Tetrapod Zoology

A substantial amount of evidence demonstrates that birds are theropod dinosaurs, and that birds evolved during the Jurassic from small, feathered maniraptoran theropods closely related to dromaeosaurids and troodontids (known collectively as deinonychosaurs) [the small dromaeosaurid Microraptor shown in adjacent image]. The precise details of avian origins remain the subject of debate, and argument continues over whether proto-birds were tree climbers, leapers, terrestrial cursors, or combined some of these lifestyles. But because basal deinonychosaurs were small animals (not much different in size from basal birds like Archaeopteryx), it seems that small size was a primitive character of the deinonychosaur + bird clade (known as Paraves). We also know – thanks to the spectacular Lower Cretaceous fossils from Liaoning Province in China – that deinonychosaurs (and other maniraptorans) were feathered like birds. In other words, small size and an extensive covering of true feathers were primitive for paravians.

Based on what we know of the dinosaur fossil record, palaeontologists generally think that most of dinosaur evolution happened at relatively large body size (by which I mean at a mass greater than a few kilos), and in animals that were entirely terrestrial. Small size (<1 kg) and arboreal habits seem to have arisen pretty late, and only within maniraptorans. I’m not saying, by the way, that all maniraptorans were arboreal; just that it’s somewhere within this clade that arboreal habits evolved.

Here, we look at a rather different proposal: the decidedly non-standard, non-mainstream Birds Come First (or BCF) hypothesis proposed by George Olshevsky. Rightly or wrongly, BCF has never been discussed in the technical literature (I have at least alluded to it in historiographical articles (Naish 2000a, b)), and all of George’s articles on it have been in the ‘grey’ or popular literature (Olshevsky 1991, 1994, 2001a, b). Thanks, predominantly, to his activity on the dinosaur mailing list (a popular discussion list for dinosaur aficionados and researchers), George’s BCF hypothesis was once well known and much discussed, and perhaps considered seriously by at least a few specialists [image below shows some of the 'dino-birds' integral to the BCF hypothesis (read on), with Archaeopteryx at far right. This illustration is © Luis Rey and originally appeared in Olshevsky (1994)]

It should be noted to start with that the name ‘Birds Come First’ is in fact rather misleading, as the hypothesis does not propose that birds in the proper sense evolved earlier than did other dinosaurs or other archosaurs: rather, it posits that small, bird-like, arboreal archosaurs were the direct ancestors of all the archosaurs that came later on (proper birds included). George was aware of this fact, and apparently considered the rather tongue-in-cheek alternative acronym GOODD, meaning George Olshevsky On Dinosaur Descendants. This was, of course, meant to be used in opposition to the also tongue-in-cheek BADD (Birds Are Dinosaur Descendants): the term George uses for the ‘conventional’ or ‘mainstream’ view of avian origins outlined in the first two paragraphs above. ‘BADD’ is bad, according to BCF, as it imagines that small size, feathers and arboreal habits all evolved very late in archosaur evolution, and exclusively within maniraptoran theropod dinosaurs [image below shows more of the 'dino-birds' integral to the BCF hypothesis. These are early forms that are imagined to be somewhere between the first archosaur and the first theropod in phylogenetic position. This illustration is © Luis Rey and originally appeared in Olshevsky (1994)].

The three ‘problems’ of the standard theory

Why does BCF have such a problem with the widely accepted, ‘standard’ view of bird evolution? Olshevsky points specifically to three problems: the ‘time problem’, the ‘size problem’, and the ‘wing problem’ (Olshevsky 1994, 2001a). The ‘time problem’ refers to the fact that basal birds (specifically, archaeopterygids) are older than other paravians. Bird-like maniraptorans (deinonychosaurs and so on) are younger than basal birds, not older than them as they should be if birds evolved from deinonychosaur-like ancestors. The fossil record indicates, Olshevsky argues, that the large, flightless maniraptorans (dromaeosaurids and so on) are more likely to be the descendants of the little, flying, bird-like forms, not the other way round. If you’ve followed the dinosaur literature you’ll know that this idea did not originate with Olshevsky: Greg Paul first proposed during the 1980s that dromaeosaurids and other bird-like coelurosaurs might be the secondarily flightless descendants of archaeopterygid-like ancestors (Paul 1984, 1988a, b, 2002).

The second problem proposed by Olshevsky – the ‘size problem’ – asserts that the apparent miniaturisation of proto-birds required for the standard theory of avian origins is problematical given that animals tend to become larger during their evolution, not smaller (the infamous Cope’s Rule). In the standard theory, no obvious explanation for the miniaturisation of the theropod dinosaurs ancestral to birds exists (Olshevsky 1994, 2001a). BCF gets round this (supposedly) by arguing that bird ancestors were small all along (going all the way back to the archosaurian common ancestor), and that the big archosaurs (ALL of them) evolved (on many, many separate occasions) from these small ancestors. In other words, BCF is compatible with Cope’s Rule [the dromaeosaurid Deinonychus shown here, from wikipedia].

Finally, BCF points to the ‘wing problem’. The standard theory requires that the avian wing evolved gradually from a limb originally used for predation. Olshevsky (2001a) argued that no convincing model ‘explaining’ this transition had ever been presented (he noted that various different hypotheses have been suggested: that wings evolved as exaptations from feathered arms originally used as insects traps, for shading eggs, for providing thrust, for aiding manoeuvrability when running, and so on). BCF therefore follows the same logic as that applied by those ornithologists who favour a non-dinosaurian origin for birds. Namely, that wings ‘cannot’ have evolved in a terrestrial context, but must instead have evolved in an arboreal setting, in animals that used their forelimbs to slow or control their movement during leaping, parachuting, or gliding.

Dinosaur phylogeny a la BCF

In order to provide a solution that takes account of these ‘three problems’, BCF kind of combines the ‘trees down’ model of avian origins favoured by some ornithologists (and by most palaeontologists in the post-Heilmann, pre-Ostrom era*) with the secondary flightless hypothesis proposed by Greg Paul. BCF proposes that dinosaurs and other archosaurs started their history as small, quadrupedal tree-climbers, already kitted out with spiny ‘proto-feathers’, and that a ‘central lineage’ of small, tree-climbing forms connected these Permo-Triassic forms with the birds proper that evolved later on. All of the key innovations that led to the modern avian condition – skeletal pneumaticity, endothermy, feathers, wings, reduction and loss of the 4th and 5th fingers, reduction of lateral toes and modification of the hallux, reduction and loss of teeth, tail stiffening and so on – evolved in this single unbroken lineage, the hypothetical members of which were referred to by Olshevsky as ‘dino-birds’. Dino-birds perhaps resembled the numerous small, arboreal theropods – the arbrosaurs – invented for Dougal Dixon’s fictional work The New Dinosaurs (a few arbrosaurs shown here). Whereas the conventional theory provides no obvious explanation for the incremental appearance of ‘avian’ characters in theropods and other archosaurs, BCF posits that they all evolved as logical improvements to the arboreal, gliding dino-bird lifestyle.

* If you’ve ever read anything about the history of bird origin theories, you’ll know that birds were originally linked with other dinosaurs back in the late 1800s, most famously by Thomas Huxley. This view remained fairly popular until the 1920s when Gerhard Heilmann’s book The Origin of Birds was published in English. Heilmann argued that birds could not have descended from dinosaurs (predominantly because dinosaurs lacked clavicles, or so he thought), and he therefore favoured the idea that birds originated from the so-called ‘pseudosuchians’: primitive archosaurs that were also thought ancestral to dinosaurs and crocodilians. This became the mainstream view until the 1970s, when a new look at the anatomical evidence (combined with new data from maniraptoran theropods) led John Ostrom to successfully resurrect the dinosaur hypothesis. The best and most complete review of bird origin theories is that provided by Witmer (1991).

Olshevsky has noted on several occasions that his model could be entirely compatible with any given ‘standard’ archosaur phylogeny. That his own, favoured phylogeny for archosaurs is decidedly non-standard is, therefore, effectively irrelevant, but it’s worth discussing it here as it hasn’t been much discussed elsewhere (I base my comments here on the details given in Olshevsky (1991, 1994), though I am aware that some of his views on the affinities discussed in those works have since changed). Olshevsky’s views on non-dinosaurian archosaurs aren’t really that weird: he imagines crurotarsans to form a clade that diverged early on from the pterosaur-dinosaur clade, and the sorts of relationships that he’s proposed for phytosaurs, rauisuchians, aetosaurs, crocodilians are so on aren’t that different from standard phylogenies (Olshevsky 1991, 1994). However, among the most controversial of his proposals is that the weird little Triassic reptiles Megalancosaurus and Longisquama [shown here] are dinosaurs, and part of a large group (termed Theropodomorpha) that also includes Marasuchus, Lagerpeton, herrerasaurids and theropods (Olshevsky 1991, 1994). Saurischia is not monophyletic in Olshevsky’s scheme, and sauropodomorphs and ornithischians are united in Phytodinosauria (an association originally proposed by Bakker, and later mooted by a few workers, but otherwise no longer maintained by anyone). Within Sauropodomorpha, sauropods are regarded as the most basal clade as, in contrast to other taxa, they possess large fifth toes (Olshevsky 2001c).

Remember that, in BCF, all of the groups of large, terrestrial archosaurs have evolved separately from arboreal dino-birds: in the crurotarsans of the Triassic, in the separate waves of big theropods, in the sauropodomorphs, and in the separate ornithischian clades, we are seeing successive and independent archosaurian radiations, all of which came down from the trees and took anew to terrestrial life.

So there you have it. It would be wrong to finish here: what about a critique of some sort? I’ll get to that next.

For previous articles on non-standard phylogenetic hypotheses see…

And for previous articles relevant to early birds and avian origins see…

Refs – -

Naish, D. 2000. Theropod dinosaurs in the trees: a historical review of arboreal habits amongst nonavian theropods. Archaeopteryx 18, 35-41.

- . 2000. 130 years of tree-climbing dinosaurs: Archaeopteryx, ‘arbrosaurs’ and the origin of avian flight. The Quarterly Journal of the Dinosaur Society 4 (1), 20-23.

Olshevsky, G. 1991. A Revision of the Parainfraclass Archosauria Cope, 1869, Excluding the Advanced Crocodylia. Publications Requiring Research, San Diego.

- . 1994. The birds first? A theory to fit the facts. Omni 16 (9), 34-86.

- . 2001a. The birds came first: a scenario for avian origins and early evolution, 1. Dino Press 4, 109-117.

