Caudipteryx Ji, Currie, Norell and Ji 1998
Diagnosis- premaxillary teeth limited to rostral half; first premaxillary tooth much larger than others; single maxillary fenestra present; twenty-two caudal vertebrae; sternal plates oval; only two phalanges present on manual digit III; metatarsal III extremely narrow proximally, but not arctometatarsalian.
C. zoui Ji, Currie, Norell and Ji 1998
?= Caudipteryx dongi Zhou and Wang 2000
Etymology- "Zou's tail feather", referring to Zou Jiahua, vice-premier of China and an avid supporter of the scientific work in Liaoning.
Barremian, Early Cretaceous
Yixian Formation, Liaoning, China
Holotype- (NGMC 97-4-A) (890 mm) skull (76 mm), lower jaws, cervical vertebae, cervical ribs, dorsal vertebrae, dorsal ribs, gastralia, twenty-two caudal vertebrae, chevrons, coracoid, sternal plate (36 mm), humeri (69 mm), radii, ulnae, mani, ilia, pubes, ischia (77 mm), femora (147 mm), tibiae (188 mm), astragali, calcanea, metatarsi (115 mm), pedes, body feathers, retrices, remiges, gastroliths
Paratype- (NGMC 97-9-A) (725 mm) incomplete skull (79 mm), lower jaws, hyoid, cervical vertebrae, cervical ribs, dorsal vertebrae, dorsal ribs, sacrum, caudal vertebrae, chevrons, scapulae, coracoids, sternal plates, sternal ribs, humeri (70 mm), radii, ulnae, incomplete mani, partial pubes, femora (149 mm), tibiae (182 mm), astragali, calcanea, metatarsi (117 mm), pedes, body feathers, retrices, remiges, gastroliths
Referred- (IVPP V 12344, holotype of Caudipteryx dongi) (896 mm) frontal, pterygoid, two cervical vertebrae, six dorsal vertebrae, dorsal ribs (100 mm), three uncinate processes (30 mm), gastralia, sacrum, eleven caudal vertebrae, chevrons, partial coracoid, sternal plates (25 mm), sternal ribs (35 mm), incomplete humeri (~73 mm), radii (~58 mm), ulnae (61 mm), semilunate carpal, radiale, ulnare, metacarpal I (13 mm), phalanx I-1 (25 mm), manual ungual I (15 mm), metacarpal II (29 mm), phalanx II-1 (18.5 mm), phalanx II-2 (25 mm), manual ungual II (18 mm), metacarpal III (27 mm), ilia (115 mm), pubes, ischia (73 mm), femora (146, 152 mm), tibiae (196 mm), fibula (181 mm), astragali, calcaneum, distal tarsal III, distal tarsal IV, metatarsal I (19 mm), phalanx I-1 (12 mm), pedal ungual I (11 mm), metatarsal II (112 mm), phalanx II-1 (25 mm), phalanx II-2 (16 mm), pedal ungual II (19 mm), metatarsal III (124 mm), phalanx III-1 (27 mm), phalanx III-2 (20 mm), phalanx III-3 (17 mm), pedal ungual III (20 mm), metatarsal IV (116 mm), phalanx IV-1 (15 mm), phalanx IV-2 (9 mm), phalanx IV-3 (7 mm), phalanx IV-4 (7 mm), pedal ungual IV (16 mm), metatarsal V (~36 mm), body feathers, remiges (182 mm), gastroliths
Diagnosis- premaxilla sharply pointed anteriorly; quadratojugal posterior process slightly developed; manual ungual II larger than manual ungual I.
