This taxon was requested by Dan Bensen, so here it is......
Liaoxiornis Hou and Chen 1999
= Lingyuanornis Ji and Ji 1999
L. delicatus Hou and Chen 1999
= Lingyuanornis parvus Ji and Ji 1999
Etymology- "small bird from Liaoxi", Liaoxi being Pinyin for the western part of Liaoning Province.
Barremian, Early Cretaceous
Yixian Formation, Liaoning, China
Holotype- (IVPP V 130723, GMV2156 is counterpart and holotype of Lingyuanornis) (82 mm) skull (20.3 mm), mandible (20 mm), eleven cervical vertebrae (15.3 mm), cervical rib, ten dorsal vertebrae (15.3 mm), twenty dorsal ribs, sacrum (11 mm), six caudal vertebrae (4.1 mm), pygostyle (16 mm), scapulae (10.1 mm), coracoids (7.5 mm), furcula (7.6 mm), sternum (3.2 mm), humeri (15.4 mm), radii (14.9 mm), ulnae (15.6 mm), metacarpal II, manual phalanx II-1 (4.3 mm), manual phalanx II-2 (1.3 mm), metacarpal III (7.4 mm), ilia (9-10 mm), pubes (6.6 mm), ischia, femora (14.5 mm), tibiae (16.5, 17.1 mm), metatarsal I, metatarsal II (9.6 mm), metatarsal III (10.4 mm), metatarsal IV (9.6 mm), pedal phalanges and unguals
Diagnosis- anterior portion of premaxilla very narrow; external naris placed posteriorly, extending almost to posterior border of antorbital fenestra; very small antorbital fenestra?; retroarticular process very elongate; very thin spike-like pygostyle; pygostyle longer than femur; no proximodorsal ischial process?.
This specimen was discovered in the late 1990's and described in 1999 as the smallest known Mesozoic bird. The specimen is complete except for some distal phalanges, cervical ribs, manual digit I and fibulae, so the total length can be measured at 82 mm. The authors think Liaoxiornis is an adult because "the skull and postcranial skeleton are well-developed". I feel there are several features suggesting the holotype is a juvenile and will discuss them after the description.
The skull is broadly similar to other basal avians, as it is triangular, with a short pointed snout and large orbits. The sutures between cranial elements are not well defined, which makes detailed description difficult. The premaxilla must be very low, with an especially narrow anterior portion. The external naris is probably placed more posteriorly and is more narrow than either Eoenantiornis or the Spanish hatchling. The maxilla has an elongate process extending beneath the anterior part of the orbit, like Eoenantiornis, but not the Spanish hatchling. Judging by the space between the naris and orbit, the antorbital fenestra was very small, although details cannot be observed. The nasals are slightly convex and shorter than the frontals. The jugal is extremely slender, with what may be a small ascending process. If so, it is more anteriorly placed than in the Spanish hatchling. The frontals are large and bulbous, with a rim above the huge round orbits. Several scleral ossicles are preserved, there were probably about fifteen in life. The parietals are both exposed, one behind the other. They are much shorter than the frontals and round in shape. The posteroventral skull isn't preserved well, so the postorbital, squamosal, quadrate and quadratojugal are not discernable. What may be the foramen magnum is enlarged, being more than a fourth of the posterior skull width. Seven teeth are present in the upper jaw, but the suture between the premaxilla and maxilla is indistinct, so how many were in each bone is not known. These are slightly procumbant and are said to be "similar to those of other early birds". They extend posteriorly only to the anterior edge of the antorbital fenestra. The dentary is very slender and straight, being more slender than Eoenantiornis or the Spanish hatchling. There are seven dentary teeth, less than Eoenantiornis (12) or the Spanish hatchling (8). The surangular is slender and posteriorly expanded, while the angular extends further posteriorly and is not as expended. Between the two is an elongate external mandibular fenestra. There is an extremely long retroarticular process on the articular, although the latter's attachment to the surangular is uncertain.
