[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Hongshanornis, "Archaeorhynchus" and ornithuromorph phylogeny



With the recent publication of Hongshanornis by Zhou and Zhang (2005), we are provided with more information about basal ornithuromorphs. They coded it and an unpublished taxon, "Archaeorhynchus", in the matrix of Clarke (2004) (NOT Norell and Clarke, 2001 as they claim). In addition, they added Yanornis, Yixianornis, Jeholornis (from Zhou and Zhang, 2002 presumably), Sapeornis (partially from Zhou and Zhang, 2002, but with new info from Zhou and Zhang, 2003), Liaoningornis and claim to have added Confuciusornis, though Norell and Clarke already included it. They provide the following cladogram-

|--Dromaeosauridae
`--+--Archaeopteryx
  `--+--Shenzhouraptor
     `--+--Sapeornis
        `--+--Confuciusornis
           `--+--Vorona
              `--+--+--Sinornis
                 |  `--+--Neuquenornis
                 |     `--+--Gobipteryx
                 |        `--Concornis
                 `--+--Hongshanornis
                    `--+--+--Liaoningornis
                       |  `--+--Yanornis
                       |     `--Yixianornis
                       `--+--Apsaravis
                          `--+--+--Baptornis
                             |  `--Hesperornis
                             `--+--Ichthyornis
                                `--+--Lithornis
                                   |--Crypturellus
                                   `--+--+--Chauna
                                      |  `--Anas
                                      `--+--Gallus
                                         `--Crax

However, after running their matrix (with or without characters 89 and 97 excluded, as they are supposed to be), I fail to retrieve their tree. Instead- enantiornithines are a polytomy; Confuciusornis, Vorona and Enantiornithes are in a poytomy with Euornithes; and the Euornithes base is a polytomy between Liaoningornis, Hongshanornis, Yanornithidae and Apsaravis+more derived birds. When I run Zhou and Zhang's complete matrix, including Patagopteryx and "Archaeorhynchus", the result is-

|--Dromaeosauridae
`--+--Archaeopteryx
  `--+--Shenzhouraptor
     `--+--Sapeornis
        `--+--+--Confuciusornis
           |  `--+--Vorona
           |     |--Sinornis
           |     |--Concornis
           |     |--Neuquenornis
           |     `--Gobipteryx
           `--+--Patagopteryx
              `--+--Liaoningornis
                 |--Hongshanornis
                 |--"Archaeorhynchus"
                 `--+--Yanornis
                    `--+--Yixianornis
                       `--+--Apsaravis
                          `--+--+--Baptornis
                             |  `--Hesperornis
                             `--+--Ichthyornis
                                `--+--+--Lithornis
                                   |  `--Crypturellus
                                   `--+--+--Chauna
                                      |  `--Anas
                                      `--+--Gallus
                                         `--Crax

As the matrix was taken from Clarke (2004), which was in turn from Clarke (2002), I decided to add the taxa Clarke (2002) did to test their positions. Upon doing so, I realized the Liaoningornis coded by Clarke (2002) and the one coded by Zhou and Zhang differed immensely. The resulting strict consensus cladogram is shown below-

|--Dromaeosauridae
`--+--Archaeopteryx
  `--+--Shenzhouraptor
     `--+--Sapeornis
        `--+--+--Confuciusornis
           |  |*-Chaoyangia
           |  `--+--Sinornis
           |     |--Concornis
           |     |--Neuquenornis
           |     |--Gobipteryx
           |     `--+--Vorona
           |        `--Liaonigornis sensu Clarke
           `--+--Patagopteryx
              `--+--Hongshanornis
                 |*-Liaoningornis sensu Zhou and Zhang (not Yanornis+)
                 `--+--"Archaeorhynchus"
                    `--+--Yanornis
                       |*-Guildavis (not a neornithine)
                       `--+--Yixianornis
                          |*-Gansus (not Iaceornis+)
                          `--+--Apsaravis
                             |*-Ambiortus (not Hesperornithes+)
                             `--+--Songlingornis
                                `--+--+--Baptornis
                                   |  `--Hesperornis
                                   `--+--Ichthyornis
                                      `--+--Limenavis
                                         |*-Apatornis (not a neornithine)
                                         `--+--Iaceornis
                                            `--+--+--Lithornis
                                               |  `--Crypturellus
                                               `--+--+--Chauna
                                                  |  `--Anas
                                                  `--+--Gallus
                                                     `--Crax