C. sp. nov.*
Barremian, Early Cretaceous
Yixian Formation, Liaoning, China
* see discussion
Specimens- (BPM 0001, referred to Caudipteryx zoui) (852 mm) skull, lower jaws, twelve cervical vertebrae, cervical ribs, nine dorsal vertebrae, dorsal ribs (~114 mm), uncinate processes, gastralia, sacrum, twenty-two caudal vertebrae, chevrons, scapulae (80 mm), coracoids (34 mm), sternal plates (~30 mm), sternal ribs (~38 mm), humeri (72 mm), radii (59 mm), ulnae (62 mm), semilunate carpal, radiale, ulnare, metacarpal I (11 mm), phalanx I-1 (25 mm), manual ungual I (16 mm), metacarpal II (28 mm), phalanx II-1 (17 mm), phalanx II-2 (24 mm), manual ungual II (15 mm), metacarpal III (25 mm), phalanx III-1, phalanx III-2, ilia (115 mm), pubes (~124 mm), ischia (72 mm), femora (145 mm), tibiae (188 mm), fibulae (188 mm), astragali, calcanea, distal tarsal III, distal tarsal IV, metatarsal I (16 mm), phalanx I-1 (13 mm), pedal ungual I (12 mm), metatarsal II (102 mm), phalanx II-1 (23 mm), phalanx II-2 (16 mm), pedal ungual II (19 mm), metatarsal III (113 mm), phalanx III-1 (24 mm), phalanx III-2 (19 mm), phalanx III-3 (15 mm), pedal ungual III (18 mm), metatarsal IV (107 mm), phalanx IV-1 (14 mm), phalanx IV-2 (8 mm), phalanx IV-3 (6 mm), phalanx IV-4 (4 mm), pedal ungual IV (14 mm), metatarsal V (30 mm), body feathers, remiges, retrices, gastroliths
(IVPP V 12340, referred to Caudipteryx sp.) (836 mm) skull, lower jaws, twelve cervical vertebrae, cervical ribs, nine dorsal vertebrae, dorsal ribs (120 mm), uncinate processes (26 mm), gastralia, sacrum, twenty-two caudal vertebrae, chevrons, scapulae (80 mm), coracoids (35 mm), sternal ribs (36 mm), humeri (69 mm), radii (~56 mm), ulnae (61 mm), semilunate carpal, radiale, ulnare, metacarpal I (11 mm), phalanx I-1 (26 mm), manual ungual I (16 mm), metacarpal II (28 mm), phalanx II-1 (19 mm), phalanx II-2 (24 mm), manual ungual II (15 mm), metacarpal III (23 mm), phalanx III-1, phalanx III-2, ilia (112 mm), pubes (~125 mm), ischia (~72 mm), femora (145 mm), tibiae (183 mm), fibulae (175 mm), astragali, calcanea, distal tarsal III, distal tarsal IV, metatarsal I (15 mm), phalanx I-1 (12 mm), pedal ungual I (12 mm), metatarsal II (102 mm), phalanx II-1 (22 mm), phalanx II-2 (14 mm), pedal ungual II (17 mm), metatarsal III (112 mm), phalanx III-1 (23 mm), phalanx III-2 (17 mm), phalanx III-3 (13 mm), pedal ungual III (18 mm), metatarsal IV (106 mm), phalanx IV-1 (12 mm), phalanx IV-2 (8 mm), phalanx IV-3 (6 mm), phalanx IV-4 (5 mm), pedal ungual IV, metatarsal V (31 mm), body feathers, remiges, retrices, gastroliths
Diagnosis- large premaxillary subnarial process; maxilla extended anteriorly with promaxillary fossa; external naris close to antorbital fenestra in length; jugal strongly concave posterodorsally; posterodorsal dentary process subequal in width to posteroventral process; posterodorsal dentary process longer than posteroventral process; no intramandibular joint; vomers do not extend past external nares?; ectopterygoid very thin and C-shaped; twelve cervical vertebrae?; ventral margin of coracoid irregular?; anterior margin of preacetabular process posteroventrally oriented.
The various specimens referred to Caudipteryx show quite a bit of variation. They will be described together and the implications for the number of species and individual variation will be discussed afterward.
