What’s wrong with this paper?

Taking a new tack with the common descent/common design theme, I’d like to sneak up on it. As Sal Cordova likes to do, I offer a hypothesis for critique. Here’s an old phyogenetic analysis of mine. Does it show common descent? If not, what’s wrong with it?

Harshman J., Huddleston C.J., Bollback J., Parsons T.M., Braun M.J. True and false gharials: A nuclear gene phylogeny of Crocodylia. Systematic Biology 2003;  52:386-402.

This one is about crocodiles. Are crocodiles all the same kind? How do you know? If they aren’t how many kinds are there within the group and how do you know? If they are, are they a whole kind or just part of a larger kind? And if the latter, what is the larger kind?

216 Replies to “What’s wrong with this paper?”

  1. John Harshman John Harshman

    Another way to state my problem with you is that almost every comment you make is an attempt at a “gotcha”, but it generally involves you misinterpreting what someone else has said.

  2. J-Mac


    Don’t forget that feathers also evolved for sexual advances… you see evolution was working on making dinosaurs to fly but forgot to curb their appetites… since they got too fat, evolution changed its mind and made them use feathers to attract potential mates… you see how clever evolution is… which obviously evolution predicts…

  3. Mung Mung

    J-Mac: Don’t forget that feathers also evolved for sexual advances…

    Thanks for the reminder. But Glen didn’t actually say that, while he did say the other two I mentioned.

  4. Mung Mung

    John Harshman: Another way to state my problem with you is that almost every comment you make is an attempt at a “gotcha”, but it generally involves you misinterpreting what someone else has said.

    Well, yes, I engage in “gotcha” but so does your side. Perhaps it’s due to my own misunderstanding. There’s plenty of hypocrisy to go around for everyone. It’s not in short supply. 🙂

    And we do have a recent example of you doing the same thing. Remember when you “corrected” me about what it is that defines a clade? Later on Rumraket made essentially the same claim that I made but you declined to “correct” him.

    Maybe both of us can do better.

  5. J-Mac

    stcordova: Yes, even when I talk about it myself, it seems comical, as if foolishness of a child’s fairy tales….

    When I was researching Noah’s Ark, I came across an interesting fact: the length to width to height ratio is 30 to 5 to 3. I found that this ratio is commonly used in large ships that are designed not for speed. The ark was designed to float and not for speed…
    How could bible writer few thousand years ago know what the ideal ratio of a huge wooden box would be for floating if that was experimentally proven less than 100 years ago and still remains as an engineering ship building formula for big slow ships…

  6. Robert Byers

    Joe Felsenstein:
    I have not had time to carefully read the paper until now.It is certainly a good and careful job.The late 1990s saw many systematists and molecular evolutionists getting nuclear gene data to complement the large amount of previous work on mitochondrial gernes.This paper is a solid advance on that front, and does a particularly thorough job of comparing the evolution of molecular and morphological characters.

    The internal consistency of the signal from molecular sites is impressive — different sites provide independent signal supporting the molecular tree.Thus there is considerable signal of common descent.

    I note that the difference between the morphological and molecular trees seems to come from the position of the gharials.Based on the runs done, I suspect that the problem is caused by covariation in the evolution of morphological characters.The analyses of morphological characters done here uses discretely-coded characters that are treated as if they evolved independently.I suspect that the postcranial characters had substantial evolutionary covariances, owing either to genetic correlations or selective covariances (covarying selective pressures).The methods available for morphological characters do not take this into account, so they are likely to be overimpressed by apparently reinforcing signal.

    If the gharial is removed from the data set, which of course we do not want to do as its position is one of the major concerns of the study, the morphological and molecular trees are entirely consistent with each other.This suggests to me that selective covariance (selective correlation) on the line to the gharial Gavialus is the issue, with selection changing multiple characters simultaneously.Genetic correlation seems less likely given the diversity of characters involved.

    The paper doesn’t compare the morphological and molecular characters evolution.
    It compares the characters of both and presumes evolution has been going on.
    Where is your evolution evidence thats being compared??

    I read carefully your post but big word combinations don’t make a point that everything in this is simple grouping of traits. Atomic or bodypland!
    Thats it and no more.

    why they score as they do is another issue.
    The trees looking the same, not enough however to not overthrow a previous conclusion, would be this way from other options.
    Why not?

    Evolutionists show thoughtful creationists they are not doing scientific investigation of the origins of biology WHEN they start off with boundaries for options for same origins.
    COMMON DESIGN must be a option for likeness in traits/characters.

    By the way. If morphology had a atomic score/code it should match genetics.
    Changing the morph changes the genes and opposite.
    for example i’m certain marsupials are just adjusted placentals.
    yet they have genetic scores with marsupial scores being quite loud.
    So this means the marsupial genes is just thrown on top after the morphing/gene change.

  7. Rumraket Rumraket

    Mung: No, I don’t think so. Science doesn’t offer us proofs, after all.

    You misunderstand me, I was agreeing with Allan that there is no “definitive” case for common descent in the sense that definitive means proof.

  8. Allan Miller


    You are running away too? Why? You must have strong faith…

    You asked a question, I answered.

  9. Allan Miller


    I think that the issues with Theobald’s paper are publicly available.

    As is his own response thereto. Doesn’t really affect what I say though.

    In his response to criticism he discusses the difference between model selection and more traditional approaches relating to null hypotheses. In the latter, you either reject or fail to reject the null, you don’t confirm the null. Proving non-relationship would be confirming the null, IMO.

