Tutorial 7: the sauropod family tree
August 19, 2009
We really should have covered this ages ago … Here we are, blithering on about brachiosaurids and diplodocoids and all, and we’ve never really spelled out what these terms mean. Sorry!
The family tree of a group of animals (or plants, or fungi, or what have you) is called its phylogeny. The science of figuring out a phylogeny is called systematics. And once you’ve got a phylogeny, the business of naming the parts of it (and of course choosing which parts to name) is taxonomy.
For a long time, sauropod systematics was completely up in the air, so that McIntosh’s (1990) review article on sauropods in The Dinosauria (first edition) said, rather despairingly, that “although recent discoveries are beginning to clarify the problems of sauropod phylogeny, were are still very far from being able to construct a cladogram” (p. 399). Happily, this changed rapidly thereafter, with the first published numerical phylogenetic analysis appearing in Russell and Zheng’s (1993) description of the new Mamenchisaurus species M. sinocanadorum. More importantly, in the same year Paul Upchurch submitted his (1993, duh) dissertation on sauropods, and this contained a much larger analysis which was subsequently published as Upchurch (1995). This paper raised the bar significantly, with an analysis of 27 taxa using 174 characters. Three years later, Upchurch (1998) published a major revision of his own work; in the same year, the other major school of sauropod phylogeny launched with a JVP memoir (Wilson and Sereno 1998), which featured only 10 taxa and 109 characters, but discussed and illustrated them in more detail. Wilson (2002) followed this up with a much larger analysis of 27 taxa and 234 characters, and Upchurch et al. (2004), in the second editi0n of The Dinosauria, saw his 27×234 and raised him to 41×309. The good news is that, by this time, the two schools’ phylogenies, having started out rather different, were converging on a consensus topology with only two significant disagreements, which we’ll come to in a minute.
Since then, Jerry Harris (2006) created a union matrix from the character scores in the Wilson (2002) and Upchurch et al. (2004) matrices, and also threw in a few additional characters from other less ambitious phylogenetic analyses. This analysis came up with a tree that was very similar to Wilson’s, and subsequent work by Wilson and Upchurch (2009) indicates that Upchurch is now also substantially in agreement with this arrangement.
So here it is!
Consensus phylogeny of sauropod, from Harris (2006)
I plucked this from Jerry’s paper, and coloured it in to show two of the more important groups. Evolution begins at bottom left, so let’s quickly tour the group.
- First of all, note the outgroups. Sauropods’ nearest relatives are the other saurischian dinosaurs, theropods and prosauropods. (They’re shown the wrong way round here, because in an unrooted tree it makes no difference. Ignore that.)
- The most basal sauropods include things like Vulcanodon and, it turns out mostly from the work of Adam Yates (e.g. Yates 2007), a whole bunch of things that, if you looked at them you’d probably guess were prosauropods.
- Sauropods as we know them really begin at the boundary of the group Eusauropoda (“true sauropods”), which is roughly speaking everything more derived than Vulcanodon. (I won’t discuss the naming of nodes and branches in detail in this post, as it would quickly get too long. Maybe in a later tutorial.) This group I have coloured pink in the diagram above.
- Basal eusauropods include quite a few genera, and the order in which they branched off the “main line” leading to the neosauropods is not clear — as the unresolved polytomy above shows. Cetiosaurus (which for some reason is not shown in this figure) is generally considered quite derived; some of the Chinese sauropods (Mamenchisaurus, Omeisaurus, etc.) may form a group of their own, but that’s not clear.
- Most of the best-known sauropods fall within the great group Neosauropoda (“new sauropods”), which is coloured purple above. A few genera float around the root of this group, including Haplocanthosaurus and Jobaria, both of which are sometimes considered neosauropods, and sometimes non-neosauropod sauropods (or what I informally call “eosauropods”, or “dawn sauropods”).
- Otherwise the great split within Neosauropoda is between the diplodocoids (on the left) and the macronarians (on the right) — the groups including Diplodocus on one hand, and Saltasaurus on the other.
- The most basal diplodocoids are the rebbachisaurids over on the left.
- Most other diplodocoids fall into the group Flagellicaudata (“whip-tails”), which is itself composed of dicraeosaurids and diplocodids. (It’s not clear where in that dichotomy, or maybe just outside it, Suuwassea falls.)
- Over in the other half of the Neosauropods, the first macronarians to diverge are the camarasaurids (which currently means, uh, Camarasaurus).
- Most of the other macronarians fall into Titanosauriformes, the group uniting brachiosaurids (yay!) with titanosaurs and their buddies. Everything closer to titanosaurs falls within Somphospondyli, and that includes Euhelopus — as it turns out. (Upchurch had found Euhelopus to fall outside Neosauropoda).