- . 2001c. Dinosaurs 2001. Column 3: Isanosaurus. Dino Press 4, 92-95.

- . 2001b. The birds came first: a scenario for avian origins and early evolution. Dino Press 5, 106-112.

Paul, G. S. 1984. The archosaurs: a phylogenetic study. In Reif, W.-E. & Westphal, F. (eds) Third Symposium on Mesozoic Terrestrial Ecosystems, Short Papers. Attempto Verlag (Tübingen), pp. 175-180.

- . 1988a. The small predatory dinosaurs of the mid-Mesozoic: the horned theropods of the Morrison and great Oolite – Ornitholestes and Proceratosaurus – and the sickle-claw theropods of the Cloverly, Djadokhta and JudithRiver – Deinonychus, Velociraptor and Saurornitholestes. Hunteria 2 (4), 1-9.

- . 1988b. Predatory Dinosaurs of the World. Simon & Schuster, New York.

- . 2002. Dinosaurs of the Air: the Evolution and Loss of Flight in Dinosaurs and Birds. Baltimore: Johns Hopkins University Press, Baltimore.

Witmer, L. M. 1991. Perspectives on avian origins. In Schultze, H.-P. & Trueb, L. (eds) Origins of the Higher Groups of Tetrapods: Controversy and Consensus. Cornel University Press (Ithaca, London), pp. 427-466.

Comments

#1 David Marjanović
June 8, 2009

“Category: speculative zoology”… is that just because of the Dixon reference, or…? B-)

I’d like to mention that the BADD guys, at least on the DML, have taken that label and run with it, soon calling themselves BAD (Birds Are Dinosaurs!). It has the neat side effect of fitting with BAND.
#2 Andrea Cau
June 8, 2009

Good post!

If Olshevsky called his hypothesis SAACF! (Small Arboreal Archosauromorphs Came First!), probably none could find it so strange… or interesting…
But once you add the vernacular term “birds” in your reptilian phylogeny, then… storm clouds fill the sky with darkness, rain came and Earth shakes: it’s the Phylogenetic Doomsday!

I’m constraining my large-scale phylogeny of dinosauromorphs in order to test BCF…
#3 David Marjanović
June 8, 2009

If Olshevsky called his hypothesis SAACF! (Small Arboreal Archosauromorphs Came First!), probably none could find it so strange… or interesting…

Well, it’s still highly unparsimonious. And the size problem doesn’t even exist, and the time problem (which was always shaky because it assumed a strangely complete fossil record) has evaporated thanks to new discoveries.
#4 J-Dog
June 8, 2009

Thank you for doing all that work and putting things all together. Great links too!
#5 Jerzy
June 8, 2009

Well, some of premises of birds-first hypothesis are poorly defined or non-testable or not argued upon in classical hypothesis. For example, arboreality is quite hard to prove/disprove. Or, how much flight should be developed, and how much of grasping ability should be lost?

More interesting is history of maniraptorian arms/wings. How wings could devolve back into grasping predatory arms? How predators which partially lost grasping power could outcompete ones with full range of movements in forearms? Why deinonychosaurids had stangely limited arm movements? Why a heck Velociraptor needed huge quill-knobs?
#6 Pete Buchholz
June 8, 2009

George has mixed up two ideas into BCF. One, IMHO, is entirely reasonable: bird flight evolved from the trees down. The other, however, is not reasonable: there is a central trunk to ornithodiran evolution consisting of small arboreal animals (that he decided to call birds for some reason!) that are never ever ever fossilized (except for a few obvious non-dinosaurs), but was always leading to an obvious end result: birds.

If he left out this central trunk business and focused on trees down people would take this more seriously…
#7 John Jackson
June 8, 2009

I’ll register a comment high up in the list this time.

Readers coming to this posting on “BCF” will be under the impression that Naish’s view is well rooted in sound science, but the following should be borne in mind:

Whereas palaeontology these days involves the analysis of statistical algorithms for trying to work out family trees (phylogenies), most paleontologists have no qualifications in any remotely related area. It is easy to misuse or misunderstand these issues, even for those with expertise, but most palaeontologists simply take the results of a simplistic computer program at face value, despite being told repeatedly, by world authorities, that this is wrong. (They then deny that they’re doing it.) That makes as much sense as treating newspaper horoscopes as gospel truth. Most people would think that interpreting complex statistical evidence with nothing but a geology degree would be rather pointless, but in fact paleontologists keep trying this on, ignoring proper experts, and hoping to get away with it. Many people have learned from recent experience that communities, indeed whole industries, can be built up over many years, by led by people without any qualifications, which are eventually shown to be seriously flawed. With banking, the truth took longer to emerge than it would have done in medicine or engineering. Palaeontology is safer though since demonstration of error phylogenies is practically impossible.

No argument supporting BCF is ever listened to or taken account of in any way by the likes of Naish, forcing the process of argument to be restarted from the beginning every time. (Partly for this reason, arguments with cladists just go round forever in circles, as if you were arguing with some moron from a Monty Python sketch. Another reason is that they don’t follow modern scientific principles.) Olshevsky is not the main protagonist now, I am. You’ll notice my detailed views are absent from this blog posting, as is my name. You may also know that my contributions to blogs on this subject are most definitely not absent, whereas Olshevsky hasn’t said much for years. My arguments are, by now, much more detailed than Olshevsky’s. One attempt to justify ignoring me is that I have not published any of my views in recognised journals. Of course I never trusted the integrity of those in a position to block my publications – until I had an early script of my forthcoming book “The Secret Dinobird Story” and gave it a try. Now I have even more disdain for the integrity of palaeontologists – but also a small but valued collection of peer-review rejection drivel to demonstrate to the world. (In case anyone is wondering what kind of scientist I am myself, I have a masters degree in info. sys. eng., I’ve devised a form of artificial mind that actually works, and I teach brain science at degree level.) No, Naish is not the one exception who actually listens to the arguments from the BCF camp. I asked him to detail exactly why he insists on taking cladograms at face value and he refused to answer. He denied that he did swallow cladograms whole, but if you investigate his opening statement here: “A substantial amount of evidence demonstrates that birds are theropod dinosaurs”, you will find not only a misunderstanding of the nature of evidence, but a slavish reliance on cladograms. However, he suggested he did use non-cladogram evidence to justify his views – something that involved oviraptors I believe – but we’re still waiting for details on that, even though I have reminded him about it.

Finally, be sure to recognise Naish’s gratuitous use of inverted commas, as in: …the ‘size problem’, …the ‘wing problem’, …that wings ‘cannot’ have evolved in a terrestrial context, etc, and of weasel words as in …decidedly non-standard, non-mainstream… . How ingenious it is to describe one side of an issues that hasn’t been settled as Controversial. Well, if you’re not using science, you’ve got to use something.
#8 john Jackson
June 8, 2009

5
Well, some of premises of birds-first hypothesis are poorly defined or non-testable or not argued upon in classical hypothesis. For example, arboreality is quite hard to prove/disprove.

It’s fascinating how people with no genuine appreciation of the problems, but also the capabilities, of philosophy of science in a historical science, like to suggest they have some special insight into it that their opponents haven’t.

There is nothing in BCF (my form at least) that is any less testable than in competing theories. “Testability” is often claimed for cladogram-based theories, but this relies on cladograms testing something, which they don’t. And why is suggesting arboreality intrinsically harder to prove than suggesting the opposite? Don’t start accusing theories on mesozoic issues of being poorly defined. A statement – ie theory – can be as poorly defined as it likes.

Much of the pseudo-philosopical defence paleontologists use are misunderstood principles borrowed from physical sciences, which don’t apply in the same way to historical sciences anyway.
#9 Kevin Schreck
June 8, 2009

Good, thoughtful post. Looking forward to the next entry!
#10 Tygo Raxx, III
June 8, 2009

OMG – AMAZING. I mean, you really do know how to write stuff that reels in the lunatic fringe. I am referring to the wonderful Mr Jackson.
#11 Andrea Cau
June 8, 2009

Dear john Jackson, the BCF hypothesis is a palaeontological hypothesis, not a “info. sys. eng.” model.
I don’t care what your degree is: I want to know what arguments you use againts Naish’s statements.
Is your critique merely rethorical or based on palaeontological evidences?
Before publishing your “The Secret Dinobird Story”, read this.

PS: Do you know Peter Mihalda?
#12 Jerzy
June 8, 2009

@john Jackson
Arboreality is, of, course, possible to establish. However, it is much more difficult to prove behaviors (arboreality, flight) than paleontological characters (especially, if only parts of skeleton are preserved).
#13 Naraoia
June 8, 2009

I only wonder how much this statistics expert and paragon of scientific integrity knows about archosaurs.

Can’t wait to see how this discussion plays out. *gets popcorn*
#14 Aleck Zhou
June 8, 2009

It seems clear the BCF theory aims to be more controversial than convincing.

Talk about a time problem, ceratosauria and sauropodomorpha are suppose to be descendents of “birds” which don’t appear in the fossil record till the Jurassic? I guess he means ornithodirs? Why are they birds again? Clearly most were small but arboreal? Only pterosaurs seem to fit the bill.

Besides Cope’s law (why’s it a law if there’s so many obvious exceptions?) is there any justification for picking random features and declaring it the ancestral condition? Why not ceratopsians came first or sauropods came first? If you don’t respect parsimony then it’s clearly possible to envision a series of evolutionary steps from any suite of characteristics to any other.

John Jackson,

I think dinosaurs evolved from humans! Prove me wrong! All theories, including yours, are accepted by pseudo-science.
#15 Michael Erickson
June 8, 2009

It’s here! It’s Here! Now I’ll read it…
#16 Michael Erickson
June 8, 2009

Excellent post! Very, very interesting. Can’t wait for the critique!
#17 David Marjanović
June 8, 2009

For example, arboreality is quite hard to prove/disprove.

I’d say it’s pretty easy to disprove: when adaptations to climbing are absent, and especially when features are present that hinder climbing, we’re not looking at an arboreal animal. At least that’s the most parsimonious option.

Whereas palaeontology these days involves the analysis of statistical algorithms for trying to work out family trees (phylogenies), most paleontologists have no qualifications in any remotely related area. It is easy to misuse or misunderstand these issues, even for those with expertise, but most palaeontologists simply take the results of a simplistic computer program at face value, despite being told repeatedly, by world authorities, that this is wrong. (They then deny that they’re doing it.) That makes as much sense as treating newspaper horoscopes as gospel truth. Most people would think that interpreting complex statistical evidence with nothing but a geology degree would be rather pointless, but in fact paleontologists keep trying this on, ignoring proper experts, and hoping to get away with it.