The first odd thing about the tree is that a Sauriurae of sorts is formed by Confuciusornis, Chaoyangia and enantiornithines (including Vorona). It's supported by- dorsal centra much longer than wide; deltopectoral crest projected dorsally; deltopectoral crest subequal to shaft in width; proximoposterior surface of deltopectoral crest concave; m. humerotricipitalis groove on humerus; brachial scar on ulna; anterior surface of metacarpal I broadly convex; proximodorsal ischial process contacts ilium; enlarged posterior trochanter; fibular trochlea laterally projected as small notch; tuberositas retinaculiextensoris present on proximal tarsus; one proximal vascular foramen on tarsometatarsus; m. tibialis cranialis tubercle on metatarsal II. Enforcing ornithothoracine monophyly is just one step longer. When this is done, Vorona can be a confuciusornithid, enantiornithine or euornithine basal to Patagopteryx.

Liaoningornis can have two positions, depending on whose codings I use.
As coded by Clarke, it's an enantiornithine sister to Vorona. This is based on- sternal keel projected (also in ornithuromorphs; coded absent by Zhou and Zhang); reduced interclavicular angle (also in ornithuromorphs; coded ? by Zhou and Zhang); elongate hyocleidium (coded ? by Zhou and Zhang); unfused scapulocoracoid (also in ornithuromorphs; coded ? by Zhou and Zhang); radius with deep longitudinal groove (also in yanornithids); distal metatarsal fusion (also in ornithuromorphs); ginglymoid metatarsal II (coded absent by Zhou and Zhang; also in Patagopteryx).
As coded by Zhou and Zhang, it's a basal ornithuromorph, in a polytomy with Hongshanornis, "Archaeorhynchus" and more derived birds. This is based on- sternal keel extends anteriorly (coded absent by Clark; also present in Neuquenornis); radius with deep longitudinal groove (only shared with yanornithids; also in enantiornithines); medial cnemial crest (coded absent by Clarke); narrow intercondylar groove on astragalus (coded ? by Clarke; only shared with Apsaravis and Yanornis; also in enantiornithines); distal metatarsal fusion (also in Vorona); metatarsal V absent (also in some enantiornithines; coded ? by Clarke); hypotarsus present (coded ? by Clarke).
So its position in Clarke's version is primarily due to the elongate hypocleidium which was uncoded by Zhou and Zhang (interpreted as the anterior sternal keel?), and the ginglymoid metatarsal II which was coded differently by Zhou and Zhang (stated and illustrated as present in original description). While its position in Zhou and Zhang's version is due to the anteriorly extended sternal keel, medial cnemial crest and hypotarsus, which were coded as absent (said to be poorly preserved, and seemingly in part interpreted as a hypocleidium), absent (the original desiption notes a small bone similar to the lateral and medial cnemial crests) and unknown respectively by Clarke. Thus it's really a matter of coding disagreements. When Liaoningornis is excluded, the only difference is that a polytomy forms between Confuciusornis, Vorona, enantiornithines and ornithuromorphs. Examination of the trees indicates both Sauriurae and Ornithothoraces are possible. Making a new Liaoningornis OTU using only the codings which agree between authors results in it being able to go anywhere inside Pygostylia (which is unresolved like when Liaoningornis is excluded) basal to Apsaravis. Using Clarke's codings (which I subjectively trust more) and constraining Ornithothoraces, Liaoningornis can be an enantiornithine or euornithine less derived than Hongshanornis. I myself find its sternal shape to be intriguingly similar to Eoalulavis, arguing for the enantiornithine identification.


Hongshanornis is resolved as an ornithuromorph more derived than Patagopteryx, so is not the most basal known, contra Zhou and Zhang. Perhaps that's why they excluded it from their analysis. This is based on- jugal-postorbital contact absent (unknown in Patagopteryx); shortened pygostyle (unknown in Patagopteryx); ossified uncinate processes (unknown in Patagopteryx); procoracoid process; humeral head convex in anterior view; midline of humeral head projected furthest proximally (unknown in Patagopteryx, and quite possibly correlated with the previous character); manual phalanx II-1 strongly compressed; proximodorsal ischial process absent; tibiotarsal condyles subequal in transverse width.

