Xenoposeidon week, day 6: so what is a “family” anyway?
November 21, 2007
OK, so it’s actually day 7: I missed my deadline yesterday due to that unfortunate necessity, the day-job, which had me in meetings for half of the day and travelling for the other half. Yes, I could have written this post on the trains and planes, but I had my reasons.
So here we are, at last. Today I’d like to talk a little bit about this idea that Xenoposeidon represents a new “family” of sauropods, which is just a little controversial. However, since this is supposed to be Picture of the Day, let’s start with a pretty picture which is not really related. For some dumb reason, the Xeno paper illustrates the specimen in left and right lateral and anterior and posterior views, but not dorsal view: so here it is!
If I had more time, I’d prepare an interpretive drawing of this, but since I don’t, let me draw your attention to a few points. All of this will probably make more sense if you look at this picture together with the left-lateral view, and its corresponding drawing, from Day 1.
Anterior is towards the top of the picture, so the flatness you see at the bottom is the cotyle in dorsal view. In reality, of course, the cotyle is gently convex, but all we can see from this angle is the dorsal margin. You’ll see that a pair of symmetrical buttresses run up towards breakages, also symmetrical. The buttresses are the pedicels of the neural arch, and the breakage is part way up the centropostzygapophyseal laminae. We can see that those laminae extend anteriorly until they meet the accessory postzygapophyseal laminae (as we interpreted them), which in dorsal view extend almost directly laterally. Obviously that lamina is only well preserved on the left side; anterior to that is “accessory infraparapophyseal lamina”, which we think is homologous with the widespread posterior centroparapophyseal lamina (PCPL) but which we didn’t feel we could call by that name as it doesn’t get anywhere near the centrum. As we look at it from above, this lamina appears to run anterolaterally, but of course it’s also ascending as it progresses anteriorly, as you can see in lateral view: so its true orientation is anterodorsolateral.
Isn’t this fun?
Well, anyway. The most anterolateral point of that lamina, which in this dorsal view looks like a blunt triangle sticking out to the left, is the site we interpreted as the parapophysis: that is, the more ventral of the two sites where the rib head articulated with the vertebra. The other of these sites, the diapophysis, seems to have been pretty much directly above the parapophysis … and is of course lost, like far too much of the vertebra: see the speculative reconstruction from Day 2. The position of the diapophysis is actually easier to make out in this dorsal view than in lateral view, in which the paradiapophyseal lamina (PPDL) appears to be oriented posterodorsally. One of our reviewers, seeing this apparent trajectory of the PPDL, questioned our interpretation of the lamina, suggesting instead that the diapophysis might be located some way posteriorly (as well as dorsally) of the parapophysis, and that the accessory poztzygapophyseal lamina might instead be an anterior centrodiapophyseal lamina. (He was quite right to raise that point, as the versions of the figures that we originally submitted didn’t point out all the relevant features that support our interpretation.) Anyway, in this dorsal view we can see that the PPDL is a sheet of bone projecting anterolaterally from the body of the neural arch and running directly dorsoventrally (i.e. into the eye of the camera), hence the directly-dorsal position of the diapophysis.
Sorry if all that was a bit dull — just wanted to clear it up before launching into the ‘What is a “family” anyway?’ discussion.
Here’s what we said in the paper on that subject, in the conclusion to the Comparisons and Interpretation section:
While X. proneneukos is clearly a neosauropod, it cannot be referred to any existing neosauropod genus, nor even to any `family’-level or `superfamily’-level group, a conclusion first reached by means of group-by-group comparisons and then verified by the phylogenetic analysis. Its unique characters indicate that it is either a highly derived member of one of the known groups, or, more likely, the first representative of a previously unknown group. While we consider this specimen to represent a new `family’-level clade, raising a new monogeneric family name would be premature; and the indeterminate position of the new genus within Neosauropoda means that no useful phylogenetic definition could be formulated.
Why did we put the scare-quotes around the word “family”? Because the notion of of a “family” in taxonomy is at best a rather slippery one. We felt it was necessary to draw attention to just how weird Xeno is compared with other sauropods, and how very uncomfortably it sits in any of the available groups, and it seemed to us that to talk in terms of a “family-level” distinction was the best way to do that: or, as I said in one of the many TV interviews, Xeno is about as different from other sauropods as bears are from dogs or cats. Of course, a statement like that is necessarily pretty fluffy, and better suited to the News At Ten than to a technical paper, but it does get the message across.
The problem is that as soon as you ask exactly what a “family” is, you start to realise that the answer is “a group that has been designated a family”. There is, and can be, no objective standard for how broad a grouping should be designated a “family”, or even of how such broadness should be assessed — by number of species, morphological disparity, whatever. We felt that the best way to look at this in the Xeno paper was to look at the existing sauropod families (i.e. group names that end with the convenional ending “-idae”) and we concluded that Xeno appears as different from, say, brachiosaurids and diplodocids as they are from each other.
Some people think that we should have avoided using the term “family” at all, notably Randy Irmis who I am as I write this in the middle of an email argument with. Actually, I have plenty of sympathy with that approach: the problem is that the suggestions of what to say instead are much worse than the disease. A common suggestion is that we say Xeno represents a new clade: but that is trivially true of every single organism. Randy Irmis, in fact, represents many new clades, just one of which is the clade of all animals more closely related to Randy Irmis than to Buzz Aldrin. A statement like that is true but contains no information. While the word “family” indicates only a fuzzy idea of morphological disparity, it does at least convey some idea, which is more than you can say for “clade” or “group”. The trick is to avoid being fooled into thinking that “new family” means something more precise than it does.
Finally, let’s remember that, assuming we’re correct in saying that Xeno is a neosauropod (and I’m sure we are), then it must belong to either Diplodocoidae or Macronaria, by the very definition of those names. And if it’s macronarian, then it must be either a basal member of that group of a camarasauromorph; and so on, as clades slide up the phylogeny. So in saying “new family”, we certainly don’t mean that it couldn’t be, for example, a bizarre deviant titanosaur or brachiosaurid: just that, even if it is one of those things, it’s weird enough to merit recognition.
I’m sure plenty of people will disagree with the approach we took; that’s fine, if we’d taken a different approach, plenty of people would have disagreed with that, instead. The use of ranks, even informally, just is controversial, there’s no getting away from it. Maybe in another decade we’ll be closer to a consensus — but I wouldn’t want to guess at this remove what that consensus will be.
Well, that’s plenty for today. Sorry if it’s been a bit hardcore in places: Matt will bring you a helping of your usual wackiness tomorrow, on the seventh and final day of Xenoposeidon Week. Thanks for sticking with it!