19 aprile 2017

L’ischio di Aurornis ricade nella variabilità di Anchiornis? [Aggiornamento]

In un commento al precedente post, Mickey Mortimer ha suggerito che la diversità esistente tra l’ischio di Aurornis rispetto a quella osservata negli esemplari di Anchiornis non sia sufficiente per considerare Aurornis un taxon distinto.
Se questa ipotesi è corretta, un’analisi della diversità morfologica tra l’ischio di Aurornis e quella degli esemplari di Anchiornis dovrebbe collocare il primo dentro la distribuzione del secondo.
Ho testato questa ipotesi usando 8 caratteri dell’ischio e mappando la loro variazione in un campione di 5 Anchiornis pubblicati rispetto alle morfologie di Aurornis, Eosinopteryx, Archaeopteryx e Caudipteryx.

Profilo laterale dell'ischio di alcuni paraviani giurassici. I pallinii colorati indicano i principali landmark dell'osso.

Innanzitutto, è bene stabilire alcuni landmark per gli ischi in vista laterale: la regione acetabolare (estremità prossimale dell’ischio), l’apice distale, il processo otturatore e quando presente, il processo dorsale. Inoltre, ho definito l’ampiezza dell’incisura distale sottesa dal processo otturatore e dall’estremità distale.
Gli otto caratteri sono i seguenti:
[1] 'Dorsal process: present vs absent.
[2] 'Distal tip: blunt vs acuminate.
[3] 'Obturator process: blunt vs acuminate.
[4] 'Distal tip directed: ventrally vs dorsally.
[5] 'Obturator notch angle: less than 90° vs wider.
[6] 'Distal tip elongation: shorter vs longer than obturator notch.
[7] 'Dorsal process distal position: proximal vs distal to obturator process.
[8] 'Obturator process proportions: deeper vs narrower than elongation of process at base.

Ho svolto una analisi delle componenti principali della diversità di questi ischi nel campione.
Limitandosi al grafico delle due componenti principali, risulta che Eosinopteryx cade dentro la distribuzione dei 5 ischi di Anchiornis (4 punti visualizzati, perché due esemplari si sovrappongono), mentre Caudipteryx, Archaeopteryx e Aurornis cadono ognuno ben al di fuori della ellisse che definisce l’intervallo di confidenza stimato per Anchiornis, ed ognuno ricade in un proprio quadrante.

Questo risultato è abbastanza chiaro: la differenza tra l’ischio di Aurornis e quelli di Anchiornis è comparabile a quella tra Anchiornis, Caudipteryx e Archaeopteryx. Quindi, l’ischio di Aurornis non può essere considerato una variazione intraspecifica di Anchiornis.
Al contrario, Eosinopteryx cade dentro la distribuzione di Anchiornis, e potrebbe essere un esemplare di quel taxon (oppure, che i due taxa non siano distinguibili nell’ischio): questo conferma ulteriormente la validità di Aurornis come taxon distinto.

Aggiornamento del 20 Aprile 2017
Istigato da alcuni commenti, ho incluso altri esemplari nell'analisi: due Archaeopteryx.
Il risultato conferma il precedente grafico: Eosinopteryx risulta dentro la variabilità di Anchiornis, ed Aurornis è molto al di fuori. Notare l'ampia variabilità di Archaeopteryx che comunque non include Aurornis.

23 commenti:

  1. Interesting, though I have to wonder if the distal tip in Aurornis is just not simply there due to preservation or perhaps poor ossification and what you have mapped as distal tip in Aurornis in this analysis of yours is just the distal end point of the distal notch.

    1. The ventrodistal curvature describing a narrow notch is in both ischia, so I assume it is not a preservation artifact.

    2. indeed. hmm. Whatabout the ossification thing. The photo of the ischium in description paper seem to shw some pitting though it's really hard to say anything about it since the photo is so small. Is there anything in the bone surface that would indicate that the ischium wansn't fully formed? Striations, pitting etc.

      Man how I wish it was a standard operating procedure that there would always be supplementary folder with full high resolution versions of the original photos used in the paper. Why this isn't the case is beyond me.

    3. Bone texture in Aurornis dismisses an immature condition. The immature condition is evident in Eosinopteryx, that shows the juvenile rugosities. Also, the individual is in the same range size of Anchiornis specimens: such a difference cannot be ontogenetic.
      Part of the ischia are probably under slab, but this cannot alter significantly the shape compared to Anchiornis ischia.

    4. Well, size is not really good indicator of ontogenetic maturity. But even dismissing that as it is all we currently have at ahnd to judge these things Aurornis is actually very much mid sized individual with largest individual currently known being PKUP V1068 with femur length around 90 mm. Asuming it represents the adult of the species this means that Aurornis with 66 mm femur still has good deal of growing to do.

    5. What I meant is that nor size neither ontogenetic-related features visible in the specimen support the an ontogenetic-based explanation for the ischium shape.

    6. PS: my hope is one day to publish a detailed description of the specimen, with hi-res photos. It is a project for the future, but unfortunately I have other projects first with priority. So many good fossils and so few time available...