Look, dude, you have fallen among the scientists. Take a published cladogram and show which misunderstanding of statistics it’s based on; or, alternatively, shut up.

“I fart in your general direction” is not an argument! Make specific (and thus falsifiable) arguments, or go home.

Partly for this reason, arguments with cladists just go round forever in circles, as if you were arguing with some moron from a Monty Python sketch.

Hmmmmmmmmm.

You know…

Another reason is that they don’t follow modern scientific principles.

Another case of “put up or shut up”.

my forthcoming book “The Secret Dinobird Story”

“Secret”? “Story”? If it’s supposed to be science, why do you dress it up as literature?

Also, in case you wonder, o engineer, I haven’t got a geology degree. I’ve got a biology degree. Palaeobiology, you see.
#18 Protest Fish
June 8, 2009

Thank you, Dave. I was wondering when someone would point out to Mr. Jackson that most palaentologists have a grounding in biology, seeing as how palaeontology happens to be a life science and not whatever the hell he seems to think it is.
#19 Owlmirror
June 8, 2009

I’ve devised a form of artificial mind that actually works

What do you call an “artificial mind”, and what do you mean by “actually works”?

Can you get your artificial mind to post here?

and I teach brain science at degree level.

Where do you teach, and why does this institution call it “brain science” instead of “neurology”?

I ask only for information.
#20 Laura Tamara Henson
June 8, 2009

I have read Tet-zoo for the last two years and it is definatly my favorate science site. I always thought Paul was onto something in his suspicion that Deinonychus was a decendant of Archaeopteryxian birds. I guess the pproblem is one of definition as Aves was always Archie + Modern birds. Personally I wouldent mind dromeosaurs being included in Aves, maybe it would finally stabalize dinosaur systamatics. I hate Gauther’s suggestion to restrict aves to the crown group, I still remember his Ornithosuchia “shudder”. Poor Onithosuchus, it no longer has an ordial name because Gauther likes redifining perfectly good names .

Olshevsky’s main problem was in insisting that the Avecephalians were early theropods. The Avecephalians truly were a weird group of tetrapods, especially Coelurosauravis. You should blog on them some time Darren.

PS. If you need some more stuff for a sequel for your april 1st dragon article you can use my paper on the Classification and Evolution of the Dragon on http://www.scribd.com/doc/16225318/The-Evolution-Classification-of-the-Dragon
#21 David Marjanović
June 8, 2009

Can you get your artificial mind to post here?

LOL!!!
#22 David Marjanović
June 8, 2009

To be fair, there are palaeontologists who have mostly a geology background. When you want to study palaeontology in Vienna, you have the choice between starting from biology or from geology; there are places, or at least have been, where geology is the only option. However, if you start from one, you have to take several courses in the other; and if you start from geology nowadays, you’re probably planning to become a biostratigrapher for petroleum geology, rather than a phylogeneticist — a palaeontologist as opposed to a palaeobiologist (as Othenio Abel meant these terms when he invented the latter).
#23 Nathan Myers
June 8, 2009

John, you’re bitter, maybe justifiably. Still, those of us here have done you no wrong, so it would be better to place your arguments and leave the bitterness out. (Its primary effect is to connect you, textually, to real crackpots, which I am now obliged to choose to assume you aren’t.) Darren didn’t have to write about your subject at all. If he uses scare quotes, it’s his blog, and it’s your job to show that the problems he doubts are real.

We’re very familiar with the phenomenon of echo-chamber fields, having seen them envelop much of particle physics, astronomy, economics, psychology, neurophysiology, and U.S. foreign relations in the present or in living memory. Paleontology is not immune to it; BAND is a holdover of such an incident. Still, paleontologists I know are rather more interested in evidence that contradicts their preconceptions than, say, present astronomers.

If you have evidence, let’s see it. Statistical evidence is hard for most people to evaluate. (See http://www.zedshaw.com/essays/programmer_stats.html for a rant in a different field; the money quote is “I’ve been studying it for years and years and still don’t think I know anything.”) It would help to get somebody (else) who really does understand statistics to vouch for your results.

On a side note, I’m curious to learn what sort of problems this artificial mind can solve. Maybe there’s a web page where I can set it problems to see how well it does, and how it learns? (I just hope to The FSM that it owes nothing to Mentifex.)
#24 Darren Naish
June 8, 2009

A brief response to John Jackson…

– The constant complaint that palaeontologists as a whole do not understand the mechanics/algorithms of cladistic methodology, and hence cannot and should not use it, is nonsense. Firstly, minor point: cladistics is not unique to palaeontology; it’s widely applied throughout the biological sciences and was invented by an entomologist. You should read the literature on extant organisms some time. Secondly, I don’t actually see why you think you need to understand the statistical algorithms and so on to be a perfectly good user of parsimony software. Please explain why you think this necessary. As someone who actually uses PAUP (also NONA, sometimes) and has generated cladograms, I think it’s far more important that you sort out character polarity, accurate coding etc. Thirdly, you are blissfully na�ve if you really, honestly think that biologists/palaeontologists truly don’t understand this stuff. Maybe you don’t know that some cladists actually write their own software, or that some have spent the better part of their careers writing about the data handling and computation involved in parsimony software.

– Another of your complaints – that the parsimony software we so doggedly adhere to will one day tumble like a house of cards and that, ho ho ho, won’t we all look oh so silly when that day comes – is also rubbish. It is universally understood that the cladograms generated by PAUP etc. are hypotheses liable to further testing. Are you seriously saying that we shouldn’t be proposing hypotheses using whatever means we have available?

– I do agree with you that BCF is essentially testable, but I do not see how this testing supports it. BCF makes direct claims about ancestral conditions at various points on the cladogram. We can test these claims by looking at the fossil evidence we have, and those fossils falsify the predictions of BCF. You’re gonna love my next article, as I discuss this matter in some depth therein.

– Finally, please remember that your perpetually condascending tone is really irritating and insulting. As per last time, I decided not to delete your comments because I think you should be allowed to have your say, but I am under no obligation to do this. As you will happily admit I’m sure, you are an amateur with no relevant qualifications. That doesn’t mean that you can’t sometimes be right, but it does mean that you have no good reason for your constant holier-than-thou, ‘we palaeontologists are all a bunch of idiots’ rhetoric. There is already one Tet Zoo regular who frequently has his angry, archosaur-themed rants deleted without hesitation, don’t make it two.
#25 DDeden
June 8, 2009

Dromaeosauridae: note typical long prehensile tail for gripping branch and “hyperextended” perching claw locked in tension, it leapt and “flyde” (guided glide + fly as controlled fall) down to stomp-tackle and bite little critters on the ground (but also vulture-like on carrion, tearing flesh with conspecifics as crocs & komodos do (which is not “conflict” but cooperation)), then leapt back up the tree like a lemur (or tree kangaroo or hoatzal). The tail flattened for flight and could encircle the tree trunk like a spider monkey, the stiff feathers help to grasp. In areas with few trees, became runner but still climbed, had very good eyesight, crepuscular/nocturnal, warmblooded, favored waterside sites but adaptable. May have eaten some sort of plant matter “fruit” occasionally, eg. gingko. My guess.
#26 Daniella Perea
June 8, 2009

DDeden: all I will say is that you must stop making such ‘guesses’. Seriously, there is a ton of evidence that is incompatible with just about everything you suggest.
#27 Hexapodoidingaedongaedae
June 8, 2009

Can you get your artificial mind to post here?

LOL!!!

Computing phylogenetic analysis of character matrix for: Aves, Maniraptoriformes, Dromaeosauridae, Troodontidae, Coelurosauria, Archaeopterygiformes, Archosauria
Using: Procrustean Elliptical Spline Transformation

dit
dit
dit
ding!

Computation complete.

Results:
75% probability hypothesized relationship is correct
17% probability Aves highly derives from Lissamphibia
15% probability Anatidae-Crocodylidae hybrid exists
02% probability Segmentation Fault – core dumped
0.01% probability contains creamy nougat filling

Warning: May Contain Nuts
WARNING ERROR DANGER DOES NOT COMPUTE DOES NOT COMPUTE
EXTRAPOLATE! EXTRAPOLATE!
+++DIVIDE BY CUCUMBER ERROR+++

Daisy, daaaiiiisyyyy…. ooooowerp.

.
#28 Aleck Zhou
June 8, 2009

“Dromaeosauridae: note typical long prehensile tail for gripping branch”

Haha have you ever taken a look at the caudal vertebrae of a dromaeosaur? I figure snapping bones would make gripping branches quite painful.
#29 Pete Buchholz
June 8, 2009

John Jackson wrote:

No argument supporting BCF is ever listened to or taken account of in any way by the likes of Naish,

Now which part of BCF do you mean John (or is it Stephen? or Peter?)?

1) Bird flight evolved from “the trees-down,” not “the ground up.”

2) The direct ancestors of birds all the way back to the divergence of the Crurotarsi were never greater than about 10 kg.

3) The direct ancestors of birds all the way back to the divergence of the Crurotarsi should a) be called “birds” and b) had some sort of teleological inclination toward becoming “birdy.”

Argument one is perfectly reasonable, testable, and in the mainstream of paleontological thought.

Argument two is not directly testable, but the principle of parsimony (or Ockham’s Razor or whatever you want to call it) can be applied, and unfortunately, argument two fails when you look at basal tetanurans. It’s perfectly possible that basal spinosauroids, basal megalosauroids, and basal carnosaurs all evolved from very small basal tetanurans, but quite frankly it’s way more parsimonious to assume that the “direct line” leading to birds went through a “body mass far greater than 10 kg” stage when Tetanurae diverged in the Early Jurassic.

Argument three…… is ridiculous, and was what got most people hung up on the BCF idea on the DML for the last 15+ years. Everyone, and I mean everyone would consider the clade {Archaeopteryx lithographica + Vultur gryphus} to be “birds.” That’s what birds are both scientifically and in common speach. Some people would also consider many non-avian maniraptorans “birds” if they saw them in the flesh. But no one would consider the MRCA of the clade {Allosaurus fragilis + Vultur gryphus} a “bird.” Even fewer people would consider the MRCA of clade {Iguanodon bernnesartensis + Vultur gryphus} a “bird.” And George Olskevsky himself would probably hesitate to call the MRCA of the clade {Pterodactylus antiquus + Vultus gryphus} a “bird.”

