Let’s open up an old can of worms…

Three and half years ago
ReptileEvolution.com got a sound thrashing from Dr. Darren Naish writing form his Scientific American blog, Tetrapod Zoology. Today with 350 more taxa added and the tree still fully resolved, let’s add some ‘post-its’ to ten images of Darren’s blogpost (Figs. 1-10) to see how things stand today. Both adversaries are still out there on the Internet. Neither has buckled under.

This will be a long post
So the gist of it is:

  1. Naish paints me as an outsider, only part of the paleo background. True – for most workers
  2. Naish uses art and hypotheses that are not in my website. Some perjorative art he uses are from other artists. He had permission to use all my artwork, so bringing in these oddities was not necessary given his headline. 
  3. Naish spends many paragraphs talking about pre-ReptileEvolution errors that I made. One reason for starting a fresh website was to rid myself of old errors, but with this ‘history’ tainting all of my present work Naish provides no possibility for future redemption or honor.
  4. Naish claims that I see certain things in fossil photos that are not there in fossils. None of these are present in ReptileEvolution.com. If so, let me know so I can remove them.
  5. Naish claims that I see certain things in fossils that are not visible in fossil photos. So, what can one do? (see below)
  6. Naish blackwashes the entire ReptileEvolution website, all of its data, all of its images, then and forever in the future. He leaves no stone unturned.
  7. Naish supports traditional trees, even those that nest pterosaurs close to phytosaurs and erythrosuchids and those that refuse to include fenestrasaurs. These are problems that need to be resolved in academia.
  8. Naish wants me to add hundreds of more characters. As we’ve seen earlier, that doesn’t statistically help much after 200 traits and can be a never ending request.
  9. Several times Naish provides high praise before he buries the knife. I won’t do that.
  10. Both of us are unmoved regarding our stances.

Bottom line:
No one is perfect. No matrix is perfect. No interpretation is perfect. Even so, superior provisional hypotheses can be advanced by increasing the number of taxa in a taxon list. The more data, the less any single error affects the rest of the matrix and the more unbiased opportunities are present for nesting. If, in the end, all sister taxa look like sisters, with gradual accumulations of traits for all derived taxa, and the tree is completely resolved with high Bootstrap scores, isn’t that what we’re all looking for? Doesn’t that more closely model actual evolutionary events?

On the other hand, if Naish is correct
If my cladogram is built on (hundreds of) thousands of errors, even after fixing tens of thousands of errors, how is a fully resolved tree demonstrating gradual accumulations of derived traits for every taxon even possible?

When Naish is writing on his blog
and as I am writing here, we lose our scientist mantles and become journalists. In that role we have the obligation to name sources, be specific, inform the reader, write about what the headline promises, and try to maintain a balanced unprejudiced view. Any variation from this becomes propaganda. To that end, if you have any questions regarding topics that arise here, look up the answers on previous blog posts or on ReptileEvolution.com or drop me a note.

All of these pages can be enlarged with a click.
Some Carl Sagan quotes break up the long page images here and give you time to digest.

Figure 1 of 10. Click to enlarge. Monitor shot of Tet Zoo blog with annotations in yellow.

Figure 1 of 10. Click to enlarge. Monitor shot of Tet Zoo blog with annotations in yellow.

Below (Fig. 2) you’ll see some bizarre Pterodactylus art
that Naish says was illustrated as if my hypotheses were correct. To be fair, I show you a Pterodactylus from ReptileEvolution.com (Fig. 1a). Why did Naish choose to show the bizarre artwork of another artist instead of using one from the website he was critical of?

Pterodactylus scolopaciceps.

Figure 1a. Pterodactylus scolopaciceps, BSP 1937 I 18, No. 21 in the Wellnhofer 1970 catalog.


Figure 2 of 10 from Tetrapod Zoology.

Figure 2 of 10 from Tetrapod Zoology.

“For me, it is far better to grasp the Universe as it really is than to persist in delusion, however satisfying and reassuring.”
— Carl Sagan —

Figure 3 of 10 from Tetrapod Zoology.

Figure 3 of 10 from Tetrapod Zoology.


The truth may be puzzling.
It may take some work to grapple with.
It may be counterintuitive.
It may contradict deeply held prejudices.
It may not be consonant with what we desperately want to be true.
But our preferences do not determine what’s true.
— Carl Sagan —

Figure 4 of 10 from Tetrapod Zoology.

Figure 4 of 10 from Tetrapod Zoology.

Here’s one Pteranodon from ReptileEvolution.com. 
(Fig. 4b) Why didn’t Naish use this one by hand instead of the Pteranodon from Hell? Did he venture into propaganda? Did he want to make my work look unworthy of respect?

Figure 3. The UALVP specimen of Pteranodon. Note the lack of taper in the rostrum along with the small size of the orbit.

Figure 4b. The UALVP specimen of Pteranodon. Note the lack of taper in the rostrum along with the small size of the orbit.

Figure 5 of 10 from Tetrapod Zoology.

Figure 5 of 10 from Tetrapod Zoology.