Skulls are illustrated for BPM 0001, IVPP V 12430 and NGMC 97-9-A. A complete skull is shown in the skeletal drawing of NGMC 97-4-A, but not illustrated in detail. A frontal and possible pterygoid are known in IVPP V 12344, but not illustrated in detail either. The skull is fairly short, with a snout taking up about half the length. Contra Zhou et al., the external naris is not larger than the antorbital fenestra once the specimens are articulated. It is much smaller in NGMC 97-9-A, but only slightly smaller in BPM 0001. It extends to the rostral border of the antorbital fenestra. The maxilla has a very small contribution to the external naris. The orbits are large and round, while the laterotemporal fenestrae are shorter and dorsoventrally elongate. The latter are broadest in IVPP 12430 and smallest in NGMC 97-9-A. The premaxilla is gently rounded anterodorsally, with a slightly convex ventral margin. There is an elongate narial fossa anteroventral to the external naris. The premaxilla is lower and more triangular in NGMC 97-9-A, with a narrower external naris, more vertical maxillary suture and shorter, narrower subnarial process. The naris is widest in BPM 0001, while the subnarial process is largest in IVPP V 12430. The ventral edge appears to have three notches in BPM 0001, while it is smooth in IVPP V 12430 and NGMC 97-9-A. The premaxillae are unfused and contain teeth in the anterior half. There are four procumbant teeth, the first is by far the largest. Teeth are serrationless and constricted at the base. The maxilla is reduced, with narrow ascending and posterior processes. The latter is much shorter in IVPP 12430 and thicker in BPM 0001. A single maxillary fenestra is present; there does not seem to be much of an antorbital fossa. BPM 0001 and IVPP V 12430 have an elongate anterior portion with a pneumatic fossa, while NGMC 97-9-A is blunt anteriorly without accessory fossae. Maxillae are toothless. The nasal is shorter or subequal to the frontal in length. The nasals lack rugosities and are unfused. The lacrimal is triradiate, with an elongate posterior process, suggesting the prefrontal is fused to it. While the processes are 120 degrees apart in IVPP V 12430, the anterior and posterior processes are 160 degrees apart in BPM 0001. There is a large pneumatic lacrimal foramen, but no rugosities or horns are evident. The jugal is relatively low, although not rod-like. The anterior process is more elongate in IVPP V 12430 and NGMC 97-9-A. The dorsal process is much more robust in BPM 0001 and IVPP V 12430 than in NGMC 97-9-A , but contacted the postorbital in all of them. The posterior process is shortest and not visibly bifurcated. The frontal is roughly triangular, with an elongate anterior section seemingly overlapped by the nasal. The orbital rim is raised and ventral impressions indicate a large brain. The frontals seem fused in BPM 0001, but separate in NGMC 97-9-A. The frontonasal suture appears anteriorly convex. The triradiate postorbital is much larger in BPM 0001 and IVPP V 12430 than in NGMC 97-9-A. The anterior process is expanded in the first two, while it is shorter and tapered in the latter. The posterior process is broad in the former two, but shorter and tapered in the latter. The ventral process is much more slender and elongate in NGMC 97-9-A, compared to the other two specimens. Because of these differences, I believe the unlabeled element preserved directly anterior to the detached ventral postorbital process in NGMC 97-9-A is actually the main postorbital body. It is much larger and more similar in shape to the postorbitals of BPM 0001 and IVPP V 12430, but the possibility remains it is a lacrimal or jugal. The supposed postorbital then may be the other quadratojugal. The parietal is poorly preserved in all specimens, but appears subequal to the frontal in length and quadrangular. The squamosal is also poorly preserved, but has an elongate tapered ventral process and a hooked posterior process that exposes the quadrate head laterally. The ventral process contacted the quadratojugals dorsal process. The quadratojugal is not triradiate and has a dorsal process more elongate than the anterior process. Both processes are narrow and tapered. The quadrate is single-headed, with a gently concave posterior margin and a deep notch ventrally. It is vertical and not pneumatic. The braincase is visible in BPM 0001 and NGMC 97-9-A, but shows no details besides several large foramina. The broad posterior portion of the pterygoid is preserved in BPM 0001, contacting the quadrate. Another similar element is unidentified in IVPP V 12430. An L-shaped element, tapered on one end and slightly expanded on the other, is identified as a possible pterygoid in IVPP V 12430. I can't see how this can be a pterygoid (although my grasp of three-dimensional palates is poor) and think it resembles a quadratojugal more. A large irregular element in BPM 0001 is identified as a palatine. This does not resemble any palatine I have seen, but is not triradiate. In BPM 0001, there are two pointed elements projecting posterodorsally from the anteroventral margin of the external naris that Zhou et al. identify as vomers. If that is true, they are very short compared to other theropods, as they extend only halfway past the external nares. A thin, C-shaped element in IVPP V 12430 is identified as an ectopterygoid. This is dissimilar from both the dumb-bell shaped ectopterygoids of oviraptorids or the hook-shaped ones of most theropods, although it is most similar to the latter. The ectopterygoid of NGMC 97-9-A is more robust, with the usual thickened portion seen in most theropods. Several other cranial elements are also difficult to identify. Two elongate elements preserved in the snout of IVPP V 12430 could be vomers or vomeral processes of the palatines. A vertical strap-like element, wider than the lacrimal and found in the antorbital fenestra, defies identification but is present in all three specimens. A small element in the naris of BPM 0001 is very similar to a coronoid, but its position makes this identification suspect. Curiously, two scleral plates are preserved in NGMC 97-9-A, but not in other specimens.
