Intraspecific variation and taxonomy

One aspect of vertebrate palaeontology that often gets overlooked is how much individuals can vary from the ‘type’. Most people focus on interspecific variation – that is, difference between species, but just as important is intraspecific variation – the differences seen within species. In a great many cases (and most especially when it comes to dinosaurs) the fossil record of an animal can be limited to one good specimen with various other bits of others that may be badly distorted, incomplete, broken, or with elements that do not overlap with others. As such you can end up with a rather fixed view of what a taxon looks like if you only take one specimen into consideration (or can only use one).

However, biology deals with a continuum, not absolutes – species are not the same as elements or atoms in their definition or reliability of diagnosis. Individuals vary – you look like your parents and siblings, but not *exactly* like them (even identical twins usually have some distinguishing features). If you look at humans across the earth (admittedly a big sample size and with more variation than many species) you can find individuals of all manner of sizes and body proportions, with no clavicles, extra or fewer fingers and toes, all manner of skin, eye and hair colours (or no hair at all) and more. Some of these are the result of odd occasional defects (like the extra fingers) but can still be hereditary and in large numbers (like missing clavicles) and in both of these cases at least, would of course show up in the fossil record. This is therefore a pretty important consideration when trying to define species and separate out differences between closely related groups.

Obviously we typically lack the sample sizes needed to try and establish this kind of variation in dinosaurs (for example) but we *can* look at patterns we see in extant creatures to see what we might expect in fossils. Does size vary a lot among living populations? Oh yeah! Right, don’t include absolute size as a criterion for defining taxa unless the differences are consistent and very big. Do the number of digits (unless obviously duplicated by a generic error) usually vary? No – so this is a viable indicator of a taxonomic difference. Does the number of teeth vary? In mammals, generally not, but in reptiles, yes (at least a little). So best not use the gain or loss of a single tooth as a factor, but the loss or gain of several is better evidence. And so on, pretty much ad infinitum, though of course just a couple of concepts (relative sizes for example) can cover an awful lot of ground and most of these are pretty obvious.

Sadly, despite the general simplicity of the concept and its application this is not always so vigorously applied. There are plenty of species (and dinosaurs especially) out there that are characterised by being a bit bigger than an otherwise identical taxon, or because they have one more tooth, or one fewer tail vertebrae. A great deal of taxonomic research is, frustratingly and time consuming-ly, a clean-up of old (and even new) bad taxonomic identifications, assignments and descriptions (aside from the normal disagreements). Obviously the field has moved on over the decades (and yes, centuries) but it is amazing how much poor stuff still comes out regularly and how much stuff has never been revised. As I have noted elsewhere, taxonomy is absolutely crucial for basic biology and we are running out of taxonomists and behind on taxonomic progress. I don’t want to be too pessimistic, but the situation is far from bright and the more good taxonomy and less bad that is produced the better off near every field in biology will be.

Share this Post

22 Responses to “Intraspecific variation and taxonomy”

  1. 1 Tor Bertin 31/08/2009 at 12:30 pm

    I was thinking about this exact issue today, comparing what constitutes a taxonomic genus in the fossil record and what we have placed within an individual genus now–coyotes, wolves, and dogs of all forms are all part of the Canis genus; that’s a lot of skeletal proportional variation! I couldn’t help but wonder how we would classify them if we knew nothing about extant dogs and had only their (fragmentary) bones to work with. And further, just how many dinosaurs we have classified as separate genuses that are in fact different species of a single genus (or a different gender within a single species, etc.).

    Considering what we have to work with, the fact that we can (relatively accurately) determine evolutionary pathways at all through the fossil record just blows me mind! It’s fascinating stuff.

    • 2 David Hone 31/08/2009 at 2:44 pm

      Well as I have noted before (somewhere on here) what constitutes a biological species can be very different to working purely under a morphological species concept for an extinct clade. I suspect that the animals you list would all be different genera if handed over to a bunch of dinosaur researchers, but that is not certain – some pretty big size differences can fit into a single genus (like Psittacosaurus for example) it’s complex.