  2. It shows the ischium of Aurornis is the most divergent of described specimens, which I don't argue against. Would you mind adding the other Archaeopteryx specimens to the graph?

    1. It is my plan to add other paravians, including other Archaeopteryx.

    2. And now that you've added more Archaeopteryx specimens, I can make my point. You could just as well argue that the top-most Archaeopteryx is a different taxon than the bottom two, but (see The Theropod Database for details) there is no consistent variability between the specimens and thus no objective way to divide them into species. It's the same for the Microraptor complex. So yes, Aurornis has a more different ischium than any (other?) Anchiornis specimen, but that doesn't mean it's a different species. I just don't think morphometry is a tool that can solve this problem.

    3. This is not morphometry, this is morphology: look at the characters used in the list. You dismiss morphological disparity because it dismisses your idea that Aurornis is an Anchiornis.
      I have shown that such is not valid, because if you assume that Aurornis ischium is another Anchiornis ischium, then we should conclude that also Archaeopteryx (and Caudipteryx) are among Anchiornis disparity... which is not a valid explanation of the data.
      I am not stating that every species must have a comparable amount of diversity, but suggest that the ischium of Aurornis is too different for being considered among Anchiornis variation. The graphs shows that among-Anchiornis variation is ditinct from the (larger) among-Archaeopteryx vaiation, and that Aurornis is outside both clusters. This is confirmed by the graph: any other Anchiornis is closer to each other than Aurornis, and all other "genus-level" taxa are closer to the Anchiornis cluster than Aurornis.
      It seems to me that you simply follow your a priori idea that Anchiornis is so variable that everything could be considered an Anchiornis.
      Then, if Aurornis is so distant from other Anchiornis (and the Anchiornis are so closely placed in the graph), what is your explanation for such disparity? Note that every Archaeopteryx results more similar to Anchiornis than Aurornis, and also note that no overlap is between Anchiornis and Archaeopteryx clusters: these are clearly taxonomic distinctions.
      If Aurornis is Anchiornis, please provide explanation of what biological criteriom supports such referral (actual data, not the mere idea that "Anchiornis is super-variable", which is just a circular reasoning avoiding to consider the features used above to estimate disparity, and that shows that Aurornis is too different from Anchiornis to be considered inside that taxon).

    4. First, I don't have an opinion of whether Aurornis is Anchiornis since I haven't examined the issue myself. This has merely been me saying IF this is the only difference, I'm skeptical.

      But note that Aurornis' ischium is basically the same as Eosinopteryx's, just with a decurved and expanded tip. And if you only decurve the tip (character 4), that forces the obturator process to be narrower (5) and provides a dorsal angle that can be interpreted as a dorsal process (1). So that one change is tripled in your scorings, which then separate the taxa by who knows how much more distance in your graph. We also have theropods that exhibit this exact interspecific variation- Tyrannosaurus CM 9380 has an expanded ischium and AMNH 5027 has a decurved ischium, while RTMP 81.6.1 has neither. So there's precedent.

      But even if there wasn't precedent and the characters weren't correlated, the ischium is just one element. I don't think there's anything wrong with one element falling far outside the normal range of variation for a species in certain individuals. Caudipteryx's ischium may be more similar to Anchiornis' given these characters, but Aurornis' entire anatomy is far closer to Anchiornis'.

      Two final points. First is that Aurornis' ischia are asymmetrical (wider, more square obturator notch with more anterior expansion of the tip in the right element) and subjectively look more deformed than natural. So I could easily imagine some pathology being to blame. The second is that we have hundreds of Anchiornis from those sediments but only one Aurornis. Could be coincidence, or maybe Aurornis didn't live close to the lakes normally, but the hypothesis Aurornis' type was an abnormal Anchiornis individual also fits that data pretty well.

    5. I mistyped and meant to say " that forces the obturator NOTCH to be narrower (5)."

    6. The ischium shows a blunt apex (acuminate in all Anchiornis), a dorsal process at level of obturator process (absent in Anchiornis), a rounded notch distal to obturator process (absent in Anchiornis). These may be pathologies, individual variation or valid autapomorphies.
      Given that my phylogenetic analysis, using all anchiornis-grade forms (including unpublished specimens) does not cluster Aurornis among the Anchiornis specimens node, I consider this taxon valid.
      The left ischium is partially covered by sediment: take this into account before stating abnormalities.
      The pathology hypothesis is not plausible given that the rest of the skeleton lacks pathologies, and the two ischia shows the same shape in the comparable elements.
      The ischial analysis is not an attempt to define the species, but to show that that bone is much different than other Anchiornis ischia, supporting a taxonomic distinction.

      Last, but not least, there are other features differentiating Aurornis from Anchiornis (in particular, in the dentition), but I cannot say anything about them now, bacause they are part of a paper in preparation.

    7. As I said, the blunt apex and ventral curvature (which forms both the dorsal process and rounded obturator notch) are known to vary in e.g. Tyrannosaurus.

      The left ischium being covered in sediment cannot explain why its notch is narrower than the right ischium. Sediment would create the opposite illusion. I'm no pathologist, but why would other elements also be expected to be pathological? Surely there could be something that affected just the distal ischia. Similarly, since both ischia are controlled by the same genes AND physically adjacent to each other, both developmental and physical pathology could be expected to affect both ischia.