The dentary is toothless with two long posterior processes. The ventral process is longer than the main dentary body in BPM 0001 and IVPP 12430, but shorter in NGMC 97-9-A. The processes are subequal in length in the first two, but the dorsal process is half as long in the latter. The dorsal process is much narrower than the ventral in NGMC 97-9-A, but wider in BPM 0001 and IVPP V 12430. Anteroventrally, the dentary is concave. Medially, a shallow mackelian groove seems to be present in NGMC 97-9-A and BPM 0001. The symphysis is well-developed, but not fused. The external mandibular fenestra is 39% in BPM 0001 and 33% in IVPP V 12430. The surangular and dentary may be fused in these specimens, but are loosely joined in the former. Ji et al. (1998) claim the intramandibular joint in NGMC 97-9-A was mobile, but this could not have been the case in BPM 0001 and IVPP V 12430. A ventral surangular process crosses the external mandibular fenestra in BPM 0001 and extends partway into it in NGMC 97-9-A, but not in IVPP V 12430. If a surangular foramen was present, it was very small. The angular is large, unlike oviraptorids, and connot be distinguished from the articular. The mandibular joint is not extremely convex, unlike caenagnathoids. The retroarticular process is moderately elongate, narrower in NGMC 97-9-A. An elongate bone ventral to the surangular (and fused angular?) in IVPP V 12430 is probably the other surangular/angular in medial view, based on a posteroventral tubercle also seen in the surangular/angular of BPM 0001. A splenial is identifiable in IVPP V 12430, it is acutely triangular with a notched posterior edge. There is no indication it could be seen in lateral view. Another much larger element is identified as the splenial in BPM 0001, but appears to be a straight narrow prearticular instead. The prearticular of IVPP V 12430 is probably visible behind the dorsal dentary process. Another elongate element in BPM 0001 is more problematic. It is a bit larger than the splenial should be, triangular and has a notch posterodorsally. Perhaps it is an oddly shaped splenial. A slender, tapered hyoid is seen in NGMC 97-9-A.
An element labeled "?" by Zhou et al. located behind the skull of IVPP V 12430 looks like an atlantal neurapophysis to me. It is triangular, with a small process on one corner. Ten cervicals are reported by Ji et al., while Zhou et al. estimate twelve. They are amphicoelous and have slender unfused ribs. The axial neural spine is prominent and expanded, the neural spine of the third cervical is tall and rectangular. More posterior cervical neural spines are low. Prominent cervicodorsal hypapophyses are said to be present. Zhou et al. report nine dorsal vertebrae. They are procoelous and reported to lack deep pleurocoels. Currie (pers. comm. 2000) states pleurocoels are present anteriorly, but not posteriorly. Details are hard to make out, but the dorsals appear longer than Nomingia, with slightly shorter quadrangular neural spines. Nine pairs of dorsal ribs are present, the fourth the longest. Three uncinate processes are preserved in IVPP V 12344, four in BPM 0001 and six in IVPP V 12430. On the latter specimen, they are present on the first six dorsal ribs. The second, third and fourth processes are longest. They are flat, slightly curved and expanded ventrally. Gastralia are also present. There are five sacral vertebrae, unfused in IVPP V 12344. The tail contains twenty-two vertebrae, none are fused into a pygostyle. The centra lack pleurocoels and decrease in length posteriorly. Twelve to fifteen have transverse processes, the last seven have elongate prezygopophyses. It appears the centra may be grooved ventrally and are not procoelous. The last chevron is after the seventeenth caudal. Dorsoventrally elongate chevrons are present until after the tenth caudal, all but the first are distally expanded. No dromaeosaur-like highly elongated prezygopophyses or chevrons are present.