  2. 3 Tor Bertin 31/08/2009 at 12:31 pm

    My mind I mean. I’m not Irish. 😉

  3. 4 Tor Bertin 31/08/2009 at 1:28 pm

    Sorry for all the comments, one last question after a bout of thinking–

    After looking at the skulls of the various species in the Canis genus, I’ve noticed that various proportional differences (like the brain case, directionality of the jaw anterior jaw elements etc.) are generally found between species, but an individual genus is more based upon skeletal elements and characters that exist in one that does not exist in the other.

    So basically, is it fair to say that in most (if not all) cases, proportionality and relative shape is a bad basis for erecting a new genus?

  4. 5 Tor Bertin 31/08/2009 at 2:12 pm

    To clarify before I had to bed–by proportionality I mean very slight differences (for example, the zygomatic in wolves arches higher above the tooth row than it does in coyotes and dingos; there are also variations in curvature of the external portions of the brain case etc.).

    But broadly all the bones are similar in how they are put together and formed–it’s just their rough shapes that differ (while the articulations are near identical).

    I really wish comments had an edit feature so I didn’t have to spam all these. 😉

    • 6 David Hone 31/08/2009 at 3:02 pm

      No problem with the comments! In answer to your question it really depends on what you are doing. If there is a lot of variation in a clade for proportions then that’s a bad choice, if they are conservative then they are a good choice. I don’t know much about canid cranial characters but if people are deliberately not using certain characters there is probably a good reason (if unspoken) for it. Of course it could equally be a ‘traditional’ thing – if the early taxonomy work focused on osteological characters then it is easier to revise and add to these as you define new taxa if they work well than to start again and add loads of new proportional characters.

      Pterosaur proportions in the wing phalanges and legs are generally pretty conservative within groups and vary between them so at higher taxonomic ranks at least they can form a very good basis for diagnosing taxa for example. There’s no problem with using them for genera and species either really provided they are clearly defined and diagnostic.

  5. 8 Michael 01/09/2009 at 2:29 am

    Great post, all very true! But I will add that we can’t go too far with intraspecific variation. Case in point: Allosaurus fragilis and Allosaurus atrox. It should be obvious their differences go too far for them to be the same species, and yet A. atrox is still constantly lumped in with A. fragilis – even though Greg Paul set it straight way back in 1988.

    • 9 David Hone 01/09/2009 at 11:15 am

      But therein lies the problem Michael. Ho do you know that “their differences go too far for them to be the same species”? What are you basing that on? (I’m not challenging you, it’s rhetorical). You need to establish what these patterns are – do you see the same differences repeatedly in the same specimens (i.e. do species of type A always have features 1, 2 and 3 and B have 4, 5 and 6?) or are the mixed (is there one A with 4 and one B with 1 and 2?) and are these ever found together or correlate with something like size (when it might be ontogenetic) or ornamentation (when it might be sexual dimoprhism). Some species are hugely variable (like humans) and taxonomic practice varies from taxon to taxon and researcher to researcher. I have a few congential quirks to my anatomy that I think some people would consider enough to mark me out as a separate species if I were found in the fossil record, but I’m still Homo sapiens (no jokes please! 😉 ).

      Just pointing out differences and calling it taxonomy doesn’t count – you need to try to rule out the other possibilities before naming new things.

      • 10 Michael 02/09/2009 at 2:04 am

        Well, I personally feel splitting A. fragilis into two species is justified because:

        1) We can rule out differences resulting from ontogeny because the two morphs are almost exactly the same size, and 2) We can rule out sexual dimorphism because Morph 1 (True A. fragilis if the species is split up) is much rarer than Morph 2 (the one Paul labeled A. atrox, although tht name seems to be a nomen dubium and I would suggest giving it a new name). The way I see it, the only options left are 1) Allosaurus fragilis is a highly variable species in the manner of Homo sapiens, or 2) we are dealing with two different species. Because there seems to be only two distinct morphs, each of which always seem to have the same recurring sets of characters, I would advocate specific seperation.

        Now, keep in mind I’m no prefessional, just a particularly dedicated amatuer, so there may something I’m missing here.


      • 11 David Hone 02/09/2009 at 8:11 am

        I’m not disagreeing with A. fragilis as such (I really don’t know or care that much about its taxonomy) – the point I was making was a general one. I.e. You can’t just say they look very different and leave it at that (not that you were, it was a useful line to take to make the point).

