a taxon which you think is not a synapsid, but others think is e.g. Caseids, the various varanopids which come outside synapsids. Do these taxa have the same scores as aerosaurus? If yes, then your oppinion that these taxa are not synapsids is based on a lack of vital characters.

        Again, Do archosaurs have the same scores as Aerosaurus? if not then to positioning of Synapsids with Archosaurs is again based on a lack of characters

      • “I don’t care how they come out.”
        I’m sorry, do I understand you right? Did you actually just say you don’t care how these characters come out? Am I to take it that you will stand by your results no matter how many characters oppose them??

        If I do understand you right, then you have lost the right to ever answer people with the claim “I change my mind.” Do you think characters are not important or something? Do you think there is no way adding new characters can change your matrix? If you do think this, then I really do dispair; your knowledge of phylogenetic analysis is worse than I ever thought.

        “These are the current results subject to new data”

        CHARACTERS ARE DATA. These characters are new data. And if you’re just going to ignore them, then you are inviting the same criticism of blinkeredness and refusal to shift your oppinion that you delight in heaping on others. You’re just proving that no matter how much alternative data (again, new characters are new data) is piled on, you will not change your position. Moreover, you will refuse to even test it!

        “Not an opinion”

        Remember what I said a couple of months ago about arrogance?

      • This is not what I mean. In other words: I’m not trying to support a particular tree. I let the results speak for themselves to tell me what the tree is. It would probably help if you formulate the character you want me to use, because I don’t want to make a mistake. What other results are you hoping the new data will create? Maybe we can focus on that? Which taxa should shift to which other node?

      • Add the characters I suggested above. The states are there (it doesn’t matter what order the states go it, the polarity will be sorted by the outgroup). As for what will move, it does depend how many you add and how complete the taxa in your tree are. The first 6 are supportive of a monophyletc synapsidaso adding it may cause Caseids and the varanopids not currently in synapsids over to synapsids. The latter 5 might enough to put synapsids as the sister to sauropsids rather than within sauropsids. But if adding characters to 300 odd taxa is a pain, then for now just see what the scores are for a basal archosaur and a caseid.

  2. Of course, in saying this I run the risk of wasting more time, since you may well delete it, but here I go again. Your constant complaint that critics need to point to specifics rather than indulge in broad-brush nay-saying ignores the point I keep making in these comments: your magic-eye DGS technique is giving you so many erroneous results that it would take hours of detailed critique to go through all of them, and sending out brief critical comments like the one I’m writing now certainly takes less time than going through the features you are claiming to find, _and_ the features you are then using in character analyses. Since you simply reject the idea that ambiguous observations of vague lumps and hollows on matrix represent ambiguity and pareidolia (cf. Cosesaurus – hey, just because Ellenberger claimed it was there too doesn’t mean we have to trust it), let’s look at your recent article on Placodus. I haven’t seen Placodus inexpectatus, but I have looked at a lot of placodont material, especially the German stuff. Based on what I’ve seen – and what others have reported and described in the literature…

    — you’re showing the prefrontal as two bones, whereas in fact it seems to be, with sections both anterior and anterodorsal to the orbit. You’ve been misled by damage to the periosteum.
    — you’re showing the postfrontal descending down the posterior border of the orbit, whereas in fact it doesn’t do that – you’ve been misled by distortion.
    — you show the jugal as if it can be clearly demarcated from the maxilla, whereas breakage and cracking makes it impossible to work out what’s going on here. Based on other Placodus specimens, the anterior part of the jugal should peter out close to the prefrontal.
    — you show three different elements in the region of the squamosal and quadrate, whereas in fact smashed bone means that we can’t see the difference between, or the extent of, the squamosal and quadrate.
    — your illustration makes it looks as if there are three bones forming the lateral side of the coronoid process, whereas everything there is ambiguous – perhaps we’re seeing a coronoid and surangular posterior to it, but you can’t tell.
    — you also show a long, low, hyoid-like slip on the ventral surface of the jaw, but we also can’t make out what’s going on there..