      Of course if there are other characters that divide the taxa, such as those in the Megamatrix or your unpublished dental characters, they could affect my conclusions. I'm just saying the ischia alone don't do it for me. In the Lori trees btw, Anchiornis, Aurornis, Eosinopteryx and Pedopenna all form a clade, which looks suspicious to a lumper like myself...

    8. "which looks suspicious to a lumper like myself..."

      So, being a lumper, I was right in writing that you a priori assume they are the same taxon.
      These forms form a clade in almost every analysis including them, called Anchiornithidae/inae (Pedopenna among scansoriopterygids in the recent TWiG iterations is not plausible, and even the Xu team agree Pedopenna is an anchiornithid).

      Again, if you call that clade "Anchiornis" (Pedopenna has priority) or Anchiornithinae is just personal taxonomy, which is not my interest to discuss.
      I have looked carefully to Aurornis, Eosinopteryx and other Anchiornis-grade forms. They differ each other in a series of features in dentition, manual proportions and robustness, ilium and ischium shape and pedal proportions. Again, we can be overlumpers and call them with a single name, but such assumption requires an explanation for such variation, not the mere statement that this is intraspecific variability.
      Among these, Eosinopteryx is clearly juvenile, and I may accept it being an Anchiornis. There are also other anchiornithines suggesting a Pedopenna-Aurornis clade. In the latter, Aurornis is not a Pedopenna because pedal toe 1 differs markedly in proportions and lacks Pedopenna autapomorphic slender P1-I (and I agree that metatarsal I is not elongate as we stated in 2013: after re-preparation, I have re-analysed Aurornis and confirms that its mt I is similar to other paravians).

    9. I don't a priori assume, but the cases I've examined in depth (Archaeopteryx, Microraptor, Sapeornis, Shenzhouraptor) have the same result. Either every well described specimen is a new species, or there was interspecific variability. So when we have the same pattern again in anchiornithines, I'd of course be suspicious. Note I think Xiaotingia is quite different though, so it's not like I just lump all basal paravians from that formation together. I haven't looked into the anchiornithines very much though, but will also note Archaeopteryx specimens also differ in all of those areas you mention. So if you support lithographica vs. siemensii vs. grandis etc. then it's understandable you'd support Aurornis vs. Anchiornis.

      My explanation for lumping is that every single Mesozoic theropod known from a lot of specimens (Coelophysis, Allosaurus, Tyrannosaurus, Deinonychus, Microraptor, Archaeopteryx, etc.) shows a lot of variability that doesn't vary consistently between specimens. In cases like Tyrannosaurus, the lit seems fine keep it one species, but in cases like Archaeopteryx, the trend is to recognize multiple species based on a subjective portion of characters. I'm just trying to be consistent.

    10. There are two main problems with these multiple specimens (not just anchiornithines): one is stratigraphic. These Chinese biota are far from being constrained stratigraphically, and this means we cannot determine how such variation is related to age differences. It seems that some people consider all these fossils as coeval, and thus referrable to a single population... something actually not demonstrated.
      The other is ontogenetic: we need to compare morphological differences with ontogenetic stages (Nanotyrannus docet).
      Something on anchiornithine histology is under study.

  3. l'impressione meramente qualitativa che avevo avuto osservando le immagini precedenti era che le differenze fossero un po' troppo marcate per cadere nell'intraspecifico (a meno di non supporre artefatti di conservazione o patologie - cose che non sono minimamente in grado di valutare) - davvero interessante vedere come una simile comparazione possa essere condotta con metodi quantitativi (a prescindere dal fatto che la mia impressione sembri confermata - lo ritengo quasi un caso)

  4. What functional reason could explain the "unusual" form of the distal tip in the Aurornis ischium? For example, based on shared anatomy, it seems that its functional may be consistent (and conserved) in Anchiornis, Archaeopteryx, and various other early paravians.

    1. I have considered the reason for the unusual shape in Aurornis.
      It is interesting to note that these basal paravians all share the shortest ischia among dinosaurs. I suspect this is linked to caudal musculature reduction prior to the evolution of the bird-like locomotory module.
      Thus, I suspect changes in ischial shape (and development of various ischial processes) among these paravians is related to muscular development.
      Note that Aurornis is also unique among these basal paravians for the larger and rectangular postacetabular process. Furthermore, Aurornis shows a distinct and rugose supratrochanteric process along dorsal margin of ilium.

      My preliminary hypothesis is that these features are linked to locomotory (and adaptive) differences compared to other paravians.

  5. David Marjanović22/4/17 10:58

    di 5 Anchiornis pubblicati

    That's the real problem, isn't it? Well over 200 individuals are known, almost nothing has been published.

    Are there any ischia accidentally shown in the sternum paper?

    1. And I am not sure all those 200 individuals are referrable to Anchiornis (this is also the opinion of Ulysse Lefevre, who personally visited one of the largest Anchiornis collections in China: there is a significant variation among the anchiornithid-grade specimens).

      Sternum paper?