Yet that’s what George wants us to do. He wants this line of MRCAs to be called “birds” in addition to real birds…. Because this line of tree-dwelling, under 10 kg, never-fossilized dinosaurs were always going to become birds, despite the fact they left lots and lots and lots of ground-dwelling, way over 10 kg, and commonly fossilized descendants.
#30 DDeden
June 8, 2009

@ 26 – evidence to the contrary? Please specify.
@ 28 – prehensile tail does not equate with curliness. I was speaking of function (like emperor penguin tail as prop, also woodpeckers IIRC) in arboreal transit/posture). Note: feathers (partly) encircled trunk, not vertebrae. http://en.wikipedia.org/wiki/Dromaeosauridae
#31 Zach Miller
June 8, 2009

I think DDeden was being sarcastic.

John Jackson and I (and a few readers) already had it out on my blog a few weeks ago about this very topic.
#32 Mokele
June 8, 2009

One issue that’s rather important here: Why go up into the trees at all? Once angiosperms and fruit evolve, there’s plenty of high-nutrient food up there (and predators will follow the herbivores), but prior to that? How nutritious *are* gymnosperm seeds, really? I mean, they’re digestible, but enough to warrant the difficulty and risk of climbing into a tree to get them.

I should also point out that arboreality is NOT always obvious from skeletal remains. Sure, there are highly adapted species like gibbons and squirrels, but can anyone look at the skeleton of Sceloporus or Peromyscus and tell how much time they spend in trees?

And remember – most small animals climb extensively, especially those in forested habitat. When you’re small, a rock is a mountain and a fallen tree is a huge obstacle. Small animals live in rugged world with many of the same challenges faced by arboreal species (slopes, impinging branches, surfaces difficult to balance on, etc.)
#33 Tor Bertin
June 9, 2009

Insect feeding? I’d imagine that many species of insects would have formed to utilize the trees as habitat and a source of nutrition. If there are very few insect-feeding Pterosaurs in the area, it’d be an excellent niche for them.

Alternatively if they fed on smaller dinosaurs/mammals, they could have used trees as a vantage point to find prey.
#34 Nathan Myers
June 9, 2009

Mokele: What would keep small animals — already adapted, as you say, to climbing over fallen trees — from climbing up still-standing trees? Risk of injury from falling is negligible for small animals. In the modern world, climbing a tree may be risking attack, but when the treetops were less populated with flying predators, it must have been a much safer place to spend time, even when there wasn’t much food there.
#35 Carlos
June 9, 2009

I think John Jackson is an insany like Peter Mihalda
#36 DDeden
June 9, 2009

@ 31 – speculative, not sarcastic, I’d thought.
@ 32 – Why up? Ask a baby komodo, they’re predators too. Gymnosperm seeds are rich in protein and lipids IIRC, though maybe it was the insects that were food for small ones.
BTW, I just read in the local paper of a salamander with a partial-prehensile tail seen and filmed climbing a tall redwood tree, not unusual, redwoods grow past 350′ (110m) tall and are very ancient gymnosperms of the coastal fog belt.
#37 Nathan Myers
June 9, 2009

Carlos: If so, he has an excuse for rudeness — he can’t help it. What’s yours?

Ad hominem remarks insult everyone. Please do your part to keep TetZoo a pleasant environment for us all.
#38 David Marjanović
June 9, 2009

prehensile tail does not equate with curliness.

It does, because “prehensile” means “grasping”. To this day, prendre (note the dropped h) is the French word for “take”.

and “hyperextended” perching claw locked in tension,

1) It is in fact hyperextended. No need for quotes.
2) Perching? Then why has it got a cutting edge!?!
3) What do you mean by “locked”?

it leapt and “flyde” (guided glide + fly as controlled fall)

Aerodynamically, gliding and parachuting are opposites. You probably can’t really do both at the same time.

then leapt back up the tree like a lemur (or tree kangaroo or hoatzal).

Neither quetzals nor hoatzins leap… unless you count takeoff, but then, all flying birds count. They don’t use leaping as a way of locomotion.

The tail flattened for flight and could encircle the tree trunk like a spider monkey, the stiff feathers help to grasp.

I have, let’s say, trouble imagining that skin muscles could be so strong as to allow using feathers as a hand.

favored waterside sites

Why?

I just read in the local paper of a salamander with a partial-prehensile tail seen and filmed climbing a tall redwood tree, not unusual

Not unusual among arboreal plethodontid salamanders!!! Put a salamandrid on a tree, and it’ll just fall down.

Really, you still have a lot of learning to do before you will be able to form informed opinions. You haven’t reached that point yet.

It’s fine to speculate. It’s just useless to speculate without being able to test one’s speculations against reality. In many cases you are evidently unaware that such tests are even possible.

Ad hominem remarks insult everyone.

To be fair, this was not an ad hominem argument (of the form “I think you’re insane, so everything you say is wrong”). It was a conclusion: “everything you say is wrong in such ways that I think you’re insane”.

“Insult” and “ad hominem argument” are nowhere near synonyms. They’re completely orthogonal to each other.
#39 Dartian
June 9, 2009

Mokele:

Why go up into the trees at all?

I can think of a number of possible reasons (in addition to finding gymnosperm seeds to eat) why a small to medium-sized Mesozoic vertebrate would occasionally/regularly climb trees:

-To find food, e.g., sap, insects (including, from the Cretaceous onwards, termites), other invertebrates, and smaller vertebrates and their eggs & young.

-To escape non-climbing predators.

-To find safe roosting and nesting sites.

-To avoid being swept away by the water in certain habitats, such as mangrove swamps and flood forests (like those in the Amazon basin today).
#40 David Marjanović
June 9, 2009

from the Cretaceous onwards, termites

Aren’t there humongous termite mounds in the Morrison Fm? I thought termites go all the way back to the Triassic…
#41 DDeden
June 9, 2009

@ 38 – prehensile is grasp, not necessarily curly. That some animals grasp by constricting the tail wasn’t my point. 1) wasn’t sure if it was at its limit. 2) lateral stability, tail provided dorsal-caudal stability, (presuming the cutting edge was at the front of the branch, not on top). 3) to prevent unwanted flexion (during rest), the other digits appear to have a dorsal notch at the point of contact just behind their claws, the branch fit between them and the front claw. I think a harrier hawk during dive is in a controlled fall, that’s what I’d meant, not parachuting. “quetzals nor hoatzins leap”, fine, AFAIK no extant animal moves the same way as it did. “using feathers as a hand”, not as a complex hand, stiff spine feathers as a clutch-brake against gravity and torsion, very little muscular strength required. Waterside, initially (aqua-arboreal), later less so. “Really, you..” Please stop projecting, preach at your own blog.
#42 Dartian
June 9, 2009

David:

Aren’t there humongous termite mounds in the Morrison Fm? I thought termites go all the way back to the Triassic…

IIRC, Grimaldi & Engel, in their Evolution of the Insects (2005), were of the opinion that those Triassic mounds were probably not made by termites, and that the earliest indisputable evidence of termites is from the Early Cretaceous. (I’m no entomologist, however; hopefully Christopher Taylor will show up and provide some more authoritative information.)
#43 Darren Naish
June 9, 2009

I’ve asked the palaeoentomologists I know about these giant Morrison termite mounds (photo of one here). Nobody ever seems to know anything.
#44 Dartian
June 9, 2009

Pedantic correction to my previous post:

those Triassic mounds

I should have written “Triassic and Jurassic mounds” (there are putative Triassic termite mounds too, if memory serves me right).
#45 Aleck Zhou
June 9, 2009

DDeden,

Come on don’t be silly. You meant “long prehensile tail” as on a spider monkey. The term prehensile has never been used for the tails of woodpeckers or penguins, and even if we’re making up definitions on the fly, they certainly aren’t long or resemble dromaeosaur tails in any way. Just man up and admit that you did not know dromaeosaur tail morphology essentially kept them immobile.
#46 johannes
June 9, 2009

If there were no termites in the Morrison Formation, what was *Fruitafossor* eating?
#47 Lou Dattilo
June 9, 2009

Thank you Darren for your contribution to thinking outside of the box imposed by Victorian era orthodoxy! I look forward to more articles.
#48 llewelly
June 9, 2009

I thought Cope’s rule was an artifact of fossilization’s tendency to preserve large forms more frequently than small?
#49 Nathan Myers
June 9, 2009

David: I thank you for correction, but I took Carlos’s remark to mean “I think John Jackson is [objectionable] … and therefore we may reasonably ignore his attempts at discourse and, further, try to drive him away with insults.” We extrapolated differently, but neither, I think, wrongly.
#50 Michael Erickson
June 9, 2009

John Jackson is a ludicrous idiot.
#51 Michael Erickson
June 9, 2009

Make that a ludicrous babbling idiot.
#52 Michael Erickson
June 9, 2009

Yes, I know that was mean and should’nt have been said, but sometimes you just gotta call ‘em as ya see ‘em.
#53 Thomas R. Holtz, Jr.
June 9, 2009

You know, between some of the JJ comments here and the Quick & Ruben paper in Journal of Morphology, it makes me wonder if I haven’t fallen through a timewarp to the 1990s…
#54 Zach Miller
June 9, 2009

Tom, you mean the new paper about bird knees? I saw the press release, but I don’t see how the paper deals with the dinobird connection at all. I mean, the press release included a lot of high-fives from the BAND, but the connection between knee-driven walking/running and dinosaurs was not discussed (in the press release). And did none of those people read the recent Hutchinson paper discussing the gradual transition from “theropod” to avian postures?
#55 DDeden
June 9, 2009

@ 45 – No, not immobile. Spider monkey tails do not have spiny pen feathers to grip as Dromasaurs did, they have ventral eccrine volar surfaces. Functionally, in tree climbing, the tails were both prehensile climbing/posture aids. I read one account of an articulated S shaped Dromasaur~ish tail, so I thought there might have been some slight curvature (or swivel?) of the caudal vertebrae around the trunk, still not sure about that. But it was the feathers that held, like penguin and woodpecker, not the frictional skin like an oppossum or Atelidae.
#56 Andrea Cau
June 9, 2009

I agree with Zach: why a hypothetical series of avian autapomorphies would challenge the well established synapomorphy-based relationship between (non-avian) dinosaurs and birds? It’s only rethorical nonsense.
Not only the very good Hutchinson’s article: did they read the recent literature on saurischian air-sacs (listed here: http://svpow.wordpress.com/2008/10/04/the-aerosteon-saga-part-1-introduction-and-background/)?
#57 David Marjanović
June 9, 2009

prehensile is grasp, not necessarily curly.