The number of vertebrae is confusing, as Hou and Chen have a different opinion than Ji and Ji. Hou and Chen estimate more than ten cervicals, eight sacrals and four free caudals. Ji and Ji say there are nine cervicals, eleven dorsals, seven sacrals and seven free caudals. As there are ten dorsal ribs, I think the number of dorsal vertebrae should be ten. This leaves eleven cervicals. Eight vertebrae are between the ilia, Ji and Ji consider the last a caudal. As having eight sacrals would make six free caudals, which is identical to Iberomesornis and Sinornis, I think that is most likely. There are eleven cervical vertebrae, which are a bit wider than long and can be seen to have wide diapophyses. The third to sixth centra are longer than other others. One cervical rib is reportedly preserved. Ten dorsal vertebrae complete the presacral column, grading from 60% wide as they are long to 20% wider than long posteriorly. Hou and Chen state the dorsal centra were round with high neural spines, but as they are in ventral view in their specimen, I don't see how this could be determined. Ten pairs of dorsal ribs are present, the fourth longest. There are no uncinate processes, sternal ribs or gastralia preserved. Eight sacrals attach to the ilia and are unfused. They maintain a constant width, but decrease in length posteriorly. The centra are broad with long transverse processes. The tail is long and six vertebrae are definitely not fused to the pygostyle. At least eight additional vertebrae can be seen making up the proximal third of the pygostyle. All caudals are very wide and have transverse processes. The pygostyle is a simple spike, tapering to a point distally. It is very long, more so than the femur. There is no hint of any chevrons.
The scapulae are very slender, with pointed anteriorly projected acromion processes. The coracoids are strut-like, fairly slender and have convex lateroventral edges. The furcula is V-shaped, with a short hypocleidium. It has an interclavicular angle of fifty-seven degrees. The sternum is single, but very tiny. It is roughly diamond-shaped, with rounded lateral corners, a slightly indented anterior corner and elongate posterior corner. No additional lateral or posterior processes are present. There is a low keel present extending from the anterior to the posterior tip.
The humerofemoral ratio is 106 and the radiohumeral ratio is 97. The humerus is slightly bowed posteriorly, with a prominent internal tuberosity, low deltopecoral crest and a slightly enlarged ectepicondyle. The ulna is slightly longer than the humerus and bowed, while the radius is about 60% of ulnar width. The proximal ends of metacarpals II and III are said to be fused to the semilunate by Hou and Chen, but reported to be separate by Ji and Ji. I cannot identify metacarpal I, but metacarpals II and III are subequally robust. Metacarpal III is bowed and extends past metacarpal II. There are two phalanges on digit II, II-2 much shorter than II-1.
The ilium has a greatly expanded and rounded preacetabular process, with a ventral projection. The postacetabular process is narrow and pointed, without a brevis shelf. It is shorter than the preacetabular process. There appears to be no supracetabular crest. The pubes are fused at the symphysis and are said to lack a foot. The ischium is reported to lack a proximodorsal process.
The femur is gently curved, but further details are lacking. The tibia is also nondescript, but has a fibular crest. It is 115-118 percent of femoral length, while the tarsometatarsus is 72 percent. The fibula is not preserved. A tarsal element is present, though I'm unsure which one. A tarsometatarsus is present (fused proximally) and fairly slender, with parallel sides except for a small proximal expansion. Metatarsal II is shortest, III the longest. Metatarsal IV is said to be slender by Ji and Ji. The pedal phalanges are disarticulated and poorly preserved, but there was a hallux with a large curved ungual.
Hou and Chen don't mention feathers, but Ji and Ji state that primaries are present and at least 2.5 times ulnar length.
As stated above, Hou and Chen consider Liaoxiornis adult because "the skull and postcranial skeleton are well-developed". I disagree based on several features-
- very small size
Not a good indicator of age by itself, but certainly makes young age more probable. Liaoxiornis was 82 mm long, compared to 100 in Longchengornis, 115 in Jibeinia, 119 in Cathayornis? caudatus, 135 in Eoenantiornis, 138 in Sinornis, 140 in Cuspirostrisornis, 145 in Largirostornis and 190-280 in Confuciusornis. Incidentally, Longchengornis and Cathayornis? caudatus both have only partially fused pygostylia as well.
- very large skull
Craniohumeral ratios of enantiornithines are- Liaoxiornis (139), Cathayornis yandica (113), Cathayornis? caudatus (106), Largirostrisornis (106), Eoenantiornis (75), Cuspirostrisornis (~66). Juvenile theropods are known to have relatively large skulls (eg. Scipionyx) and Liaoxiornis' skull is at least 30% larger than other enantiornithines.
- poorly ossified elements
Most of the skull elements are indistinct, there are no sternal ribs, the distal manual and pedal phalanges are not completely preserved and the fibula isn't preserved. The extremely small sternum is also likely caused by incomplete ossification. There are long ventral edges of the coracoids that would normally attach to the front of the sternum, but the sternum is much too small. I think there was cartilage present around the ossified section in life. The poor ossification also means that the absence of uncinate processes cannot be used to prove enantiornithines lack such structures. This may also explain the short hypocleidium and lack of a proximodorsal ischial process.