  6. 12 Michael 01/09/2009 at 9:26 am

    Sorry if that last comment sounded know-it-allish or super authorative, I was in a hurry when I wrote it. 🙂

  7. 13 Mickey Mortimer 02/09/2009 at 11:39 am

    Michael- The Allosaurus fragilis vs. atrox issue as presented by Bakker and Paul was solved by Chure in his thesis. The basic solution in that the single “fragilis”-style skull was reconstructed incorrectly (that’s why it’s so rare!), and that statistical studies have shown supposed differences cited by Paul do not exist (so there are no same recurring sets of characters). In more detail…

    (sorry for the long comment, Dave)

    It has become common (e.g. Paul, 1988; Britt, 1991) to recognize two types of normal-sized Morrison Allosaurus- one with a shorter snout and pointed lacrimal horns (usually called A. fragilis) and another with a long snout and rounded lacrimal horns (usually called A. atrox, and sometimes separated as Creosaurus). While A. fragilis sensu stricto has been based mainly on the A. fragilis topotype, the ‘A. atrox’ morphology has not been based on the Creosaurus atrox holotype. Thus, the long-snouted morphotype will be called ‘A. atrox’ for this section. Regarding the supposedly shorter skull of A. fragilis, Chure (2000) notes the only short skull known is that of USNM 4734, which was found disarticulated. When it was reconstructed by Gilmore (1920), he had to “compromise in regard to the exact articulation of the elements”. There are large plaster filled gaps in the specimen, the contact between the maxilla, jugal and lacrimal is missing, the dentary is from another specimen (USNM 8335), the other mandible is plaster, the palate is fragmentary, and the postorbital regions are distorted judging by their asymmetry. Chure notes the maxilla is reconstructed too far posteriorly, as the lacrimal articulation of the dorsal process is projecting into the antorbital fenestra. The angle between the maxillary body and its dorsal process is similar to other Allosaurus specimens, which wouldn’t make sense if the snout were shorter. Similarily, the angle between the anterior and ventral lacrimal processes is in the middle of the range Allosaurus exhibits, with Cleveland-Lloyd ‘A.atrox’ specimens showing marked variation. The nasal of USNM 4734 is
    broken and the anterior part moved dorsally and rotated ventrally. The lacrimal horn shape shows many intermediates between tall and triangular (USNM 4734) and low and rounded (DINO 2650). There is an example of a triangular lacrimal on a long skull (MOR 693). Contra Paul, triangular lacrimals are known from the Cleveland-Lloyd quarry (eg. UU 40-581). Though Paul claimed ‘A. atrox’ has a more robust neck, therte is no difference when cervical width/length ratios are compared. Similarily, though Paul claimed ‘A. atrox’ has a more robust forelimb, no difference was noted when humeral circumference and length were quantitatively compared (circumference/length ratio .45 in A. fragilis, .36-.49 in ‘A. atrox’). Finally, both “species” are found in the same quarry, as evidenced by AMNH 600 (referred to A. fragilis by Paul) and AMNH 666 (which he referred to ‘A. atrox’). This is contrary to the stratigraphic distinction supported by Bakker and others. In conclusion, there is no evidence for the fragilis/atrox dichotomy advocated by Paul and Bakker. All Allosaurus are long-snouted.

    I know I pointed this out to you on Dinoforum earlier, so please take it to heart this time. Paul was ahead of the curve on some issues, but a lot can change in two decades.

    • 14 David Hone 02/09/2009 at 1:40 pm

      Hi Mickey, thank for all of that, and no problem for the long comment! I do realise that the Musings is going to end up hosting these kinds of debates and discussions between people from time to time, though as a very *general* rule, I really want to engage with people who are not really into their dinosaurs or up on their palaeontology and care more about basic working definitions, ideas and practical science. (That’s a comment to everyone on here, and NOT a criticism of your comment Mickey, which is welcome).

      I’m well aware that people get really into their taxonomy and systematics and if I have raised an issue on here that people want to discuss, that’s fine. However, comment-ers should realise that a) a lot of readers won’t be able to follow it and b) I’m really not that fussed, and I think there are better forums for this kind of thing.