Of course it is. Woodpecker tails don’t do any grasping.

presuming the cutting edge was at the front of the branch, not on top

How would that work? The entire ventral edge of the claw was cutting, all the way to the tip.

the other digits appear to have a dorsal notch at the point of contact just behind their claws

That’s to make (slight) hyperextension of those claws possible.

And holding a branch, with one’s body weight, against the dorsal side of two toes, against the extensor tendons, strikes me as… unhealthy and painful.

Waterside, initially (aqua-arboreal)

Why?

Please stop projecting

preach at your own blog.

Ain’t got none. You’re projecting.

I read one account of an articulated S shaped Dromasaur~ish tail,

Yes, Velociraptor.

so I thought there might have been some slight curvature (or swivel?) of the caudal vertebrae around the trunk

Far too little to grasp anything.

==================================

Thank you Darren for your contribution to thinking outside of the box imposed by Victorian era orthodoxy!

Erm.

As Darren writes loud and clear, it’s George Olshevsky’s contribution, and it’s dead wrong…

You might like to read the post again…

I thought Cope’s rule was an artifact of fossilization’s tendency to preserve large forms more frequently than small?

That could contribute, but even if that’s ignored, Cope’s “rule” is an artefact of doing science without math, and of doing evolutionary biology without a good phylogeny. Any time people use a method other than eyeballing to look at a case where the “rule” was claimed to apply, it evaporates. That includes my unpublished MSc thesis (which I need to do over with yet more data and with better methods, however).
#58 Rewi Kemp
June 9, 2009

Marvellous discussion, thank you, everybody.
Always kick the ” “/ball not the wo/man.
Cheers, Rewi Kemp
#59 William Miller
June 9, 2009

Re “Cope’s Rule”: I thought Dave Hone had a paper out recently (2007 maybe?) supporting the rule. He pointed out, IIRC, that tests of the rule often assume that the *smallest* species should get larger for it to count as an example, but that this really doesn’t make sense: small niches don’t go away. And clearly many groups’ average and maximum size increase over time.

I think that the main thing is that Cope’s rule can’t be used unmodified: it should be something like “in the absence of mass extinction events and effective competition in the large size niches, a lineage’s average size will tend to increase”. This would account for things like sauropods getting smaller between the Late Jurassic/Early Cretaceous and the Late Cretaceous – they were largely marginalized by ornithischian herbivores and so could not attain their full potential.
#60 Metalraptor
June 9, 2009

Ugh, its like the hydra, whack one, and another grows in its place. Anyway, excellent article Darren. What I find interesting is the fact of how Archie seems to have more and more in common with the deinonychosauroid maniraptorans as we dig deeper into its natural history. I mean the recently discovered “Thermopolis specimen” is said to have a hyperextendable sickle claw. If Archaeopteryx and its kin became a dead-end lineage, I bet that modern day paleontologists would just chalk them up as strange dromaeosaurs or possibly more basal deinonychosauroids.

As for Cope’s Rule, I’ve been noticing that several paleontologists have been mentioning this lately (both in a for and against context). News flash, Cope’s rule doesn’t exist, or at least not in the terms of modern paleontology. A lineage is just as likely to increase in size over time as decrease, it just does one or the other based on environmental factors influencing natural selection. Dwarf elephants, anyone? Or do I need to mention how there are still tiny little hyraxes living alongside elephants and formerly mammoths. Or better yet, compare a hummingbird to Herrerasaurus. Cope’s Rule indeed.
#61 Christopher Taylor
June 9, 2009

(I’m no entomologist, however; hopefully Christopher Taylor will show up and provide some more authoritative information.)

Flattery, as they say, will get you everywhere…

I don’t have much to go on either beyond Grimaldi & Engel (2005), who state that Isoptera body fossils are not known before the Cretaceous. Triassic wood borings have been attributed to termites, but Grimaldi & Engel suggest they may have been made by beetles (wood-boring beetles were definitely around at the time). I don’t know if they mention the Morrison mounds – first I’ve heard of them, to be honest – but a quick bout of Google-fu tells me that similar structures have also been found in the African Clarens Formation (also Jurassic). No fossils have ever been found in association with them, so it seems they could still be just about anything. One abstract seems to be implying that they may not even be biological in origin.

A big potential problem with a pre-Cretaceous origin for termites would be that it’s not just making a ghost lineage for the termites. The fossil record for the cockroach crown group (which termites are part of) also doesn’t seem to go back any further than the late Jurassic at earliest. Crown Dictyoptera (termites, cockroaches and mantids) don’t seem to go back much earlier. Stem-Dictyoptera (the Palaeozoic to Jurassic “cockroaches”) are quite readily distinguished from crown-Dictyoptera because they still have an ovipositor, and wouldn’t have produced the distinctive egg-cases of crown Dictyoptera. I’m not really buying Béthoux & Wieland’s (2009) recent identification of some Carboniferous fossils as stem-mantids, but only because that would be just as much (or maybe even more) of a trip for our understanding of dictyopteran stratigraphy as the identification of Protoavis as closer to modern birds than Archaeopteryx would have been for avian stratigraphy.
#62 Adrian Morgan
June 9, 2009

It can’t be a coincidence that you posted this article just as doubts about the bird/dinosaur link are back in the news. In fact, my reason for this particular visit to your blog (well, last night’s visit, actually, but I decided to sleep on it) is that I’d just read http://www.physorg.com/news163760732.html and for the sake of my ongoing education I wanted to read a balanced critique of it. If anyone had written such a thing, I thought, it would be Darren Naish. Are you planning to go there?
#63 Aleck Zhou
June 9, 2009

The problem with Cope’s rule is there’s really no scientific mechanism behind it. Rather it’s a correlation based on observation.

It is true that we do see an increase in organismal size over very long periods simply because the niche at the very top is always open. At a smaller size, niches are more likely to be occupied. Furthermore most extinction events have a stronger impact on larger organisms, freeing up those niches. However when niches are empty, a decrease in size can be just as common as an increase in size (island dwarfism, island giantism for example).
#64 Nathan Myers
June 9, 2009

Michael Erickson: It’s always been easy to be polite to the powerful and the admirable, but that tells nothing about us. We get our chance to demonstrate grace by showing politeness to the weak, the misunderstood, and the loony.
#65 DDeden
June 10, 2009

@ 57 – The tail feathers brace, the tail muscles grasp, probably using the same muscle movement as when landing on a branch. The two large claws and tail provided a secure triangle base on a branch. The ventral claw cutting edge surface assisted in lateral position but did not have weight on them, the weight was behind them on ventral pads. The smaller claws had ventral and dorsal knuckle pads, parallel to knucklewalking African apes (which have mid-finger tissue padding both ventral and dorsal), protecting tendons etc. That the lower claws could hyperextend strengthens this idea, allowing for different size branch circumference. I doubt it would be any more painful than large male gorilla knucklewalking on dry ground.

Waterside, as parallel to African apes, vertical tree climbing, part-time brachiating (loose parallel to arboreal controlled fall-flight), some frugivory, some carnivory, atypical ground locomotion (knucklewalking qpal & bipal in ape, hyperextended claws in dromasaur).

I don’t know if English is your native tongue, but I was not projecting, I was asserting that you were projecting, and apparently preaching, on a Science blog, rather than at your own pulpit.

The vertebrae may have slightly encircled the stem, but it was the axial and radial spiny pen feathers that provided actual friction, and did so while the tail was generally downwards. Obviously, I’m not referring to the largest species, which derived from these smaller arboreal ones.
#66 DDeden
June 10, 2009

Re. tail feathers: “Once he’s picked up some speed, he tucks his wings in only to pull out of the death-defying descent at the last minute by abruptly spreading his tail feathers”.
#67 Dartian
June 10, 2009

Thanks for the information, Chris!

Johannes:

If there were no termites in the Morrison Formation, what was *Fruitafossor* eating?

Other insects, apparently. It should be noted that in their original paper, Luo & Wible (2005), strictly speaking, do not suggest that Fruitafossor was a termite-eater; rather, they say that its teeth were most similar to those of modern armadillos. And most armadillo species don’t really feed that extensively on termites (or on ants, for that matter) – their diets are more generally insectivorous, or even omnivorous.

(Also, the small size of Fruitafossor makes it somewhat unlikely that it fed extensively, never mind exclusively, on colonial insects. Most extant mammalian termito- and myrmecophages are relatively large and some even have body armour. This probably has to do with the fact that most colonial insects defend their nests vigorously.)

Reference:

Luo, Z.-X. & Wible, J.R. 2005. A Late Jurassic digging mammal and early mammalian diversification. Science 308, 103-107.
#68 Darren Naish
June 10, 2009

Adrian Morgan (comment 62) writes…

It can’t be a coincidence that you posted this article just as doubts about the bird/dinosaur link are back in the news.

It is entirely coincidental. I was not aware of the new Quick & Ruben (2009) paper when I wrote the BCF stuff, nor was my article at all inspired by James & Pourtless (2009).

In fact, my reason for this particular visit to your blog (well, last night’s visit, actually, but I decided to sleep on it) is that I’d just read http://www.physorg.com/news163760732.html and for the sake of my ongoing education I wanted to read a balanced critique of it. If anyone had written such a thing, I thought, it would be Darren Naish. Are you planning to go there?

Sorry, I’m not planning to cover it, predominantly because I don’t find it at all interesting nor worthy of review. In recent years Ruben and colleagues have made a career of publishing papers in which they assert that ‘birds cannot be dinosaurs because of [whatever, blah blah blah]‘. Quick & Ruben (2009) assert that non-avian theropods were fundamentally different in abdominal morphology from extant birds, and they hypothesise (note: hypothesise) that the sub-horizontal avian femur and its associated musculature might be required to prevent collapse of the lateral abdominal wall: non-avian theropods evidently moved their femora a lot during normal locomotion, and hence, say Quick & Ruben, could not have had abdominal air sacs. All of this is extremely questionable or just flat-out wrong (sternal movement etc. almost certainly was present in non-avian theropods, the ‘mobile thigh inhibits abdominal air sacs’ just doesn’t make any sense, and the authors ignore evidence for abdominal pneumaticity in non-avian saurischians): if the authors have set out to demonstrate anything, it is that evolution cannot happen.