The poor graduate student at his or her Ph.D. oral exam
is subjected to a withering crossfire of questions
that sometimes seem hostile or contemptuous;
this from the professors who have the candidate’s future in their grasp.
— Carl Sagan —

Figure 6 of 10 from Tetrapod Zoology.

Figure 6 of 10 from Tetrapod Zoology.

Every kid starts out as a natural-born scientist,
and then we beat it out of them.
A few trickle through the system with their wonder and enthusiasm for science intact.

— Carl Sagan —

Figure 7 of 10 from Tetrapod Zoology.

Figure 7 of 10 from Tetrapod Zoology.

I think people in power have a vested interest to oppose critical thinking.
— Carl Sagan —

Figure 8 of 10 from Tetrapod Zoology.

Figure 8 of 10 from Tetrapod Zoology.

It is undesirable to believe a proposition
when there is no ground whatever for supposing it true.
— Carl Sagan —

Figure 9 of 10 from Tetrapod Zoology.

Figure 9 of 10 from Tetrapod Zoology.

At the heart of science
is an essential balance between two seemingly contradictory attitudes –
an openness to new ideas, no matter how bizarre or counterintuitive,
and the most ruthlessly skeptical scrutiny of all ideas, old and new.
This is how deep truths are winnowed from deep nonsense.
— Carl Sagan —

Figure 10 of 10 from Tetrapod Zoology.

Figure 10 of 10 from Tetrapod Zoology.

In the end
is it good to have an an adversary/nemesis? Does conflict help advance Science? I think of it as necessary and invigorating. An adversary can be like a cleaner fish, helping one get rid of errors. I think it is also important to recognize value where present and to discuss specifics and details whenever possible.

On the subject of outsiders and insiders
JL Powell 2014 writes: “Luiz Alvarez was another resented outsider, resented by author and geologist Charles Officer. 

Outsiders do not know the mores for a given field and would be unlikely to uphold them if they did. Outsiders can resurrect a question that insiders have long stopped asking, 

Insiders may find it almost impossible to change their mind publicly, which may be equivealent to renouncing their life’s work. Even if insiders do not cast aspersions on the character and training of an individual outsider, they may still resent the implication that their discipline is incapable of solfing its won problems using its own methods.” 

Powell JL 2014. Four Revolutions in the Earth Sciences: From Heresy to Truth. Columbia University Press.



8 thoughts on “Let’s open up an old can of worms…

  1. “The last common ancestor should look like its proximal descendants”

    Only if we have a complete fossil record. We don’t. In fact for a lot of the clades in the terrestrial realm in the Permian and Triassic the fossil record is pretty dire, so you would expect seeminingly large jumps due to missing data. The fact that you have a seeingly beautiful set of transitions despite the fact that massive gaps exist seems to me extremely fishy; you have to assume that despite so much missing data you jus happen to have sampled the important transitional forms

    Also you keep harping on about how well resoved your tree is. This has no bearing on its accuracy, and if anything I would say the fact that such a large sample is so well resoved is a mark against it; there are situations where it is mathematically impossible for parsimony to resolve a clade (see the work by David Bapst entitled “When Can Clades Be Potentially Resolved with Morphology?”), so the fact that you are fully resolving your tree when you have such a largesample size should hint that something is wrong.

    But I’ve said these things to you more than once and been ignored

  2. We don’t have a complete fossil record — but it is complete enough to recover this tree. Unfortunately, Neil, you may be working with an outdated paradigm or mindset with regard to cladistic theory and its limitations. All I can tell you is: this works — it appears to accurately reflect evolutionary events. All of the sister taxa here gradually blend with one another, certainly better than in competing cladograms (where cladograms differ). Errors have been present in the data set and I have repaired them as best as I could. I’m sure more repairs will be forthcoming. At present, even convergent taxa are separated here. And yes, there is still room for additional transitional taxa to be added between every pair of leaf taxa on this tree.

    I invite you to tell us/me which taxa do not belong together, if you think there are errors here. They should be obvious if present. If you can find errors in the matrix, please do. We’ll test them. And you’re right… it doesn’t pay to gripe — if you don’t have a better (more parsimonious) solution to offer. I thank you for your interest and your comment.

    • Ok now you’ve irritated me. And not only irritated. You’ve insulted me.

      I’ll tell you why. I may not be particularly hot on the anatomy of these critters. I’m not an anatomist. But I have devoted my career to studying the quality of the fossil record of these animals. The completeness of their fossil record, and the influence of sampling bias on our interpretations of it, is literally what I have spent the last four years of my life looking at. Not only that, but I have devoted a considerable amount of effort into familiarising myself with phylogenetic methodology.

      And then someone like you, who may have more experience with the anatomy, but clearly has only the most basic understanding of quantitative palaeontology, tells me I’m working with an outdated paradigm? Seriously?

      You have ad hoc, with no evidence, stated that the fossil record is fine. I have four years of research containing actual quantitative analyses into specifically this issue telling me it’s not (you’re following me on researchgate, you can check it yourself).