"Archaeorhynchus" is more derived than Hongshanornis based on- pneumatic articular (unknown in Hongshanornis); reduced caudal zygapophyses (unknown in Hongshanornis); sternal carina extends anteriorly (unknown in Hongshanornis); hypocleidium absent; lateral coracoid process; ulnohumeral ratio ~100; metacarpal III reduced in width compared to metacarpal II; medial cnemial crest (unknown in Hongshanornis); tarsometatarsal proximal vascular foramina absent (unknown in Hongshanornis; only shared with Yanornis and Yixianornis). Some local autapomorphies of the taxon include small nares, a noncondylar quadrate-quadratojugal articulation, undivided quadrate head, amphicoelous cervicals, and only seven sacrals. It's also toothless, presumably where the name originates.

Yanornis is the next most derived taxon, based on- premaxillary, maxillary and dentary teeth (near certainly primitive, but optimized as derived in this tree); elongate posterodorsal dentary process; dorsal centra longer than wide; dorsal pleurofossae (unknown in "Archaeorhynchus" or Hongshanornis, probably symplesiomorphic for pygostylians); nine or more sacral vertebrae (also in Patagopteryx, unknown in Hongshanornis, probably symplesiomorphic for ornithuromorphs); posterior coracoid surface lacks fossa (unknown in more basal ornithuromorphs, absent in Apsaravis); acrocoracoid process hooked medially (absent in Apsaravis); transverse groove on humerus (absent in "Archaeorhynchus", but present in enantiornithines, so possibly an ornithothoracine symplesiomorphy); distal humerus compressed anterposteriorly (absent in Patagopteryx, but present in enantiornithines, so possibly an ornithothoracine symplesiomorphy); deep longitudinal groove in radius (only shared with Yixianornis); V-shaped ulnare (absent in Apsaravis); carpometacarpus fused distally (absent in Hongshanornis, but present in Patagopteryx, so possibly an ornithuromorph symplesiomorphy); antitrochanter posterodorsal to acetabulum (unknown in Hongshanornis or "Archaeorhynchus"); distal vascular foramen of metatarsus closed (absent in "Archaeorhynchus", but present in Patagopteryx, so possibly an ornithuromorph symplesiomorphy); proximal metatarsal III displaced plantarily.

Guildavis is at least as derived as Yanornis, but is outside Neornithes. This is a wider possible distribution than given by Clarke (2004), because basal ornithuromorphs were more poorly known then.

Next is Yixianornis, which does not come out in Yanornithidae as advocated by Clarke et al. (2002) and Zhou and Zhang. This is based on- scapula subequal or longer than humerus; extensor process present; pisiform process present; pubic foot absent (also in Patagopteryx, though not Yanornis or "Archaeorhynchus", so maybe an ornithuromorph symplesiomorphy); distal tibiotarsus with well developed posteroproximally extending crests (unknown in more basal ornithuromorphs); metatarsal II shorter than IV (also in Hongshanornis; absent in Apsaravis).

Gansus is at least as derived as Yixianornis, but is not a member of the Iaceornis + Neornithes clade. Again, this is a wider range than Clarke (2004) suggested, due to basal ornithuromorph feet being better known now.

Apsaravis is more derived than previous taxa based on- completely heterocoelous cervical vertebrae; ten or less dorsal vertebrae; at least ten sacral vertebrae; deltopectoral crest projected dorsally (only shared with Ambiortus and Ichthyornis); pit on bicipital crest (absent in Yanornis, but also in enantiornithines, so possibly an ornithothoracine symplesiomorphy); m. humerotricipitalis groove (also in Confuciusornis and enantiornithines); ulnar brachial scar (also in Confuciusornis and enantiornithines); radius with ventroposterior muscle impression (also in "Archaeorhynchus", but not Yanornis or Yixianornis); metacarpal I distal articulation non-ginglymoid (also in Yanornis, but not Yixianornis); intermetacarpal process at least present as scar; obturator flange (also in Patagopteryx); pubis compressed mediolaterally in section; pubic symphysis absent (also in Patagopteryx); patellar groove present; tuberositas retinaculiextensoris present on proximal tarsus; proximal vascular foramen in tarsometatarsus present (also in Patagopteryx); m. tibialis cranialis tubercle on lateral surface of metatarsal II.