    In my response to you I was not thinking of Theobald, but of that kind of ‘null’ approach. I was thinking just in terms of species pairs. If a species (or simply a sequence) pair has no alignment beyond the expectation of two ‘random’ sequences, you haven’t proven their non-relationship, just failed to find any relationship.

    Theobald is a little more subtle: it’s about treelike topology – not just species pairs, but the concordance of the overall pattern with the expectation of common descent – which, bizarrely, you still seem to think has no expectation: But it does: trees.

    A statistic that favors common descent would mean that we know what to expect from common descent, but we don’t know what to expect from common descent. You cannot predict the outcomes of “common descent.” You cannot say which outcomes of “common descent” are more probable than others.

    Of course you can. On common descent, a treelike topology is more probable than a ‘random’ arrangement, because common descent means that species pairs tracing back to a common ancestor will have that ancestor’s genes, albeit somewhat scrambled by subsequent mutation. This holds through additional bifurcations, building up from single nodes to trees. So, if you find a treelike topology in a set of character states – better yet if you find the same broad tree in another set – you have pretty strong evidence that common descent lies at the heart of it.

  10. Corneel Corneel

    stcordova: The species specific “-omes” were created with species specific functions, but many of the new alleles that weren’t created heterozygously into the kinds but were added by mutation are probably function-compromising alleles. Sadly we see the effect of the function compromising alleles that are damaged -omes in various human heritable diseases.

    So we have

    1) new (deleterious) alleles introduced by mutation.


    2) radiation (branching evolution) of lineages from a few ancetral kinds

    And now for the million dollar question. Are you ready?
    How exactly will those mutations be distributed across the resulting lineages. Will that resemble a nested hierarchy perhaps?

  11. Corneel Corneel

    stcordova: But, I should have specified the “lower organisms” were definitely the microbes.

    So crocks are NOT lower organisms? What exactly do you mean with lower organisms? Are we talking scala naturae here?

    stcordova: There are paralogs of INFL3 in the human. One working paralog is NG_042193.1, the paralogous pseudo gene I’m not sure is annotated, but you can find the first exon starting at coordinate 45,136 and the end of the last exon at 46,919 in AC011445.6. A mutation in the Adenine at coordinate 45,633 in AC011445.6 converts the pre-mature stop codon (TAA) to a Tyrosine (TAC). Individuals who get this nice back mutation “allele” have better immune defense as this paralog interferon is active in them and they don’t catch colds and other viral infections as easily.
    That is an example of what I meant of how even we humans can tolerate some mutation, suffer, but not immediately not go extinct by these new disease associated alleles. That hopefully reconciles that apparent contradiction with random mutation creating new alleles, and yet the human (and other genomes) also having sophisticated proteomic “-omes”.

    Let me get this straight: Humans started out with a perfect polyconstrained genome (plus all the other -omes stacked on top). Still, we were able to tolerate numerous mutations that created, among other things, pseudogenes without suffering severely debilitating effects. Right? And then because we have acrued a number of deleterious mutations, backmutations create beneficial alleles by restoring the original function.

    Hence you concede that junk DNA exists (pseudogenes) and the possibility of beneficial alleles by random mutation. Did I understand that correctly?

  12. phoodoo

    Mung: Salvador will be pleased to hear that.

    I fully understand how little serious argumentation actually goes on around here and don’t try to pretend otherwise. Take for example your recent response to Erik:

    And why should I take seriously anyone who claims that feathers evolved for insulation and that feathers also evolved for flight? Do they sound like they are really putting forth a serious argument?

    One of the two uses was lucky, either the flying part or the insulation part, the other one was…. Was evolved for!

  13. John Harshman John Harshman

    So, to summarize: there’s nothing wrong with the paper, and it either is or is not evidence for common descent of crocodylians, depending on which creationist is talking and, apparently, the relative humidity at the time, because Jesus. I had hoped for more.

  14. J-Mac

    This one is about crocodiles. Are crocodiles all the same kind? How do you know? If they aren’t how many kinds are there within the group and how do you know? If they are, are they a whole kind or just part of a larger kind? And if the latter, what is the larger kind?
    The real beauty of Hirschman’s paper that it fits any imaginable scenario, including the option that crocks didn’t go to the Ark and just floated in the water… No matter what idea one has about crocks’ origins, harshman’s phylogenic analysis fits it…
    Who can argue with this piece of experimental science?

  15. RoyLT

    J-Mac: How could bible writer few thousand years ago know what the ideal ratio of a huge wooden box would be for floating if that was experimentally proven less than 100 years ago and still remains as an engineering ship building formula for big slow ships…

    Except that no wooden ship has ever been built to those dimensions. Johan’s Ark in the Netherlands (which is a full-scale replica) needs the support of 21 steel barges to keep it afloat. Even Huibers’ half-scale replica is barely seaworthy and cannot stand waves of any significant height.

    And length-to-width and length-to-draft ratios of transport ships vary widely across the industry depending upon application. I’m sure you can find a subset which adhere roughly to the ratio that you have stated, but that hardly constitutes a magic number.

  16. RoyLT

    stcordova: Jeanson has argued there is a genetic signature for domesticated animals that shows a bottle neck from a foundry population less than 10,000 years ago. If we see the same pattern in other animals, this would be consistent with Noah’s ark, or some comparable miracle.

    A bottleneck suggesting a foundry population for domesticated animals less than 10,000 years ago doesn’t sound particularly far-fetched to me. Archaeological evidence puts the Neolithic Revolution at the end of the Holocene around 12,500 years ago when agriculture and widespread animal domestication appear to have been invented.

    Good luck coming up with evidence for a similar bottleneck in wild animals though.

Leave a Reply

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