    I could go on… the point is that all of your DGS-based images are like: a significant percentage of what you’re reporting is erroneous or ambiguous. This is wholly relevant to everything you’re doing, since the raw data you use cannot be trusted; your cladograms, and biomechanical comments and so on, are based on your DGS results! (You’re right about the scale bars though – what they report as 10 mm is indeed meant to be 100 mm).

    And, by the way, are you seriously comparing yourself to E. O. Wilson?

    In case this comment gets deleted, I’m posted it (and the previous one) on Tet Zoo. I think this stuff should be recorded online: again, because I’m concerned that you’re misleading naive readers. People should know that what you’re claiming to be ‘good’, ‘repeatable’, ‘evidence-based’ etc is actually wholly unsatisfying, subjective and erroneous.

    • re: Placodonts
      1. I corrected the antorbital portion of the prefrontal to better match my original tracing. Not sure why I didn’t do that earlier. Palatodonta and Paraplacodus both separate the lacrimal and prefrontal.

      2. Several placodonts have the postfrontal descending down the posterior orbit, so unchanged there.

      3. I think the bone break occurs at the jugal/maxilla border. If not we’ll figure this out when the next undamaged specimen appears.

      4. The smashed bone around the squamosal and quadrant is unclear. Still, enough appears to make a tentative assignment, which is okay in Science. You’ll note I changed the color of the posterior jugal drifted bone, which now makes more sense as an anterior qj.

      5. The posterior green (third) bone posterior to the coronoid process is the only visible port of the pterygoid crushed beneath the coronoid.

      6. The hyoids (although broken) seem clear to me. Other placodonts have them.

      Thanks for your help, Darren.

      • why didn’t you respond to the meat of Darren’s questions?
        He says your results are a problem because of your techniques.
        You responded with an answer based on your technique.

        E.O. Wilson was attacked by people with political differences, for the most part, not scientific. As far as I can tell, not a single critic of your work has a political agenda.


      • Paul, if my techniques were questionable to wrong, then my results would be not likely be as evolutionarily ordered as they are. If my observations were wrong, then, as I’ve said many times, it’s up to you readers to point out the errors so I can fix them. If the observational errors are not pointed out, then what can I do? You are welcome to reorder the reptile or pterosaur tree to the correct topology and we’ll compare results.

  3. Good job on specific stuff, Darren, that’s the ticket! I’ll take another look at your suggestions.

    When you say ambiguous, that’s understood. It can be ambiguous, even with the fossil in front of you. What I offer here is a suggestion based on related taxa. When you say erroneous, you might be right, or not. I’ll check it. In any case, the placodont nesting was confirmed, errors or not.

    DGS has its limits. But sometimes it offers solutions, especially to really difficult roadkills and pterosaur embryos. To your point, in some placodonts the lacrimal and prefrontal may be fused. In others they are not. Best to look at them individually to make that determination. I’ll take a closer look.

    Not comparing myself to EO Wilson, but I liked what he had to say. Didn’t you?

    And I’m glad you think my data cannot be trusted. That’s to encourage you to collect your own data on these specimens. Then we can come to consensus or argument. To your point, I often trust data, but sometimes I can’t (like the scale bar situation noted above) or I find that mistakes were made. I make them. So do others. It happens. I don’t see the need to call into question all of a colleagues observations, like you have done to me. Just take note of the problems and deal with them individually.

  4. Dave – I think you’re deliberately missing or ignoring the point. My specific comments above address one taxon (Placodus inexpectatus), and it’s clear from your answer(s) that you’re happy to accept that there is ambiguity in your proposals and identifications; even that things are provisional. My point – and it’s a major point that shouldn’t be downplayed or glossed over – is that this goes for every single one of the DGS-based reconstructions that you’ve produced: in other words, yes, in your case, there IS a need to “call into question all of a colleague’s observations”. Every time I look at one of your reconstructions, I think “that’s not right”, “that’s unknown”, “that’s ambiguous and probably not right”, “that’s not right either”, etc etc etc. I’m sure you understand that I don’t have time to submit similar comments about all of your DGS reconstructions (even the Placodus comments above don’t touch on all the things I think you’ve mis-represented).

    As I’ve said before, given that these observations are key to character choice and character coding, and thus to tree form, and thus to hypotheses, it’s all an untrustworthy house of cards. Sorry, but this is unavoidable and it partly explains why nobody follows or uses your reconstructions.