As for James & Pourtless (2009): these authors use cladistics to test the hypothesis that birds are deeply nested within coelurosaurian theropods, and argue that they use an unbiased approach where non-dinosaurian archosaurs and other reptiles are included too (they include Longisquama among archosaurs for some reason, and even imply that it’s a proto-bird [p. 37]). The paper is full of really weird claims (e.g., that theropods can only be diagnosed by their intramandibular joint) and does a lot of stuff that’s bound to skew the results: they coded all characters of disputed homology as ‘unknown’ (p. 14), for example (and, as usual among those disputing the theropod affinities of birds, they ignore evidence showing that the disputes about homology are erroneous anyway). This is wrong because it makes an a priori assumption about homology, and it introduces loads of new question marks in the matrix for character states where we do have data. Furthermore, the choice of taxa is weird: it’s wrong to analyse theropods and other archosaurs without including at least some non-theropod dinosaurs. Finally, the trees they generated are entirely uninformative (they are mostly polytomies) and don’t provide support for any hypothesis, so quite how the authors can say that they found weaknesses in the ‘birds are theropods’ hypothesis is really not apparent. As an impartial test of archosaur phylogeny, this study fails miserably.

Refs – -

James, F. C. & Pourtless, J. A. 2009. Cladistics and the origins of birds: a review and two new analyses. Ornithological Monographs 66, 1-78.

Quick, D. E. & Ruben, J. A. 2009. Cardio-pulmonary anatomy in theropod dinosaurs: implications from extant archosaurs. Journal of Morphology doi: 10.1002/jmor.10752
#69 Adrian Morgan
June 10, 2009

Thanks for that.

When I read articles on science news sites that strike me as dubious, I often search ScienceBlogs in the hope of reading some knowledgeable blogger’s analysis of why the relevant claim is wrong, unlikely, or (as I’m open to any eventuality) actually plausible. I do the same when I read articles that strike me as interesting but sketchy on details. Such searches yield results often enough to be worth trying (although it would be nice to be able to restrict the results to recently-published items only).
#70 John Jackson
June 10, 2009

Let’s put the trash out first.

This is supposed to be a science blog, and the best reply to a comment you disapprove of is to disprove it. If you can’t do that, simple jeering is not a last resort, it’s not even an option; to do so indicates not only that you don’t have an answer but that you don’t have the kind of mind to produce a relevant answer. Genuine science-capable readers eager to save time will be able to judge the worth of a comment by the name, and will soon recognise the likes of Tygo Raxx, III, Naraoia, Owlmirror, Hexapodoidingaedongaedae, Carlos, and Michael Erickson, as literally a waste of space.
#71 John Jackson
June 10, 2009

I’m still awaiting Naish’s reply to the questions I posed as comments to an earlier blog, so there’s no point pretending we have a dialogue, but I will address his points:

24
you are an amateur with no relevant qualifications.

And I suppose Elliott Sober is also an amateur since Naish would say he isn’t a qualified palaeontologist. Perhaps that’s how cladists justify the way they’ve ignored for decades Sober’s recommendations not to accept cladograms at face value. A field such as palaeontology is made up of specialist sub-fields, and calling oneself a professional palaeontologist does not entitle you to some kind of privileged unassailable authority on any subject within the area you define, nor does it mean those without whatever the appropriate membership card is, are simply ‘amateurs’.

The interpretation of cladograms is at the heart of dinobird palaeontology. Whatever tickets one may or may not hold, an adequate level of expertise counts. There is more to cladogenesis than running a program, or even than choosing characters well, and it is simply not characteristic of a professional to accept cladograms at face value. When Naish stops doing so, that’s when I’ll stop criticising him for doing it. While he continues, he will be condemning himself more strongly than anyone else can.

24
Secondly, I don’t actually see why you think you need to understand the statistical algorithms and so on to be a perfectly good user of parsimony software. Please explain why you think this necessary.

It’s because you have to know how far their outputs can be trusted. I’ll admit the basic principle of selecting the simplest tree is hardly an algorithm at all, and that it’s irrelevant that the attempts to achieve this do become rather involved, but understanding the extent to which the simplest tree can represent evolution is the vital skill. With no reversals or convergences, cladograms would be 100% trustworthy; with too much, they begin to lie. What calculation can you do that warns you when this starts to happen? Is it when the number of nodes in the cheapest tree divided by its number of character changes is more than 10% of the number of characters? Or does it have to be more than the square root of the number of leaves? If we start to approach some such criterion, does that mean we can make no use of the tree, or can we tentatively hold on to it until it gets twice as bad? Does it help if some characters appear to misbehave at random but others seem to work together as two or more competing teams? What does it mean if one of the teams contains more characters than all the others? How do we know when the largest team is not the most honest? How do we even know when characters form a team? Does the age of the fossils have any relevance at all, and if so, how can it be used? Which of these issues are best addressed theoretically, and which by means of simulations?

If you have experience of simulating evolution, if you have enough experience and understanding of using algorithms, statistics and evolution to be able to write a program that weaves and tests its own theories out of patterns of observation that demonstrably aid its survival, then you’ll be able to approach the challenges above with a reasonable chance of success, and your attempts at tasks like reconstructing a tricky phylogeny will be worth considering. If you simply say things like ‘This possibility was tested using the following animals and characters, and taxon X was found to fall within taxon Y’, confusing the hypotheses nature of a cladogram with some ability to act as a test while repeatedly denying it, then your attempts will not be genuinely authoritative.

Time for breakfast.
#72 Darren Naish
June 10, 2009

I cannot believe that you are still using this ‘unquestioned acceptance of cladograms’ argument. Here is a slightly edited version of the response I provided last time (from here)…

————————–

I’m not really interested in your approach to this area given that it involves stuff which is beyond my interest and experience. Do I ‘support’ the ‘conventional’ phylogeny because it is found to be most parsimonious by computational phylogenetic analyses? Well, partly, but not necessarily. Forget cladistics and all of the assumptions about data analysis that you have a problem with: I return to the point I’ve made before – that examination (or comparison, or whatever) of the morphological data (you know, the raw information that our phylogenetic hypotheses are based upon) does not find ‘dromaeosaurs are nested within the Archaeopteryx + modern bird clade’ to be the best explanation for the data. You do a very good job of ignoring this, or of remaining ignorant of it.
#73 David Marjanović
June 10, 2009

I thought Dave Hone had a paper out recently (2007 maybe?) supporting the rule.

Yes… using a method that ignored the phylogeny. The pterosaur paper just used a linear regression, for crying out loud.

There is more to cladogenesis than running a program

Cladogenesis is divergent evolution. Splitting of a lineage. “Speciation” under some species concepts.

You pontificate about cladistics, but don’t understand its most basic terms? How… reassuring.

More later.
#74 David Marjanović
June 10, 2009

With no reversals or convergences, cladograms would be 100% trustworthy; with too much, they begin to lie. What calculation can you do that warns you when this starts to happen? Is it when the number of nodes in the cheapest tree divided by its number of character changes is more than 10% of the number of characters? Or does it have to be more than the square root of the number of leaves? If we start to approach some such criterion, does that mean we can make no use of the tree, or can we tentatively hold on to it until it gets twice as bad? Does it help if some characters appear to misbehave at random but others seem to work together as two or more competing teams? What does it mean if one of the teams contains more characters than all the others? How do we know when the largest team is not the most honest? How do we even know when characters form a team? Does the age of the fossils have any relevance at all, and if so, how can it be used? Which of these issues are best addressed theoretically, and which by means of simulations?

Why do you act as if those questions had never been investigated by cladists?

Because you don’t read the primary literature, that’s why. There are entire journals that you don’t seem to know exist. Yet you make grand pronouncements about how it’s all wrong. Argument from ignorance. Shame on you.

––––––––––––––––––––––

The tail feathers brace, the tail muscles grasp, probably using the same muscle movement as when landing on a branch. The two large claws and tail provided a secure triangle base on a branch.

On a trunk, you mean? The way Chatterjee imagines dromaeosaurids climbed?

(BTW, using the tail as a prop, the way woodpeckers do it and the way Chatterjee imagines dromaeosaurids did, is not grasping. I don’t understand why you glue yourself to that word.)

The ventral claw cutting edge surface assisted in lateral position but did not have weight on them, the weight was behind them on ventral pads.

Fine, but why is the 2nd toe (except for the claw) so short, then? And why does the hyperextension happen between the first and the second phalanx, not between the second and the third (which is the claw)?

The smaller claws had ventral and dorsal knuckle pads, parallel to knucklewalking African apes (which have mid-finger tissue padding both ventral and dorsal), protecting tendons etc.

Is that speculation testable by bones alone? Probably not…

Waterside, as parallel to African apes,

Why, why, why? Why should dromaeosaurids parallel African apes in any way at all? What do they have in common skeleton-wise?

(And how much does “waterside” even apply to African apes?)

I don’t know if English is your native tongue, but I was not projecting, I was asserting that you were projecting, and apparently preaching, on a Science blog, rather than at your own pulpit.

It’s not my native tongue, but I understood full well what you’re saying, and I maintain that it’s you who’s projecting here. You project your lack of knowledge on the rest of the world.

You speculate and speculate, floating high above any data, building castles in the air. Half of what you confabulate is untestable (and therefore useless to science), and the other half is already disproved by data you didn’t know or even imagine exists. You are the one who are abusing a science blog to launch one trial balloon after another (not even noticing that half of them are heavier than air).

And then you accuse me of doing all that? That’s just funny.

While I was already laughing, I tried to continue the joke by saying you projected having a blog onto me.

So far, you’re a source of entertainment. Just try to learn more anatomy. There’s actually much on it online.
#75 Michael Erickson
June 10, 2009

Nathan Myers – you are absolutely right. Sorry.
#76 Michael Erickson
June 10, 2009

John Jackson – I must say you are right on that part. My comments were uncalled for, I apologize. No more waste-of-space comments from the likes of me.
#77 Nathan Myers
June 10, 2009

John Jackson: As I understand it, the demands on rigor in interpreting cladograms are much relaxed if they are used as a source of hypotheses, or to explore a space of hypotheses, rather than as proof for one. The hypotheses really need other support, anyhow, based on deep, detailed anatomical knowledge. No cladogram, however rigorous, can come close to substituting for that.

Sometimes a body of raw facts, such as DNA analysis, can trump both a rigorous cladogram and the entire population of experts. Then the experts have to adjust, but (mostly) they do.
#78 David Marjanović
June 10, 2009

%lt;sigh>

As I understand it, the demands on rigor in interpreting cladograms are much relaxed if they are used as a source of hypotheses, or to explore a space of hypotheses, rather than as proof for one.