The scapula is gently curved and gradually expanded distally to 2.4 times minimum shaft width. The acromion forms a prominent, though broad, anteriorly projecting process. The scapulocoracoid suture is broad, unfused in NGMC 97-9-A and IVPP V 12430, but fused in BPM 0001. The coracoid is subrectangular and taller than long, though not as much as in dromaeosaurids. A prominent triangular posteroventral process is present. There is a large coracoid tubercle and a foramen located near the scapulocoracoid suture. The glenoid seems to point mostly posteroventrally. The furcula is broad (interclavicular angle ~90) and U-shaped, probably without a hypocleidium. Two oval sternal plates are present, smaller than the coracoids. There is no keel or lateral processes. Several sternal ribs are preserved, they are straight, flatter than dorsal ribs and longer than the sternum.
The humerus is poorly described, but lacks a pneumatic fossa and a well-developed olecranal fossa. It is relatively straight, slender and has a low deltopectoral crest. The radius is slnder (~60-80% of ulnar width) and the ulna is bowed posteriorly. The carpus consists of a large semilunate, triangular radiale and small rounded ulnare. There are three unfused metacarpals. The first is 39-45% the length of the second. It lacks an extensor process. Phalanx I-1 is longest, so the first digit reaches well past metacarpal II. The first manual ungual is moderately curved with a well-developed flexor tubercle, but lacks a proximodorsal lip. It is slightly larger than the second manual ungual in BPM 0001 and IVPP V 12430, but smaller in NGMC 97-4-A and IVPP V 12344. Phalanx II-2 is longer than phalanx II-1. The second manual ungual is similar to the first, but has a lower flexor tubercle. Metacarpal III is very slender, straight and slightly shorter than metacarpal II (82-93%). There are only two phalanges in manual digit III. They are both very small, the second smallest, and do not reach past the midpoint of phalanx II-1. There is no ungual.
The pelvis is propubic and unfused. The ilium has a long ventrally expanded preacetabular process and rounded, posteroventrally sloped postacetabular process. The preacetabular process is 15-29% longer than the postacetabular process. The anterior margin is anterodorsally oriented in IVPP 12344, but posteroventrally oriented in IVPP V 12430. The anterodorsal margin is higher than the posterodorsal margin and is rounded anteriorly. The pubic peduncle reaches further ventrally than the ischial peduncle does. It is slightly anteroventrally oriented and notched ventrally, although this is not seen in lateral view. The m. cuppedicus fossa is shallow and reduced. The pubis is nearly straight and has an anterior foot much larger than the posterior. The pubic symphysis extends about halfway up the shaft, which is not compressed mediolaterally. The ischium is 58% of pubic length. It lacks any posterior processes, but does have a large triangular obturator process placed 60% down the shaft. It is concave posteriorly.
The greater and lesser trochantors are separated by a small notch, while the greater trochantor is well separated from the head. There is no distinct fossa for the capital ligament, and no transverse ridge bounding the popliteal fossa. The tibia is anteroposteriorly elongate when viewed proximally. The fibula is very slender, but extends distally to contact the calcaneum. The astragalus and calcaneum are not fused to the tibia, and are separate from each other as well. The ascending process of the astragalus extends 22% up the tibia and is broad and triangular. There is a shallow groove or fossa separating the process from the condyles. Two unfused distal tarsals are present. The unfused metatarsus is elongate and although the third metatarsal is very narrow, it is not arctometatarsalian. Rather, it expands proximally after reaching it's narrowest point at midlength. The first metatarsal is placed two thirds down on the posteromedial surface of the second metatarsal and has a ball-shaped distal end. It is at least partially reversed. The phalanx is short and stout, while the small ungual is more curved than the others. The second metatarsal is 90-91% the length of the second, the fourth is 94-95%. The second digit shows no predatory specializations- there is no proximoventral heel on phalanx II-2 and the ungual is subequal in size to the others. Unlike eumaniraptorans, the first phalanx of digit II is less than 90% of phalanx III-1. The fifth metatarsal is 27-29% of the third in length.