    Again, I’ll post this on Tet Zoo if you delete it. All the best.

    • Bottom line, for all my “mistakes” [and no one has validated your list yet] nesting of the new taxon was identical to the original paper.

      To your other point, untrustworthy is not only okay, that feeling you have is encouraged! As I said before, I’m offering alternatives. This only works if others find my alternatives valid (remember I did locate the pteroid on the proximal syncarpal for photos and drawings, confirmed by Kellner’s team and bipedal pterosaur tracks have been published).

      To the future: Let me know when someone tests Huehuecuetzpalli in pterosaur studies. Or Vancleavea in thalattosaur studies. Or Casea in millerettid studies. Or Stereosternum and Wumengosaurus in ichthyosaur studies. Then we’ll see.

      Key point: Your hedge, and I quote, “partly explains,” tells me a deeper part of you understands that the main reason others don’t test Huehuecuetzpalli, Diandongosuchus and Vancleavea as I’ve suggested is largely due to the fear of someone who has never actually seen the fossil being right. Most of what I’m proposing doesn’t come from DGS studies. You can toss all the ambiguous observations out the window and still come up with the same tree topology. I’m campaigning for more taxon inclusion. And you can’t lead a rally against that.

      The beauty of phylogenetic analysis is, you can be wrong on a few points, or missing large chunks of data, or you can overweight certain traits, and you’ll still come up with the same topology because the suite or weight of all the correct entries typically (not always) overwhelms the blanks and errors.

      Don’t be afraid to test the alternatives, just because you don’t like/don’t trust the author. Test the taxa, not me.

      Keep those specific comments coming and remember, you’re not always right either.

  5. Sorry for slow responses, cannot keep up. In all the discussion that I and others have had about reptile phylogeny and about you and your writings, the thing that’s constantly bemoaned is the lack of analyses that include sufficient taxa: yes, it’s widely recognised that more inclusive datasets are needed; we all agree on this and your complaints about this issue are valid and worthy.

    Alas, doing this sort of work right takes a huge amount of effort (and funding) and a huge amount of time. As I’ve made clear, basing observations on a unique digital interpretation of photos fails – erroneous results all round, so your work as goes character choice and coding can’t be trusted. Furthermore, a perusal of ALL the comments made here on your blog and elsewhere online demonstrate pretty convincingly that you aren’t compiling or running analyses that are at all trustworthy or reliable either. So, the idea that people reject your results and ignore your hypotheses is nothing to do with disliking you, it’s because we can smell the garbage. Again, remember that your stuff fails peer-review because it’s not good enough or is clear nonsense, NOT because there’s a vendetta against you or because you’re “blackballed” as you claim.

    For you to claim that people aren’t testing, say, Huehuecuetzpalli in the same analysis as pterosaurs because there’s a “fear of someone who has never actually seen the fossil being right” is insulting nonsense: sure, there are strong reasons for thinking that this result in particular is erroneous (as I think I’ve said here before, there are numerous reasons for regarding the idea of a squamate-pterosaur link as absurd, based on what we already know about character distribution), but the lack/rarity of these sorts of analyses is also due to the fact that doing this sort of work (and doing it right) is hugely time consuming and difficult. Look at the comments on the Tet Zoo mesosaur article – progress is happening, but it’s slow and difficult. As is clear, I don’t think that any of your major claimed discoveries are valid. By the way, there ARE (unpublished) trees out there that have aimed, empirically and in totally unbiased fashion, to test your hypotheses. They do not support them.

    • Why be cagey? Connect us tree builders together. BTW, Huehuecuetzpalli is not a squamate. Unfused ankles and a long list of other traits set it apart, and that’s according to Reynoso (1998). And what are the numerous reasons for the absurdity of Huehue and pteros? Being dodgy like this just shows you have no bullets in your pistol, Darren.