- Once you get past Euhelopus, you’re into Titanosauria (though there are various definitions which place the entry point differently).
- And once inside Titanosauria … well, all bets are off at this stage. There is a rough consensus that things like Malawisaurus and Andesaurus are pretty basal and Saltasaurus is, sort of by definition, derived. But apart from that, different studies have come up with wildly different phylogenies, with that of Curry Rogers (2005) being particularly left-field.
Without a doubt, Titanosauria is where the action is right now. As alluded to in the comments of Matt’s Isisaurus post, it’s a big, big group, encompassing many genera and huge morphological range. It’s also a long-lived group, spanning the whole of the Cretaceous; and it’s where most new genera are being named, as Argentina seems to be packed full of ’em.
Well, that’s all for now. Sorry it’s been wordier than usual — probably not much fun to read, but hopefully useful to refer back to in future.
Here’s the famous 8th cervical vertebra of the “Brachiosaurus” brancai lectotype HMN SII, this time in a left-lateral close-up of its left prezygpapophyseal ramus, showing the many pneumatic excavations. Enjoy!
"Brachiosaurus" brancai lectotype HMN SII, 8th cervical vertebra, left prezygapophyseal ramus in left lateral view.
References
- Curry Rogers, Kristina. 2005. The Evolutionary History of the Titanosauria. pp. 50-103 in: K. Curry Rogers and J. A. Wilson (eds.), The Sauropods: Evolution and Paleobiology. University of California Press, Berkeley.
- Harris, Jerald D. 2006. The significance of Suuwassea emiliae (Dinosauria: Sauropoda) for flagellicaudatan intrarelationships and evolution. Journal of Systematic Palaeontology 4: 185-198.
- McIntosh, John S. 1990. Sauropoda. pp. 345-401 in: D. B. Weishampel, P. Dodson and H. Osmólska (eds.), The Dinosauria, 1st edition. University of California Press, Berkeley and Los Angeles.
- Russell, Dale A., and Zheng, Zhong. 1993. A large mamenchisaurid from the Junggar Basin, Xinjiang, China. Canadian Journal of Earth Science 30(10/11): 2082-2095.
- Upchurch, Paul. 1993. The Anatomy, Phylogeny and Systematics of the Sauropod Dinosaurs. University of Cambridge, unpublished Ph.D. dissertation. 489 pp.
- Upchurch, Paul. 1995. The evolutionary history of sauropod dinosaurs. Philosophical Transactions of the Royal Society of London Series B, 349: 365-390.
- Upchurch, Paul. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological Journal of the Linnean Society 124: 43-103.
- Upchurch, Paul, Paul M. Barrett and Peter Dodson. 2004. Sauropoda. pp. 259-322 in D. B. Weishampel, P. Dodson and H. Osmólska (eds.), The Dinosauria, 2nd edition. University of California Press, Berkeley and Los Angeles. 861 pp.
- Wilson, Jeffrey A. 2002. Sauropod dinosaur phylogeny: critique and cladistic analysis. Zoological Journal of the Linnean Society 136: 217-276.
- Wilson, J. A. and Paul C. Sereno. 1998. Early evolution and Higher-level phylogeny of sauropod dinosaurs. Society of Vertebrate Paleontology, Memoir 5: 1-68.
- Wilson, Jeffrey A. and Paul Upchurch. 2009. Redescription and reassessment of the phylogenetic affinities of Euhelopus zdanskyi (Dinosauria – Sauropoda) from the Early Cretaceous of China. Journal of Systematic Palaeontology 7: 199-239. doi:10.1017/S1477201908002691
- Yates, Adam M. 2007. The first complete skull of the Triassic dinosaur Melanorosaurus Haughton (Sauropodomorpha: Anchisauria). pp. 9-55 in: Paul M. Barrett and David J. Batten (eds.), Special Papers in Palaeontology 77: Evolution and Palaeobiology of Early Sauropodomorph Dinosaurs. The Palaeontological Association, U.K.
Sauropod Vertebra Picture of the Week 
August 19, 2009 at 4:53 pm
This is very, very welcome. I was just trying last week to assemble the sense of a picture like this from sources at hand, but was drowning for lack of context.
August 19, 2009 at 5:10 pm
“And once you’ve got a phylogeny, the business of naming the parts of it (and of course choosing which parts to name) is taxonomy.”
Arguably only the “choosing which parts to name” is taxonomy (although taxonomy need not be derived from phylogeny — it’s just much nicer that way). “Naming the parts of it” is nomenclature.