      If you really like your dinosaurs, there is the DML, Vert Pal list, various FaceBook pages and other blogs etc., not to mention Ask A Biologist that allow for that level of detail and you will also find an audience who are more willing to discuss things and contribute their knowledge. I guess you could say I don’t mind this kind of discussion, but I’m not keen to encourage it since I think it’ll turn off a lot of people including me most of the time. I want to push this as a niche *between* the ‘I love dinosaurs and they are just great’ ultra-basics and the ultra-hardcore ‘but character 26 in that analysis was coded as a ? for Dilophosaurus’. I think that gap is largely overlooked and I think it contains a lot of people who want to learn more. All are welcome, but in general try to remember who else might want to read what you are writing and how they might want to discuss it.

      Sorry that was a mammoth comment and really just prompted by the fact that this is the first time this kind of thing has come up in depth, while I do note a general increase in very technical discussions on here which is something I’ve specifically tried to avoid. It therefore seemed a good moment to bring it up and get the idea out there. Do all keep things coming in please, don’t be put off, but try to keep to a middle line as far as possible.

      • 15 Michael Stearns 03/09/2009 at 6:10 am

        I’m sure I fit somewhere in that range you’re targeting, but I gotta tell you, looking at those pictures, I can’t tell the difference at all, but I’m also fascinated and want to know!

        The issues you’ve described in these last few posts have been ones that I’ve been curious about for a long time actually, and I often have a skeptical response whenever a “new” Dinosaur is announced, or even just looking at a lineup of Ceratopsians–I’m thinking how different are these guys REALLY? I don’t think there is much effort out there to explain to the layperson how dinosaurs are different from eachother, I suppose at that level you’re expected to dig through the scientific journals and (if you’re an artist) even do your own reconstructions, but that’s a jump I’m not interested in making.

      • 16 David Hone 03/09/2009 at 8:13 am

        Mike Benton has estimated that on average about 16% of dinosaur genera get synonymised over time (i.e. one is sunk into another) and a huge 30% are invalid as being non-diagnostic.

        That looks really bad, but remember that includes all manner of mad descriptions and names going back to the 1700s. Cope and Marsh between them managed to give one sauropod something like 13 names at one point. In short, these stats look bad but are heavily biased by old practices and are not half as bad now. As such, if someone names a new taxon from decent material in a decent journal, it’s most probably valid. Most people know what they are doing and get proper peer review.

    • 17 Richard G. Baer 16/11/2009 at 7:07 am

      Re: the differences between Allosaurus fragilis and Allosaurus atrox.

      Rather than beg the two-species-vs.-one question at the outset, let me speak of Como Bluff Allosaurus vs. Cleveland-Lloyd Allosaurus. There are at least two discernible cranial differences: (1) the rostrum parasphenoidalis of the Como Bluff Allosaurus is consistently shallower dorsoventrally (i.e., is markedly more “pointed”, cultriform) than that of the Cleveland-Lloyd Allosaurus; (2) the diaphysis of the quadrate is consistently straight in the Como Bluff Allosaurus but consistently sigmoid in the Cleveland-Lloyd Allosaurus. I haven’t assembled enough statistical data to rule out the null hypothesis (viz., that these differences are not genuinely systematic but are entirely within the range of single-species individual variation), but I predict that eventually the data will support a species separation between the Como Bluff Allosaurus (A. fragilis) and the Cleveland-Lloyd Allosaurus (A. atrox).

  1. 1 Ontogeny and taxonomy « Dave Hone’s Archosaur Musings Trackback on 01/09/2009 at 11:10 am
  2. 2 Synonymy « Dave Hone's Archosaur Musings Trackback on 05/10/2010 at 8:39 am
  3. 3 Notes on the taxonomy and identity of Zhuchengtyrannus « Dave Hone's Archosaur Musings Trackback on 01/04/2011 at 9:47 am
  4. 4 Intraspecific variation « Dave Hone's Archosaur Musings Trackback on 01/05/2012 at 9:14 am
  5. 5 Can Olympians teach us about dinosaurs? « News in Briefs Trackback on 10/08/2012 at 2:45 pm
Comments are currently closed.

@Dave_Hone on Twitter


Enter your email address to follow this blog and receive notifications of new posts by email.

Join 567 other followers

%d bloggers like this: