Eosinopteryx – part 2 – Better Resolution = Better Reconstruction

Updated October 23, 2015 with new skull data.

Yesterday we looked at Eosinopteryx (Godefroit et al. 2013, Middle-Late Jurassic, Tiaojishan Formation) and discussed a possible new nesting site (Fig. 1) based on a lack of included short-coracoid taxa preceding Archaeopteryx to compare it with. By comparison, Cosesaurus has a “flapping”-type coracoid and it has much less wing tissue trailing its front limbs. So the long, locked-down coracoid in bird predecessors was among the last traits to evolve, post-dating the appearance of elongated forelimb feathers.

Supplementary information

Figure 1. Supplementary information from Godefroit et al. (2013) showing their nesting of Eosinopteryx with Anchiornis, but the tree lacks several short-coracoid taxa that might provide more parsimonious nesting sites.

Addendum: The analysis of Godefroit et al. (2013) was based on and provided only a segment of an earlier analysis that DID include these more primitive taxa. Thus my doubt is reduced considerably as all pertinent taxa were included. The skull on DGS

Not sure if a skull reconstruction was provided by Godefroit et al. (2013) or not. (not) I’d still like to see the paper. (I have the paper now. ) Even so, in order to create yesterday’s reconstruction I applied DGS (digital graphic segregation) to the skull (Fig. 2). I don’t know birds as well as dinosaurs and did not attempt a palate reconstruction. If I made any mistakes, please send me an email. (Figure 2 is an update based on higher resolution images of an earlier posted figure.)

Figure 2. The skull of Eosinopteryx in situ (above) and reconstructed using DGS.

Figure 2. The skull of Eosinopteryx in situ (above) and reconstructed using DGS.

The skull of Eosinopteryx traced using Photoshop, a process known as DGS or Digital Graphic Segregation. For followers of this blog, these updated images reflect the importance of high resolution data in using DGS. Much like the Hale telescope, greater resolution enables the identification of finer lines and bones. This demonstrates that its not the mechanics of the technique so much, as the intimate knowledge long months of study provides when employed, and higher resolution really helps. (Higher resolution image did provide improved data.)

The premaxilla includes at least three elongated teeth (here seen from the inside). The saddle-shaped nasal was broken into at least four pieces during crushing. The maxilla is decayed beyond the normal fenestration. Here (Fig. 3) I reconstruct it conventionally. There is no elongated posterior lacrimal process and vestigial anterior process as Godefroit et al. (2013) reported. Rather the lacrimal is similar to that of other theropods. They did not provide a reconstruction. An earlier reconstruction of mine misplaced the quadrate articulation, which is repaired here. It also did not recognize the posterior mandible, which in situ is separated from the articular area. But here (Fig. 3) this has been repaired, thanks to theropod expert, M. Mortimer, for pointing this out. It was also overlooked by Godefroit et al. (2013, Fig. 2).

Godefroit P, Demuynck H, Dyke G, Hu D, Escuillié FO and Claeys P. 2013. Reduced plumage and flight ability of a new Jurassic paravian theropod from China. Nature Communications 4: 1394. doi:10.1038/ncomms2389
Paul GS 2010. The Princeton Field Guide to Dinosaurs. Princeton University Press 320 pp.


9 thoughts on “Eosinopteryx – part 2 – Better Resolution = Better Reconstruction

  1. Skull is disarticulated, and two halves are skewed relative to one another. Proportions, unlike in pterosaurs, are likely to be skewed as in other birds. Thus, given the uncertain, though stated, ontogenetic status of the specimen, one can not only make erroneous assumptions about the specimen’s valid proportions, so too the shapes are likely erroneous. Your intended shape of the postorbital is ridiculous: you seem to have flattened the aspect of the slab as thouigh it were in lateral view, though the skull is preserved in a form where the dorsal and lateral faces of the skull are on the same plane, and the left side (which you’ve reversed to the right) is displaced rostrally. You do not even attempt to account for this taphonomy, and take the preserved appearance for granted. Both premaxillae are preserved, and visible, but you seem to have taken the more anterior preserved of the pair as the true shape, ignoring the other, which is articulated to the maxilla in more or less its natural position, due to its probable and thus more parsimonious opposition to the mandible, which itself isn’t apparently skewed out of shape as much. Only one maxilla is apparent, and it is the opposite element. Your DGS thingie failed.