The body was covered with small plumulaceous feathers up to 14 mm long. At least fourteen remiges are present on metacarpal II, phalanx II-1 and phalanx II-2. They lengthen proximally (30, 63.5 and 95 mm long starting with most distal), are symmetrical and have well-developed rachis and vanes. Barbules seem not to have been present. Eleven retrices are present on each side of the tail. They are attached to the last six caudals. These are also symmetrical. Small rounded gastroliths are preserved in all specimens, measuring up to 4.5 mm in diameter, although most are less than 4 mm.
Comparison of specimens-
Five specimens of Caudipteryx have been described. NGMC 97-4-A, NGMC 97-9-A and BPM 0001 are referred to the type species, C. zoui. IVPP V 12344 was referred to a new species, C. dongi. IVPP V 12430 was referred simply to C. sp.. Zhou and Wang differetiated C. dongi from C. zoui based on the smaller sternum and longer first metacarpal. Most differences I can see between the specimens are cranial, although this may be due to the fact the skulls are well illustrated, while the postcrania is not. Are these differences real or preservational? A large amount of the variety seems to be due to crushing and distortion. For instance, there is no way the lacrimal of IVPP V 12430 could have had such a small angle between its anterior and posterior processes in life. Similarily, the posterior postorbital process of BPM 0001 is much too long, as it would extend well past the quadrate when articulated. The anterior squamosal process of that specimen is much too large and bulbous, as it would reach through the postorbital and into the orbit. More evidence that distortion has occured might come from the asymmetry in specimens. The dentaries visible in medial view is always deeper and more decurved, with a more pronounced chin, than the dentaries visible in lateral view. This is very unusual and I lack a good explanation. Perhaps Caudipteryx had an asymmetrical lower beak in life (like crossbills- Loxia), although the larger side is the left in two specimens and the right in the other. I provisionally consider this oddity due to distortion, although the other possibility is quite intriguing. The dorsal cranial elements (nasal, frontal, parietal) are often distorted and asymmetrical. The differences least likely to be due to distortion or individual variation support BPM 0001 and IVPP V 12430 being separate from NGMC 97-9-A. Characters these two specimens share not found in the latter are- premaxilla blunt anteriorly; large premaxillary subnarial process; maxilla extended anteriorly with promaxillary fossa; external naris close to antorbital fenestra in length; jugal strongly concave posterodorsally; quadratojugal posterior process not developed; posterodorsal dentary process subequal in width to posteroventral process; posterodorsal dentary process longer than posteroventral process; no intramandibular joint. Most of the postcranium is not figured in sufficient detail to determine morphological differences in specimens. The differing number of reported cervical vertebrae might be due to misinterpretation, as Zhou et al. state "there are estimated twelve cervical vertebrae". The coracoid has a smoothly rounded ventral border in NGMC 97-9-A, unlike the irregular border of BPM 0001, although the significance of this is uncertain. The orientation of the anterior preacetabular edge differs in IVPP V 12344 and IVPP V 12430, but as the skull of the former is fragmentary, it cannot be determined if this is correlated with the cranial differences noted above. Contra Zhou and Wang, no significant differences in postcranial ratios is evident. Most ratios vary within a few percentage points of each other, so fall within the expected range of individual variation. The sternal plates are 24% of femoral length in the holotype of C. zoui, 17% in the holotype of C. dongi and an intermediate 21% in BPM 0001. A three percent difference in size does not seem to fall outside the range of individual variation. The first metacarpal of BPM 0001 and IVPP V 12430 is 39% of metacarpal II length. In IVPP V 12344, the ratio is 45%. This difference might be considered diagnostic if not for NGMC 97-4-A, which has a 42% ratio. Although stated to be "about .4" in Zhou and Wang, this figure comes from Ji et al., who only measured to the tenths place. The exact ratio, as mentioned above, is intermediate between the more divergent specimens. Once again, the 3% difference is considered insufficient to diagnose a species. The ilium is much shorter in NGMC 97-4-A (69% of femoral length) than in BPM 0001, IVPP V 12344 and IVPP V 12430 (77-79%). This is due to the broken anterior edge in the former specimen however, as can be seen in my photo on the Dinosauricon. The only potentially significant proportional difference between specimens is- manual ungual I vs. manual ungual II (106% in BPM 0001 and IVPP V 12430, 84% in IVPP 12344). Although the first manual ungual of NGMC 97-4-A is incomplete, it was much smaller than manual ungual II, so seems to match IVPP V 12344 better. A complete chart of proportions can be found at the end of this post (copy and paste to a text program such as Notepad to view correctly). It therefore seems that BPM 0001 and IVPP V 12430 share several cranial characters not seen in NGMC 97-9-A; NGMC 97-9-A has a slightly different coracoid morphology than BPM 0001; BPM 0001 and IVPP V 12430 have a different preacetabular morphology than IVPP V 12344; and that BPM 0001 and IVPP V 12430 have different manual ungual ratios than IVPP V 12344 and NGMC 97-4-A. Two groups of specimens are suggested by these differences- IVPP V 12344, NGMC 97-4-A and NGMC 97-9-A are one group, while BPM 0001 and IVPP V 12430 are in the other. The inclusion of IVPP V 12344 and NGMC 97-4-A with NGMC 97-9-A is far from certain, but the fact they all differ from the other two specimens and that the latter two have similar ungual ratios suggests this may be the case. Are these differences due to ontogenetic, sexual or taxonomic variation? The minute size variation (femur 145-152 mm) suggests it is not ontogenetic. Settling whether two sexes or species are involved is not easily resolved with only five specimens to work with, all from different localities. One potential way to decide this would be if the groups are not sister groups in a phylogenetic analysis.