- lack of fusion
Although the metatarsals, distal pygostyle and possibly carpometacarpus are fused, nothing else is. The sacrals are unfused and vertebrae are still visible for the first third of the pygostyle. Additionally, the tarsal is large enough to be an unfused astragalus.
So, I think the evidence is pretty conclusive that Liaoxiornis is juvenile and not the smallest Mesozoic bird, as has been claimed. An interesting idea would be to compare it with the Spanish hatchling (Sanz et al., 1997) and the new juveniles found in the regurgitated pellet (Sanz et al., 2001).
Liaoxiornis vs. Lingyuanornis-
One month after Liaoxiornis delicatus was described by Hou and Chen, Ji and Ji described Lingyuanornis parvus. Over two years later, Harris discovered they were based on part and counterpart of the same specimen (pers. comm. 2001). Creisler determined Liaoxiornis has priority (http://www.cmnh.org/fun/dinosaur-archive/2001Jun/msg00051.html), but is their synonymy proven? They are determined to be the same specimen based on the following.
- They were both discovered in Lingyuan, Liaoning Province.
- They are both in the exact same position, anterior back and neck curved to the right so that the snout intercepts the right elbow. The long bones are all also at the same angles. Liaoxiornis is is ventral view, and Lingyuanornis is in dorsal.
- They are very similar in size. Humerus- 15.5 vs. 15 mm; ulna- 15.6 vs. 17 mm; metacarpal III 7.4 mm vs. 8 mm; femur- 14.5 vs. 16 mm; tibia- 17.1 vs. 18 mm; tarsometatarsus- 10.4 vs. 11 mm.
- They are very similar in morphology, the differences present probably due to the more schematic illustration of Hou and Chen or the differences in perspective. For instance, Lingyuanornis isn't nearly so perfect looking or symmetrical and the mandible is hidden. A few bones at the extremities are missing in Lingyuanornis (phalanges, distal pygostyle), suggesting either these were restored in Liaoxiornis, or more of the specimen got stuck on the Liaoxiornis side.
Thus, the two are synonyms and should be referred to as Liaoxiornis delecatus.
Hou and Chen refer Liaoxiornis to the Sauriurae based on unfused skull bones, teeth and "sauriurine-like shoulder girdle and pelvis". As the Sauriurae has been universally dismissed in phylogenetic analyses as being a paraphyletic grade of basal avians, Liaoxiornis' relationships must be looked for elsewhere. Ji and Ji refer Lingyuanornis to the Enantiornithes, though the Chinese text prevents me from examining why. This is probable, though it should be noted there is a distinct possibility this clade is paraphyletic. Caution must be used when examining the presence of characters in Liaoxiornis due to its juvenile status. Remember, I use the following phylogeny for basal pygostylians-
Liaoxiornis is more derived than confuciusornithids based on- less than twelve dorsal vertebrae; less than seven free caudals; V-shaped furcula with narrow interclavicular angle (<70 degrees); strut-like coracoid; ulna subequal or longer than humerus; metacarpal III longer than metacarpal II; reduced manual unguals. These characters easily override the plesiomorphically long tail and absent proximodorsal ischial process, to firmly establish Liaoxiornis as a derived pygostylian. It's less derived than Protopteryx based on- short hypocleidium. It's more derived than Protopteryx based on- manual phalanx II-2 shorter than II-1; more than seven sacral vertebrae. It is more derived than Jibeinia based on- less than three phalanges in manual digit II. It is less derived than ornithothoracines based on- interclavicular angle >50 degrees. Potential enantiornithine characters are difficult to identify, as the description is short and so many former enantiornithine characters are now known in more basal pygostylians, as well as basal euornithines (eg. Apsaravis). The coracoid is laterally convex at its ventral end and the fourth metatarsal is reduced in width. This makes it most parsimoniously an enantiornithine, although the large interclavicular angle makes this conclusion somewhat tenuous. Possible interrelationships within that clade are difficult to establish without an accepted phylogeny of enantiornithines. The juvenile status of the holotype also makes this difficult. There is a possibility Liaoxiornis is a juvenile of one of the many better known Liaoning enantiornithines (which are probably oversplit), but such comparisons are beyond the scope of this post. There are several characters distinguishing it from other enantiornithines, although these could change in ontogeny. For now I recommend classifying Liaoxiornis as a provsionally valid, probable enantiornithine.
Anyone who wants a scan of Liaoxiornis' skeleton or a pdf of Lingyuanornis' description, contact me offlist.