Of course, some people use the term a bit differently. YMMV.
August 19, 2009 at 5:21 pm
“The most basal sauropods include things like Vulcanodon and, it turns out mostly from the work of Adam Yates (e.g. Yates 2007), a whole bunch of things that, if you looked at them you’d probably guess were prosauropods.”
That’s an artifact of how Wilson and Sereno (1998) defined _Sauropoda_: as a branch-based clade with _Saltasaurus_ internal and _Plateosaurus_ external. Salgado (1997) used a node-based definition with _Vulcanodon karibaensis_ and _Eusauropoda_ as specifiers — under that, _Anchisaurus_ et al. would not be sauropods. Ditto for Sereno’s (2005) branch-based redefinition, with external specifiers _Jingshanosaurus xinwaensis_ (!) and _Mussaurus patagonicus_.
Maybe I should save these comments for Tutorial 8, though….
August 19, 2009 at 5:36 pm
Yes, of course it’s true that everything depends on how you define the clades. The in-press chapter on Sauropoda, from the Phylonyms volume, will follow Yates’s (2006) definition as (Saltasaurus loricatus not Melanorosaurus readi), and thereby cut out some of the “prosauropods” that are included in Sauropoda by the definition of Wilson and Sereno. Nevertheless, this definition of Sauropoda includes forms like Blikanasaurus and Antetonitrus, which would appear pretty much prosaurodish to the naked eye.
August 19, 2009 at 6:29 pm
Not a bad middle-of-the-road solution. (I always thought an apomorphy-based definition would probably reflect the traditional usage best — something related to columnar limbs, perhaps — but we all know the problems with that type of definition.)
August 19, 2009 at 10:13 pm
This was a very well done post. For the novice or non-sauropod loving folks, this explains the sauropod family tree very clearly and to the point. Good job!
August 19, 2009 at 10:39 pm
Nice summary of the history of phylogenetic work, too!
August 20, 2009 at 1:11 am
Ugh, cladistics….
I’m with Greg Paul.
August 20, 2009 at 1:14 am
If you want to know why I reject cladistics, it’s basically for the same reasons Greg rejects them in Dinosaurs of the Air.
August 20, 2009 at 7:56 am
Anonymous,
I wasn’t sure whether to let your comments through moderation or not, but decided to be charitable. They are essentially content-free: by “I’m with Greg Paul” you just mean “I agree with Greg Paul’s reasons for rejecting cladistics”; but unless you tell us what those reasons are and why you find them convincing, we’ve learned nothing — only that some anonymous person unknown to us agrees with Greg Paul regarding something unspecified. Please explain what point you were making — and sign your name! Thanks.
August 20, 2009 at 2:37 pm
Basal position of Chinese sauropods outside Neosauropoda may reflect Jurassic Angara isolation.
August 21, 2009 at 7:29 am
JS Lopes said:
“Basal position of Chinese sauropods outside Neosauropoda may reflect Jurassic Angara isolation.”
Or simply the excellent Middle Jurassic record of China. There are a lot of other basal Eusauropods from other parts of the world: Australia, Africa, South America, India and Europe, although many are not as complete as the chinese examples.
August 21, 2009 at 11:23 pm
I look at these diagrams and realise that everything I knew about dinosaurs when I was 10 is wrong.
August 23, 2009 at 9:28 am
Matt,
Why would you say “Sorry it’s been wordier than usual — probably not much fun to read, but hopefully useful to refer back to in future.”?
IMO if folks don’t find it fun to read they should probably be … elsewhere!
On a more serious note, as you have chosen to touch on systematics and presented a phylogeny, I should like to broaden the discussion a little, if you will allow it, and ask for the Team’s collective thoughts regarding the methods employed in phylogenetic analyses such as the ones you mention.
I can appreciate that, in principle, cladistic analysis is the most sensible basis for understanding evolutionary relationships. However, I must say that I struggle with the way it is applied, especially in terms of character definition. Firstly, I have only come across binary characters – seems ludicrous when you choose a character like ‘external mandibular fenestra is greatly reduced’ for Eusauropoda. Secondly, how can you have, as a character, ‘Addition of at least two cervicals to the vertebral column (ten to twelve cervicals)’ – again for Eusauropoda. Isn’t a character supposed to be an observable characteristic of a specific organism?
Thanks.
Neil
August 23, 2009 at 12:09 pm
A rather anglophone review – Salgado, Calvo and Coria were also publishing important phylogenetic papers in 1997…
August 23, 2009 at 8:19 pm
True, Paul; and I could also have mentioned the analysis in Calvo and Salgado’s (1995) description of “Rebbachisaurus” tessonei, or Sanz et al. 1999 and several others (all of them conveniently published in English, too!) But you have to draw the line somewhere, and for a tutorial I figured that less would be more.