  2. Your unfamiliarity with theropods is quite apparent here. Just concerning the premaxillae, what you interpret as one down-angled premaxilla is actually two premaxillae roughly aligned with the maxilla behind them, as stated by Jaime. The bottom one has three preserved teeth, which you have inaccurately angled forward. The expected fourth tooth was probably in a broken off anteroventral corner. What you show as an alveolar edge containing four smaller teeth is actually the posteroventral edge of the subnarial process that would have articulated with the maxilla. Your lateral narial fossa is just a normal area on the medial surface of the left premaxilla. Similarly, on the dentary you completely miss the left dentary preserved above the right one with continuous, closely-spaced teeth. The dentary teeth are not on the outside of the bone. There you were drawing the ones in medial view, where the labial alveolar edge is higher than the lingual edge. The purple stuff below the ectopterygoid body is part of the ectopterygoid body, which has no penetrating fenestra. Your articular is just part of the angular and surangular, with the real articular (or at least the surangular which covers almost all of it laterally) being untraced above the long pink element in back, which is actually the posterior end of the angular. The sclerotic plates above the frontal are just feathers. I could go on, but it’s another point against DGS.

    See the image here http://postimage.org/image/x9zfcqqh3/ with your tracing on the left, the photo in center, and my tracing on the right.

    Anyone familiar with theropods will see your reconstruction and do a double take, because of weird features caused by mistakes like those above. This is one of the reasons I critiqued your reptile analysis and why doing a useful one would be hard- interpretation of features demands familiarity with the group in question.

  3. And that’s why I peppered the post with comments about not being familiar with theropod morphology – and why that’s so important – and it’s not the technique, which you used very well, so much as the familiarity with the subject. So, this is not an indictment of DGS. This is proof that it works, because I gather you have not seen the specimen except in photos also. I am right?

    • Your recon#2 doesn’t address most of the errors I pointed out, and includes new ones like the premaxillary body being shallower than the anterior dentary, and the premaxilla’s subnarial process interrupting the external ascending process of the maxilla (this interruption is probably a fragment of sediment). As for my using DGS, no one’s denying we can trace photos, we’re just saying you do it overzealously- finding structures not indicated by the photo, not accounting for perspective and distortion, and tracing inaccurately if Eosinopteryx is any indication.

  4. As I explained in the text, the resolution is low and the knowledge is low, both keys to making a successful tracing, which is all DGS really is. I thank you for your help. I am not done with this project. The Nature paper just came in and I will be applying that data. So, the only ones who may be overzealous are the ones attacking photographic tracing, which they admit to practicing themselves. I am just a student who is learning as I go. If I have not accounted for perspective and distortion, where is the undistorted data? All of the images I have seen appear to be official undistorted images.

  5. It’s improving, but still has issues. You continue to misunderstand the theropod subnarial process, this time putting two alveoli in it. That surface articulates with the maxilla. It is not part of the alveolar edge. Here’s a color-coded comparison to Utahraptor’s premaxilla- http://postimage.org/image/vcoqgq4ur/ . Neither the purple or pink surfaces could have teeth. You also continue to put the articular part of the mandible too far forward. This time you traced the posterior portion, but ignored it in your reconstruction. See http://postimage.org/image/s0atctnj1/ . There are other issues like your quadrate’s dorsal articulation being in the middle of the squamosal, while in theropods the quadrate head articulates with a cup on the posterior squamosal process. Or the broad surangular-angular articulation in front of the external mandibular fenestra, which never happens in theropods and suggests something is wrong there.

    I think the vast differences between your first and current reconstructions is a perfect illustration of how you are overzealous with tracing. If you were properly cautious, the two would be roughly similar, only with less detail and identified elements in the low res tracing. Instead, bone’s shapes and articulations are so different that the skulls wouldn’t end up close to each other if included in an analysis. Consider how many of your tracings are based on low resolution photos, and how different the high resolution tracings would be if you had them, and how these would change your analysis’ outcome.

    • I will attend to your notes, but to tell me that all my work is suspect is poor detective work and clearly prejudicial. Each image was created individually and will have to be tested individually. Again, it’s my lack of knowledge, not the technique (which you are using very well) that could be the problem.

      That the taxa in my various trees greatly resemble one another (much more so than in traditional trees) falsifies your suggestion, despite the variation in resolution offered. Not saying there aren’t some things that might change with greater resolution. There is no foresight or planning while tracing, yet the parts go together like their sisters do, which supports the hypothesis by its consistency. It would be better if you put your magnifying glass on the bizarre reconstructions, like Bennett’s little monster, or Unwin’s uropatagium, that have no match to morphologies in putative sisters than to continually call my body of work into question. Specifics, yes. The entire body, no.

  6. Pingback: DinoAstur - » Un nuevo dinosaurio emplumado del Jurásico: Eosinopteryx

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s