Cladograms are not used for any of these possible purposes. Ideally, you start from a dataset (a matrix: taxa x characters), and then have the computer calculate all possible trees for these taxa (automatic generation of all possible hypotheses)*, calculate the length of each tree (that is, how many assumptions of evolution — of character-state changes — it requires), and then pick the shortest one(s) based on the principle of parsimony, also known as Occam’s/Ockham’s Razor, which is one of the two parts of the scientific method. The other part, outright disproof by evidence incompatible with the hypothesis in question, is not possible in phylogenetics**, because, in principle, anything can evolve convergently. There is no absolutely reliable character.***

So, cladistic analyses ( = phylogenetic analyses) are used to test very large numbers of hypotheses against a given dataset using the principle of parsimony; not for generating, proving, or (strictly speaking) disproving hypotheses.

(Proof is impossible in science anyway. Sure, you can prove beyond reasonable doubt, but you can’t define “reasonable”.)

That “against a given dataset” part is important. Adding taxa and/or characters or simply correcting mistakes in a dataset often changes which hypothesis is most parsimonious. The “deep, detailed anatomical knowledge” you mention is necessary to make a dataset for cladistic analysis in the first place.

Cladistic analysis can be, and is often, done on DNA sequence data. Even mixed datasets are possible and have been analyzed in published papers.

Both empirical and theoretical studies show that more data — bigger datasets — is better.

Any more questions?

——————

* This is where “ideal” and “real” almost always diverge. The number of unrooted trees grows extremely fast with the number of taxa; an exhaustive search as described here is impossible for more than 11 to 12 taxa because it would take forever. Instead, other algorithms are used which do not look at every possible tree but still find all most parsimonious trees… almost always.
** This is a point Mr Jackson likes to harp on ad nauseam. He’s entirely right about it… but then he acts as if he had never heard of parsimony…
*** Though SINE/LINE insertions come very, very, very, very close.
#79 Naraoia
June 10, 2009

Both empirical and theoretical studies show that more data — bigger datasets — is better.

Aren’t there cases where more data just lend more support to the wrong phylogeny?
#80 Christopher Taylor
June 10, 2009

Aren’t there cases where more data just lend more support to the wrong phylogeny?

Yes, there can be – for instance, if you’re comparing taxa that have undergone high amounts of convergent evolution, and/or if your dataset includes a lot of correlated characters (if a change in one feature automatically causes a change in another feature, you may be counting as two separate changes something that really should be counted as just one). There are ways of getting around these – increase the number of taxa coded to break up long branches, etc. – but unfortunately not all of these will be available in every case. For instance, if you’re working on living taxa, checking the results of multiple data types against each other (e.g. morphological vs. molecular) can act as a test of any individual data type (if they disagree, you’ll want to know what makes them disagree and why). Stratigraphy and biogeography may also act as checks – while they’re not normally included in data sets for analysis (perhaps more a question of how to code them in a way that the computer can handle the analysis appropriately than any other problem), any researcher who finds a Triassic taxon nested well within an otherwise purely Tertiary clade, or a taxon from Mauritius as the sister to a taxon from Tahiti, is going to be at least double-checking his data (the result may not necessarily be wrong, but a closer look never hurts).

Recognising correlated characters can be a bit trickier, but it usually comes down to familiarity with the organisms you’re working with, plus a bit of common sense.

Their invocation of convergence, off-hand, has always been one of my major points of confusion about the BAND crowd. It has been their favoured “explanation” for the morphological similarities between (non-avian) theropods and birds, but I’ve never heard a suggestion as to why two such ecologically different groups should be so strongly convergent in the first place.
#81 Nathan Myers
June 10, 2009

David: Thank you for the correction. But, what would you call the output of a cladistic analysis? It’s the best (by one criterion) of a universe of hypotheses, but it’s still a hypothesis, and one you might not have happened upon otherwise. Also, deep knowledge is manifestly not necessary to make a dataset; it’s needed to make a useful dataset. Deep knowledge you don’t have yet (nobody knows everythign) is needed to test the output independently. No?
#82 DDeden
June 11, 2009

@ 74 – IMO birds did not evolve from the ‘ground-up’ from fast running animals that gradually took to the air. Flyers (insects, birds, bats, pterosaurs), brachiators and gliders all derived from arboreal (green plant) waterside ancestors (except flying fish). Theropods derived similarly, sauropod types did not. So… grasping as in tensional attachment, as opposed to a terrestrial sauropod locomoting compressionally, where the tail is not used as frictional anchorage. The large-clawed second toe would be weaker if longer, poorer leverage in hyperextension. The closest extant parallel would be a fore-clawed owl or hawk, perching near and above water to ambush prey. Not impossibly it may have fed on ginkos and dropped some, ‘accidentally’ baiting the area, as chimps do with monkeys. Some saber-tooth cats probably had a similar aqua-arboreal ambush niche, but nearer the ground. The larger, later ones became more like savanna lions and T rex, due to selection in prey for improved water conservation. African ape and our ancestors formerly were 20ma part-time wetland foragers that then increased arboreal frugivory/omnivory (and so sharing some patterns with dromasaurs and birds). Later, savanna chimps, mountain gorillas and humans diverged from that pattern, and adapted unique terrestrial locomotion (knucklewalking ape, upright human). Similarly, dromasaurs retained the odd scythe claw, probably using it similar to a saber cat’s canines to slash exposed arteries in ambush attempts, which is why most non-humungous dinos had head and dorsal armor, to prevent back-grasping (= tree grasping) predators from getting a good grip without unwanted ventilation. BTW, your use of the word ‘projecting’ is different than what I’d meant. I meant in the psychological sense. Don’t be so defensive, I’m merely tying up loose ends, noting the parallel niches through time filled by anatomically different but functionally similar species.
#83 Nathan Myers
June 11, 2009

DDeden: Why “waterside”, again?
#84 Nemo Ramjet
June 11, 2009

DDeden,
Wow, man, I think you are on to something here! These projecting people know nothing. I agree that small dromaeosaurs climbed waterside gingko trees and occasionally ate waterside nuts. But I think it was their s-shaped necks that did the climbing – the spiny, dextrous tail was obviously used to crack the tough nuts. I know this is true because a frog do the same.
#85 David Marjanović
June 11, 2009

Aren’t there cases where more data just lend more support to the wrong phylogeny?

Yes. That happens in two different cases:

1) If you have correlated characters in your dataset. That’s the same as having one character that is given more weight than the others for no good reason.
2) If you have characters that have undergone such fast evolution that their phylogenetic signal got overwritten. This leads to long-branch attraction* and long-branch repulsion**. That’s a common problem with molecular data, but not unknown for morphological data either.

The solutions to 1) are statistical tests for correlation and “deep, detailed anatomical knowledge”. For example, is it even possible to have an antorbital fossa without having an antorbital fenestra first? “Deep, detailed anatomical knowledge” might answer that.

The solution to 2) are statistical tests for correlation*** and… yet more data! More taxa, so that long branches get broken up; more and clearly independent characters (for example completely different genes) so you see what other evidence says. It is, after all, not likely that different characters contain the same random false signal. It’s also good to look at what signals different subsets of the character set contain; if they point in different directions, that needs to be explained. So, what Christopher said.

====================

what would you call the output of a cladistic analysis? It’s the best (by one criterion) of a universe of hypotheses, but it’s still a hypothesis, and one you might not have happened upon otherwise.

Correct.

Of course, that it’s a hypothesis is not a bad thing. It’s too small to be a theory anyway, and a fact is something different in the first place…

Also, deep knowledge is manifestly not necessary to make a dataset; it’s needed to make a useful dataset.

That’s unfortunately true.

Deep knowledge you don’t have yet (nobody knows everythign) is needed to test the output independently. No?

Yes, because it’s required to 1) find mistakes in the dataset that was used**** and 2) make an even more useful dataset.

* When a clade with lots of autapomorphies = too fast evolution has lost too many of the character states it once shared with others, it finds itself attracted to the root. When there are several, they are attracted to each other and the root…
** When two clades that are in fact closely related lose too many of their shared character states, they will be found apart, because random similarities to other clades will outweigh the vanished evidence for keeping them together.
*** Characters that correlate too much to each other are probably just that, correlated characters, and should be turned into a single character (or how many correlation sets there are). Characters that correlate too little to any others probably just contain random noise and no phylogenetic signal. – Unfortunately, the only rigorous statistical test for both of these at once is so hard to implement that it has only been applied twice, in Robin O’Keefe’s thesis and in O’Keefe & Wagner 2001 (a paper in Systematic Biology). There’s no software available for it, you’d have to program it yourself.
**** Typos are horribly common in published data matrices, and so are similar mistakes. See here for an example where 35 % of a published matrix was, to varying degrees, wrong; the results changed drastically when we tried to correct that.

=============================

IMO birds did not evolve from the ‘ground-up’ from fast running animals that gradually took to the air. Flyers (insects, birds, bats, pterosaurs), brachiators and gliders all derived from arboreal (green plant) waterside ancestors (except flying fish). Theropods derived similarly, sauropod types did not.

Fine.

But what do you base this opinion on?

BTW, your use of the word ‘projecting’ is different than what I’d meant. I meant in the psychological sense.

That’s exactly how I mean it.

I’m merely tying up loose ends

I’ve been trying to point out that most of them are neither ends nor loose.

Again: half of what you say is untestable and therefore useless except if you want to write a Raptor Red-like book (which, I hasten to add, would be a worthwhile endeavor in its own right!), and the other half contradicts evidence that you happen not to know.
#86 Nathan Myers
June 11, 2009

Deep knowledge you don’t have yet (nobody knows everything) is needed to test the output independently. No?

Yes, because it’s required to 1) find mistakes in the dataset that was used and 2) make an even more useful dataset.

Are facts that can’t be represented as matrix entries beginning to be discounted as irrelevant?

The word “hypothesis” seems much wider than “a possible output of cladistic analysis”. Even within cladistic analysis, it seems common to have an interest in a single inheritance relationship, and to run analyses with varied input — e.g. assuming each side of a dispute — to see how robust that detail is against such uncertainties. The hypothesis of interest, then, isn’t a particular output tree, but a conjecture about how output trees will be affected by input details.

My impression is that the poster child for failed cladistic analysis must be the placement of falcons among raptors, flatly contradicted by DNA evidence despite perhaps the most thorough character analysis ever conducted. Did I misunderstand that result?
#87 David Marjanović
June 11, 2009

Are facts that can’t be represented as matrix entries beginning to be discounted as irrelevant?

What would these be?

Maybe you’re thinking of continuous characters, whose values are real numbers rather than integers. There are ways of dealing with these in data matrices; I’ve done it myself.