August 23, 2009 at 9:45 pm
(re: Neal!)
Observable in a specific organism or a related group of organisms whose ancestors lack that character.
If I’m using a correct analogy (obviously evolution is more complex, but I think it will work all the same!), think of it as locating where you are within a given country–as you move in from the broadest defined classification (the country’s name), you slowly work your way into more separate and distinct internal classifications (state, town, bar, etc.). And at the end of the day, they’re both useful abstractions. ;-)
August 23, 2009 at 10:08 pm
Hello Tor,
I beg to disagree – a character must be observable in a specimen and by inference a specific organism. Doesn’t make any sense otherwise.
Regards.
Neil
August 23, 2009 at 10:27 pm
Well yeah, they’re scored on the individual level (to map its evolutionary location), but individual characters are shared by masses of organisms to suggest that the individual belongs to that more broadly defined clade.
Maybe we’re both saying the same thing and I’m missing it!
August 25, 2009 at 5:43 pm
Thanks for the introduction into sauropod systematic, I remember when I got my first dinosaur book in the 70’s, it had a whopping 2 sauropods, one the infamous swamp creature “brontosaurus”. The way people are digging up new ones, I better study any book I hand out to my kids first so I don’t embarrass myself too much.
September 4, 2009 at 3:00 pm
No, phylogenetics.
No, nomenclature.
If you want to make a difference between systematics and taxonomy, the first is “the art of inserting species into an existing classification”, while the second is “the art of making a classification”, but most people use these terms interchangeably, notably including terminology freaks like Peter Ax (who starts off his books by defining animals philosophically, as a kind of natural bodies…).
If you use phylogenetic nomenclature, and if you use cladistics to generate and test* the phylogenetic hypotheses PN needs, all classification falls flat, and systematics as well as taxonomy become completely superfluous.
Phylogenetics used to be a tool of systematics and taxonomy, but it can be used on its own, and nowadays it more and more often is. We no longer torture ourselves with the question of how we should “translate a tree into a classification”, we just print the tree and (using PN) tie labels to defined places on it.
* Against parsimony. Real falsification isn’t possible in phylogenetics, because there is no perfect character (though SINE/LINE insertions do come very, very, very close).
How is this possible? What improbable sample of the literature is accessible to you? It’s really hard to find a matrix without multistate characters.
That’s just lazy wording (though you’re right that it should be avoided, for the reasons you mention, and for the additional reason that all states of a character must be defined). It means “number of cervicals: 10 (0), 12 or more (1)”.
Finally…
So you reject it because you don’t understand it?
Because that’s what GSP’s argument boils down to. In DA, he goes so far as to almost assemble a data matrix (that long, long table in the appendix) that would certainly be interesting to play with (like: what happens if certain characters are downweighted, what happens if reweighting is used, etc. etc. ad infinitum vel nauseam). And then he just stops and laments about homoplasy. It’s sad.
September 4, 2009 at 3:21 pm
I’m not going to keep going round this one with you, David. It’s apparent that you use these words rather differently from, well, everyone else. But to the world at large, taxonomy is an art and nomenclature is, well, stamp-collecting. Shoving a name on any random node or branch of a tree would be nomenclature; having the taste, experience and insight to know which parts of the tree to name (and what names to use) is taxonomy. If you really want to, you can smush the distinction and redefine the latter to be part of (an expanded understanding of) nomenclature, but that seems like an extreme manoeuvre just to achieve your goal of destroying all taxonomy :-)
September 4, 2009 at 4:22 pm
Taxonomy can be an art and/or a science, depending on your criteria for dividing life into sets (i.e., taxa). (Admittedly, though, at some level small enough, it’s pretty much always an art.)
I don’t really see how nomenclature is like stamp-collecting. It’s more akin to defining the length of a meter or the mass of a gram — arbitrary, yes, but if you can get everyone on board….
September 4, 2009 at 4:28 pm
Yes, those are better metaphors than stamp-collecting. What I was trying to convey is that it is, in itself, a mechanical matter that does not require judgement. Which is why it can be governed by codes (both rank-based and phylogenetic) while taxonomy can not.
September 4, 2009 at 4:30 pm
Well, to be fair taxonomy is more of a distinct science that requires an artistic touch. I’m sure that’s what you meant, but I thought I’d briefly weigh in all the same. ;-)
September 4, 2009 at 4:39 pm
Traditional taxonomy (i.e., Linnaean and rank-based) is more of an art that requires a scientific touch. Phylogenetic taxonomy is almost entirely a science, *except* for the part where you decide what your smallest units are.