  6. “Shows I have no bullets in my pistol”… Well — I’m tired of repeating myself, I struggle with time (unlike you, it seems) and I’m pretty sure I listed the problem characters before (ok, maybe it was on Tet Zoo). Your characters for linking taxa like Huehuecuetzpalli and Pterosauria include attenuate tail and long premaxillae, whereas the characters that put Huehuecuetzpalli close to squamates (sorry, Reynoso was ambiguous, calling it a squamate in the title but having it as a ‘possible squamate’ in the text, and showing Squamata for the Huehuecuetzpalli + crown-squamate node in his cladogram) include fused, hourglass-shaped frontals, straight frontoparietal suture, streptostylic quadrate, absence of gastralia and so on. You’re no doubt going to come back with “aha, but those features are present in [supposed transitional taxon x] and [supposed transitional taxon y]”, yet with the supporting characters being based on magic-eye DGS, AND with character support that’s less compelling and less robust than that proffered by the studies (Reynoso, Gauthier et al.) that link Huehuecuetzpalli with squamates.

    As for connecting tree builders together.. I’m not going to share the details of unpublished work when I don’t have permission, I’m sure you can understand. And, at the risk of sounding mean, I’m pretty sure that your behaviour on this blog, refusal to give up on DGS etc. shows that you might be a little hard to work with. Indeed, I know from experience that it’s impossible to agree with you about anatomical observations, the basic currency we need to get sorted before we start playing with data sets. You see teeth and novel bones where there are cracks and smudges (viz, Istiodactylus, Germanodactylus, that unnamed Brazilian azhdarchoid, etc. etc.), extra bones where there are hollows and scratches (viz, Cosesaurus), and entire tails and limbs where there are indeterminate shards of nothingness (viz, Longisquama). And, lest we forget, dorsal frills and babies, babies, babies everywhere until recently…

    • No, I’m easy to work with. Just send evidence.

      Squamata = Scleroglossa + Iguania. Huehue falls outside that clade according to Reynoso, who unfortunately, did not include (standard refrain) other taxa from the third clade of lepidosaurs. Nor did he even realize (did not occur to him) that such taxa might be pertinent. This is the value of the large reptile tree. The taxa nest were they do most parsimoniously, minimizing the effect of personal choice.

      The key to winning your argument and making me a convert to your cause, Darren, is providing evidence (with photos, lines and arrows) that we can discuss. I encourage you to do this. All your emails without evidence, as you realized a long time ago, are essentially a waste of your time. I stand by results. It’s not fun sometimes. But I don’t see any alternatives that are more parsimonious.

  7. Well, I’m not trying to “convert” you – I know that that can’t be done: you’re unshakable in your view that your observations on anatomy are valid (whereas I remain amazed that you trust the ambiguous and erroneous results you produce on a daily basis: I just looked at the Dorygnathus articles and see more of the same) etc. Rather – as I’ve said many times now – I want naive readers to see that you’re being challenged: you’re so insanely productive, so good at swamping the internet with your stuff, that you’re leading a one-man campaign of misinformation.

    Squamata: yup, there’s a node-based definition, but .. as to whether your ‘tritosaurs’ group with them – it’s not looking this way based on the work of other researchers.

  8. Blackwashing you’re good at, Darren. Producing valid specific objections, not so much. I encourage you to have those “other researchers” contact me before publication to avoid the sort of embarrassment Benton and Hone (2007, 2008) and several others went through. I can spot red flags before publication and they can choose to include or exclude my suggestions.

    • “I encourage you to have those “other researchers” contact me before publication to avoid the sort of embarrassment Benton and Hone (2007, 2008) and several others went through. I can spot red flags before publication and they can choose to include or exclude my suggestions”
      What embarrassment? I don’t think you are the final arbiter of all things pterosaurian. Even if you were, why would they feel like they need to get your stamp of approval before submitting a manuscript?

      • Rob, I encourage you to look at the Hone and Benton (2007, 2009) papers then check out my response to them here:
        In short they were unable to nest Choristoderes within the Choristodera and unable to nest Squamata within the Lepidosauromorpha and Cosesaurus was the outgroup to Proterosuchidae and the Archosauria, among several other bungles.

        Later Bennett 2012 reported, “Similarly, their inability to replicate Peters’ (2000) analysis of a modified Bennett (1996) data set was the result of seven coding errors in retyping Peters’ published data matrix plus two 9s treated as a distinct character state.”

Leave a Reply

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

You are commenting using your 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.