My impression is that the poster child for failed cladistic analysis must be the placement of falcons among raptors, flatly contradicted by DNA evidence despite perhaps the most thorough character analysis ever conducted. Did I misunderstand that result?

On the one hand, I don’t know if anything might be wrong with those cladistic analyses of DNA sequence data. (Molecular analyses of bird phylogeny often produce trees with very poor resolution.)

On the other, the gihugrongous analysis of morphological data by Livezey & Zusi (2007) is… brace yourself… not gihugrongous enough in terms of taxon sampling. It contains almost no fossils. I can’t help thinking that this is why, for example, the loons and the grebes were found as sister-groups.
#88 Nathan Myers
June 11, 2009

Are facts that can’t be represented as matrix entries beginning to be discounted as irrelevant?

What would these be?

I guess that means yes.

If (nearly?) the gihugrongousest analysis ever conducted isn’t giwhatsit enough, what does that say about every smaller analysis, e.g. only involving fossil specimens?
#89 DDeden
June 12, 2009

@ 83 – Because that’s where the optimal combination of water, thirsty prey focused on water, tall trees (shielded against pterosaurs) with low branches (due to open water) for climbing, co-existed. I’m not saying this was the only place, just a favorable one for a dromasaur type predator (as well a raptor, saber-cat and clouded leopard), I’d think.

@ 84 – not sure about the nugivorous frog, but snake scales and dromasaur pen feathers may have had the same frictional climbing ability, just like ‘fishscale’ pattern on skis, but the feathers function better on thick fissured tree bark, most likely.

@ 85 – Comparative physiology, comparative anatomy, optimal apex niche occupation of flora, distance from available metabolic fluid replenishment of prey/predator, availability of alternative foods (fruit/insect), etc. Why are you so defensive against sensible speculation? Allergy?
#90 David Marjanović
June 12, 2009

I guess that means yes.

Really, are there any phylogenetically informative characters that can’t be put into a matrix? I can’t think of one. Help me.

If (nearly?) the gihugrongousest analysis ever conducted isn’t giwhatsit enough, what does that say about every smaller analysis, e.g. only involving fossil specimens?

Well. That analysis was meant to tackle a really huge problem, and it did so with an uneven taxon sampling. Analyses with more fossils in them are usually supposed to handle much smaller problems (smaller parts of the tree of life), and their taxa are more evenly distributed in time.

======================

Because that’s where the optimal combination of water, thirsty prey focused on water, tall trees (shielded against pterosaurs) with low branches (due to open water) for climbing, co-existed. I’m not saying this was the only place, just a favorable one for a dromasaur type predator (as well a raptor, saber-cat and clouded leopard), I’d think.

OK, “a favorable one” sounds better. But why climbing? Why do you think dromaeosaurs climbed?

BTW, a dromasaur is this.

Comparative physiology, comparative anatomy,

Sounds good — what is it? All you’ve explained so far is that you think dromaeosaur feet and tails did not make climbing impossible (…and I explained that I disagree about the feet).

optimal apex niche occupation of flora

What do you mean? I don’t understand… for example, an ecological niche is not a physical space…

Why are you so defensive against sensible speculation? Allergy?

I’m trying to get through to you that most of it is not sensible, and that you have to test it.
#91 DDeden
June 12, 2009

@ 90 – “dromaeosaur” rather than dromasaur, thanks. “But why climbing?” Better visual and launching vantage point above the ground. “what is it?” Hooks, like sloth claws and gibbon fingers, but for upright climbing/perching not hanging/brachiation, to allow fast “spiral staircase” vertical climbing/hopping up gymnosperm radial annual branch nodes, moving upward and forward around the stem to the preferred height of ambush, then waited like owls. Again, waterside trees, (unlike dense forest trees), retain their low branches, so are easily climbed this way. They did not have opposable thumbs, so they evolved opposable ankle claws and bracing tail feathers. I guess their wings-arms had curved claws.
#92 David Marjanović
June 12, 2009

“But why climbing?” Better visual and launching vantage point above the ground.

Sorry, I didn’t make myself clear enough. Why do you think dromaeosaurs did climb? Why don’t you agree that they lived on the ground?

Hooks, like sloth claws and gibbon fingers

Then why the cutting edge, and why only one such claw per foot? And why does the rest of the anatomy not allow any serious climbing?

hopping up gymnosperm radial annual branch nodes, moving upward and forward around the stem to the preferred height of ambush, then waited like owls. Again, waterside trees, (unlike dense forest trees), retain their low branches, so are easily climbed this way.

Ah. Now it gets interesting! Thank you, this makes sense (…even though I fear it’s still not testable, so if it’s wrong, we can’t find out that it’s wrong).

They did not have opposable thumbs, so they evolved opposable ankle claws

None of the toes are opposable.

I guess their wings-arms had curved claws.

This guess is correct. The finger claws were similar to the toe claws of eagles.
#93 Nathan Myers
June 12, 2009

Really, are there any phylogenetically informative characters that can’t be put into a matrix? I can’t think of one. Help me.

Really you should ask some old-timers, e.g. Darren . But I doubt it handles uncertainties or conditional facts cleanly. As a completely artificial example, because I’m pig-ignorant: Is that a 1st or 2nd digit? If it’s a 2nd digit, the 1st is lost and the 5th is present; otherwise that’s the 4th.

Many facts aren’t phylogenetically informative but remain biologically important.

By the way, I don’t agree that nothing in biology makes sense except in light of phylogenetics. As I understand from people working on bacteria and archaea, their phylogenetics are a near total muddle as a result of them (the cells, not the people) slurping up and incorporating random environmental genetic scraps for a half-billion years. People studying them just have to keep busy in the phylogenetic dark. And they do.
#94 DDeden
June 13, 2009

Why climb? food. insects, fruit, waterside ambush. They probably nested on the ground or treetrunk hollow. Early ones were like saber cats and cloud leopards, later ones became open-ground adapted, like lions and cheetahs, as the prey were selected to minimise water use or got armored.

Neandertals used the same climb & ambush method at beaver ponds, assisted by herd drivers and backfloating spear-jabbers. A bowhunting friend on his ladder-stand at waterside got off a bad leg shot, so he leapt down on the deer with his hunting knife and slit its throat. Stomp, stun & slash, old method, works until prey evolves armor.

One sickle per foot gave lateral stability (not weight bearing) so the arm-wings could tuck and fold for perch & rest. Troodonts were similar but nocturnal, keen ears, may have used mouth more prehensilely.
#95 David Marjanović
June 13, 2009

Is that a 1st or 2nd digit? If it’s a 2nd digit, the 1st is lost and the 5th is present; otherwise that’s the 4th.

Aaah. In cases where “unknown” or partial uncertainty (like “states 0 or 1 but not 2″) don’t work, all you can do is run the analysis twice, once coded one way and once the other, and then compare the results. I’ve seen that done in some publications.

As I understand from people working on bacteria and archaea, their phylogenetics are a near total muddle as a result of them (the cells, not the people) slurping up and incorporating random environmental genetic scraps for a half-billion years.

That happens (and has certainly happened for much longer than a mere half-billion years!), but it’s not that bad. For example, there’s a nice paper in the latest Systematic Biology on the root of the tree of life.

Why climb? food. insects, fruit, waterside ambush.

I repeat: I’m not asking why anyone would climb. I’m asking why you think that dromaeosaurs actually did climb so often that they could be called scansorial, let alone arboreal.

Evidence for or against that idea has to come from the skeleton. You brought up the sickle claw, and I showed why it doesn’t count. You brought up the tail, and it’s clear that alternative explanations can’t be excluded for that either (balance, outright rudder for running, sexual selection…). Can you find any other evidence that indicates that dromaeosaurs really did climb?

Neandertals used the same climb & ambush method at beaver ponds, assisted by herd drivers and backfloating spear-jabbers. A bowhunting friend on his ladder-stand at waterside got off a bad leg shot, so he leapt down on the deer with his hunting knife and slit its throat.

Nice story, but how can you show it really happened?

(In those places where neandertalers and trees co-occurred. Which aren’t that many.)

One sickle per foot gave lateral stability (not weight bearing)

Then why has no other climber evolved such a claw, and why has it got a cutting edge? And why is it so big?

Wouldn’t something like a primate/marsupial/drepanosaurid foot work much better for stability in all directions?
#96 DDeden
June 14, 2009

Hang a sickle and a gaffing hook on a branch, the hook can slide laterally, the sickle cannot, tetrahedral tripodal bipedal lateral stability. Today’s birds have a secure sleep-proof rear claw/s which does away with the need for a vertebral tail prop and sickle stabilizer in upright orthogonal stance. Once the rear grasping claw evolved, long vertebral tails and teeth were lost and birds flew high, which gradually displaced dromeaosaurids and troontids from apex arboreal ambushers, they became large ground runners. Long vertebral tails initially braced, not for running balance or rudder (weight), or sexual selection (did not protect hutch unlike peacock), and females probably had the same tail prop.

Neandertals hunted waterside via ambush, young skinny males in trailside trees, older fat males in water with long spears, females prodding herd to waterhole. Later, more open herd hunting, like lions and wolves.

Primate digits didn’t evolve for “falling” but for climbing and plucking. The dromaeosaur’s advantage was a fast controlled stomp tackle, it wasn’t a great climber, just an adequate one.
#97 David Marjanović
June 15, 2009

Today’s birds have a secure sleep-proof rear claw/s which does away with the need for a vertebral tail prop and sickle stabilizer in upright orthogonal stance.

Yes. Why didn’t that evolve right away?

birds flew high, which gradually displaced dromeaosaurids and troontids from apex arboreal ambushers

Not remotely comparable. It took a long time till birds reached that kind of size; no flying birds of such size exist today, for example.

not for running balance or rudder (weight),

Thin and muscle-poor as the weight is, I doubt that’s an argument.

or sexual selection (did not protect hutch unlike peacock)

Irrelevant. Traits for which there is sexual selection are usually useless; that’s even the point.

Neandertals hunted waterside via ambush, young skinny males in trailside trees, older fat males in water with long spears, females prodding herd to waterhole. Later, more open herd hunting, like lions and wolves.

What makes you think so?

Primate digits didn’t evolve for “falling” but for climbing and plucking. The dromaeosaur’s advantage was a fast controlled stomp tackle, it wasn’t a great climber, just an adequate one.

As far as I can see, they were rather inadequate climbers.
#98 Graham King
June 20, 2009

Thanks, commenters, for my several prolonged and hearty LOL reading the above. But much of it is also beyond my ability to comment.

So here is a photo which I hope will make it all clear..