Arguably the PhyloCode does govern some aspects of taxonomy. It does not govern what the operational units are (an artificial decision), nor does it govern the structure of phylogeny (a natural pattern), but it does govern how these two things are translated into a taxonomy.
The PhyloCode isn’t alone here, either! The International Code of Virus Classification and Nomenclature does govern taxonomy, although it is the only rank-based code to do so.
September 4, 2009 at 4:42 pm
Err, that should be “will govern”, not “does govern”, of course.
September 4, 2009 at 4:42 pm
If we had defined a clade Xenoposeidonia as the branch-based clade consisting of all organisms more closely related to Xenoposeidon proneneukos than the Diplodocus longus, Brachiosaurus altithorax or Saltasaurus loricatus, that would have been nomenclature. But we refrained from doing so, judging that it was unnecessary: that was taxonomy.
September 4, 2009 at 5:20 pm
Uh, not doing stuff isn’t anything.
September 4, 2009 at 5:27 pm
No indeed; not doing stuff isn’t anything. But deciding not to do stuff is something.
September 4, 2009 at 5:35 pm
Now you just sound avant-garde.
September 4, 2009 at 5:58 pm
And I still think it’s nomenclature. You refrained from defining a name, not from defining a clade (Nature does that). That may seem like a petty difference … and it is. You got me.
September 4, 2009 at 7:43 pm
The more things you decide to refrain from doing, the better the scientist you are.
“Perfection is achieved not when there is nothing left to add, but when there is nothing left to take away.” — Antoine de St. Exupéry.
I was perfect until I posted this.
September 4, 2009 at 8:29 pm
The early history of planetary observation was dominated by people seeing stuff that wasn’t there. Lowell’s Martian canals are the canonical example. A young British amateur astronomer was once congratulated by a prominent older colleauge for his observations of Venus.
“But I hardly ever see anything at all!” he protested.
“That is exactly my point”, said his senior.
By the way, I’m coining Wedel’s Law: as any internet discussion of dinosaur taxonomy grows longer, the probability of someone attempting to define “taxonomy” approaches 1. 1st Corollary: this attempted definition will satisfy no-one, and will completely derail the original conversation. 2nd Corollary: when this happens, the thread becomes less compelling to most readers but hyper-compelling to the original discussants.
Oh, look, there’s the door.
September 4, 2009 at 8:39 pm
Let’s just define it as the stuffing of dead animals’ skins to as to provide the semblance of the animal in life and forget about it. ;P
September 4, 2009 at 9:12 pm
LOL!
(And when I say that, I mean it literally.)
September 4, 2009 at 9:23 pm
Let me interject here a quote from the blog of the redoubtable* Australian Christopher Taylor, “Catalogue of Organisms”, subtitled “An inordinate fondness for systematics”: “… but it pays not to look too carefully at the taxonomy“. Indeed.
(By “redoubtable” I mean to evoke the second definition, “worthy of respect” rather than the first, “causing fear or alarm”.)
September 4, 2009 at 10:03 pm
*Phylogenetic taxonomy… wish posts had an editing feature!
September 6, 2009 at 8:20 am
David Marjanović said:
>I have only come across binary characters
How is this possible? What improbable sample of the literature is accessible to you? It’s really hard to find a matrix without multistate characters.
David, you may well be correct but if I look at the characters given for sauropodomorpha in the Dinosauria none are obviously multistate – could you point me to such a matrix?
>Secondly, how can you have, as a character, ‘Addition of at least two cervicals to the vertebral column (ten to twelve cervicals)’
That’s just lazy wording (though you’re right that it should be avoided, for the reasons you mention, and for the additional reason that all states of a character must be defined). It means “number of cervicals: 10 (0), 12 or more (1)”.
I’m sorry but what you are saying does not change my objection – even with your re-wording 12 or more cervicals is recognized as a character state!
Finally…
If you want to know why I reject cladistics, it’s basically for the same reasons Greg rejects them in Dinosaurs of the Air.
So you reject it because you don’t understand it?
I agree – but then this was not part of MY post :-)
September 6, 2009 at 8:23 am
As a postscript, apologies to David for my tardy response – I looked for additional comments for a few days and then assumed that the thread had died.
Neil
September 11, 2009 at 7:19 am
David Marjanović said:
>I have only come across binary characters
How is this possible? What improbable sample of the literature is accessible to you? It’s really hard to find a matrix without multistate characters.
Forgot to mention that the example I gave “character: ‘external mandibular fenestra is greatly reduced’ for Eusauropoda.” is defined as a binary character.
Regards.
Neil