      And did you even read the paper I cited? From 2013, so not what I’d say out of date, and highlighting in a clear manner the issue of when certain clades should not be resolved properly.

      And no I will not give you a better solution if you’re going to define better as being more parsimonious (and particularly if you mean “more parsimonious with my data set”). Parsimony is not the be all and end all, and if you had spent more than 5 and a half minutes actually researching phylogenetic methodology you would know that. And particularly if you ignore 90% of the anatomical characters.

      • I will attempt an illustrated reply to your concerns in today’s blogpost, Neil. To your points: the list of anatomical characters is endless…literally endless. You have set before me (adding 9x more characters) a task only Sisyphus would be worthy of. In counterpoint, I have had success with the 228 ramshackle traits employed here that have successfully lumped and split a larger list of taxa.

        I downloaded the paper you suggested, but it is over my head and purely theoretical. Sorry. I can’t absorb it. I appreciate that you’ve done quantitative analyses for several years.

        When I say the fossil record is ‘fine’, that does not resonate as ‘fine’ to you. All I’m saying is, the large reptile tree demonstrates that the fossil record is good enough to get what you’re getting with the large reptile tree, which resolves many former enigmas and is unprecedented in its scope using specimen-based and species-based taxa.

  3. If you don’t understand a paper on the considerations behind a method you are using, maybe it’s a sign you should listen to people who get the method better than you? The “I don’t understand it so I’ll ignore it” doesn’t work for scientific methods any better than it works for my German phone bill.

    Just forget the idea of a minimum number of characters needed to fully resolve the tree. Forget any rules on how many you need.

    The reason you should want to add more character is simply to add more information. The more morphological information you have the more reliable the tree. End of story. You shouldn’t want to have the number of characters needed to resolve the tree, you should want the maximum amount of morphological information.

    The thing, I think, that riles people, is when you come up with a different result to them and assume it’s because you added more taxa. How do you know it’s not because you missed key characters? So many times people have said that you haven’t fully tested their ideas because you didn’t include the characters that defined the clade they are proposing. No one asks you to add every possible morphological detail in existence, but they do expect more than the barest minimum effort if you’re going to accuse them of being wrong.

    I remember a while ago I asked you to test the effect of 5 characters. Not 500, not 50, five. Barest minimum effort. They were characters which others had used to define synapsids, a group you found to be polyphyletic. You refused. That tells me you’re not interested in producing the best possible tree, you’re not interested in actually trying to work out reptile relationships, you’re only interested in telling people they are wrong.

    Oh, and my argument on the completeness of the fossil record. I wasn’t saying it was too incomplete to do a phylogenetic analysis, I was saying it was too incomplete for you to use he argument “in Joe blog’s tree the sisters don’t look like each other”. We don’t expect them to because the fossil record is incomplete

  4. When you say “add 5 traits” that sounds easy, Neil, but in reality it means checking each trait for every one of the 360 at that time to 642 taxa today. So that adds up to 1800 to 3210 scores presently. And if I do it for you, then I do it for everyone else who is unsatisfied by my character count, plus yours. It may never end.

    More to the point, I would have to add another character IF it were needed to lump and split some otherwise unlumpable / unsplittable pair. That has not happened yet.

    Even more to the point, I’ve seen matrices with 1500 characters and all too often the scoring doesn’t lump and split traits, as it does for me, it sometimes sprinkles traits about in singletons that are not parsimony informative. And these giant studies still result in clades without resolution. Obviously I’m pulling at low hanging fruit here. Some traits are helpful.

    Actually I do expect my sister taxa to look alike… enough. That’s why my basal therapsid looks very more like Haptodus/Ophiacodon/ Archaeothyris, rather than Dimetrodon, etc. etc. which everyone else prefers, overlooking the autopomorphic dorsal sail.

    With regard to understanding a paper… I don’t have to understand the ‘why’ or the theory when a computer does the work without prejudice what I could never do by eyeballing relationships, like they used to do prior to 1985 or so with those hand drawn bubble trees. I can’t tell you how many times I’ve had the revelation, when the software spits out a new relationship, that appears so obvious in hindsight.

    With regard to key characters – some, like a synapsid (lateral temporal) opening – are shared by unrelated taxa, like mesosaurs, thalattosaurs, caseasaurs, bolosaurs, lanthanosuchians and the related protodiapsids. No one knew that – tp that extent – before the large reptile tree. In other words, key characters for you may not be key characters after all. Convergence is rampant. I trust unbiased software has the ability to split diapsids into two unrelated convergent clades, and all the other lumping and splitting it has successfully done so far.

    I hope this helps. I’m going with what works.

    • You genuinely don’t see the irony in you criticising other workers for not adding the taxa you suggest, yet you refusing to add the characters they suggest?

      The impression I got from your first paragraph is that your only reason for not adding more characters boils down to “meh that sounds like effort.” Yes, it will take a while to score a new character for more than 600 taxa. That’s just tough. You chose to make a huge tree, you have to deal with the more tedious aspects that this entails.

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s