Ambiortus is at least as derived as Apsaravis, but is not as derived as Hesperornithes. This agrees with Clarke (2004).

Songlingornis is weakly placed as more derived than Apsaravis based on its lack of a fused dentary symphysis (which like toothlessness, is probably optimized incorrectly in this tree).

Finally, sixteen characters support Hesperornithes + Carinatae, thirteen characters support Carinatae, two place Limenavis more derived than Ichthyornis, five place Iaceornis more derived than Limenavis, and twenty-seven support Neornithes. Apatornis is more derived than Ichthyornis, but is not a neornithine.

Overall, this is the best look at basal ornithuromorph phylogeny yet. It's complicated by the fact Yixian taxa are more basal than Patagopteryx and Apsaravis in some ways, yet more derived in others. It's almost as if Yixian taxa were taking a separate path to get to the ornithurine (sensu Chiappe) level of specialization. It's almost not surprising Sauriurae forms when you consider how primitive Hongshanornis and "Archaeorhynchus" are in many ways. And what does my analysis say? The 90% majority rule consensus (excluding more basal taxa) is-

|--+--Concornis
|  |--Enantiornis
|  `--+--+--Eoenantiornis
|     |  `--LP-4450-IEI
|     `--+--+--Eocathayornis
|        |  `--+--Aberratiodontus
|        |     `--Jibeinia
|        `--+--Sinornis
|           `--+--Vescornis
|              `--+--Longirostravis
|                 `--+--Longipteryx
|                    `--Protopteryx
`--+--+--Eoalulavis
  |  `--Liaoningornis
  `--+--Neuquenornis
     `--+--Iberomesornis
        `--+--Patagopteryx
           `--+--Songlingornis
              `--+--Yanornis
                 `--+--+--Yixianornis
                    |  `--+--Apsaravis
                    |     `--Ambiortus
                    `--+--Boluochia
                       `--+--Ichthyornis
                          `--+--Hesperornithes
                             `--+--Iaceornis
                                `--+--Hongshanornis
                                   `--"Archaeorhynchus"

Here Liaoningornis clades with Eoalulavis, though it should be noted I haven't incorporated Zhou and Zhang's codings for the former. My tree shares (Patagopteryx (Yanornis (Yixianornis, Apsaravis, Ambiortus (Hesperornithes, Ichthyornis, Iaceornis)))) with the results above. It differs in the placement of Songlingornis, but we saw that was weakly supported above. The placement of Apsaravis and Ambiortus sister to Yixianornis is a small shift in position, and my tree changes the interrationships of Hesperornis, Ichthyornis and Iaceornis. The largest change, however, is the placement of Hongshanornis and "Archaeorhynchus" as sister to Iaceornis. This is based on the shortened sternum and bowed third metacarpal (both at least in Hongshanornis), while placing that clade sister to Hesperornithes is based on the elongate and deep subnarial process, posteriorly displaced external nares, and absent dorsal jugal process (again, all are only known for Hongshanornis). None of these characters were used by Zhou and Zhang. In fact, our character lists are largely mutually exclusive, mine having many more cranial, axial and manual characters, while Clarke's has many forelimb and hindlimb characters. Hongshanornis and "Archaeorhynchus are sister taxa based on their small external nares, toothless maxilla and dentary, amphicoelous cervicals and absent extensor process on metacarpal I. These are reversals besides the toothless elements, which combined with the primitive local autapomorphies of the taxa (noncondylar quadrate-quadratojugal articulation and seven sacrals in "Archaeorhynchus"; two phalanges on manual digit III in Hongshanornis) suggest they are more basal taxa which have developed derived characters in parallel to the lineage leading to recent birds. I predict adding more characters to my analysis will demonstrate this for Hongshanornis, though I hesitate to say much regarding "Archaeorhynchus" until we have more than a set of codings to go on.

Mickey Mortimer