Caudipteryx has been included in several phylogenetic analyses. Much of the following is modified from my earlier e-mails.
Ruben and Jones (2000) support the hypothesis Caudipteryx is a secondarily flightless bird. They argue against the theropod status of Caudipteryx by refuting the three "unambiguous characters" cited by Ji et al. (1998) that birds have and it lacks. They say the quadratojugal cannot be proven to have been sutured with the quadrate, as they do not contact in the holotype and show a photo of the specimen which differs in this aspect from the figure in Ji et al.. Oviraptorid specimens GIN B and ZPAL MgD-I/96 (Maryanska and Osmolska, 1997) show a cotylar articulation between the two bones and Velociraptor has a reduced loose contact (Barsbold and Osmolska, 1999), so the character wouldn't mean much even if they are correct (the cranial elements of Caudipteryx are loosely articulated and sutures between most cannot be seen). They claim the quadratojugal is far too short to contact the squamosal. New specimens confirm this is incorrect. Finally, "an obturator process may or may not exist in Caudipteryx, but, in any case, a similar structure is also known to have occured in some birds (e.g., Concornis)." First of all, they articulate the ischia upside-down, which is confirmed as being incorrect by the new specimens. Secondly, they misstate the character, which was to have a reduced or absent obturator process. I personally don't think Archaeopteryx has a reduced obturator process, but Concornis has a very reduced process. Ruben and Jones present no convincing evidence Caudipteryx is a bird.
Martin and Czerkas (2000) also argue Caudipteryx is a bird, one more derived than Archaeopteryx. The evidence presented is-
- teeth with expanded roots. Also in troodontids, Archaeornithoides, Microraptor, alvarezsaurids (which Martin considers ornithomimosaur relatives), etc.
- primary feathers. This is circular reasoning, as feathers are the only reason the authors argue against Caudipteryx being a dinosaur.
- carpus with at least four bones. Actually, there are three, but more are present in allosaurids, segnosaurs, tyrannosaurids and ornithomimids anyway.
- absence of pubic foot. Completely false (Zhou and Wang, 2000; Currie pers. comm. 1999).
- reflexed hallux. The hallux is said to be partially reversed, but a fully reflexed hallux is present in Rahonavis and MIcroraptor in any case.
- shortened tail. While the tail of Caudipteryx is shorter than any other non-avian theropod (and Archaeopteryx, Rahonavis and Yandangornis), it is not fused at all. The character "shortened tail", properly quantified, would be a potential Caudipteryx-pygostylian synapomorphy.
It's more derived than Archaeopteryx based on:
- no maxillary or dentary teeth. Enantiornithines and ornithurines primitively lack this. Is Caudipteryx supposed to be a carinate? Many dinosaurs (ornithomimids, oviraptorosaurs) also have this.
- external mandibular fenestra present. Plesiomorphic and present in virtually all theropods, while absent in a couple enantiornithines and maybe Archaeopteryx.
- enlarged premaxilla. No more than oviraptorids.
- reduced maxilla. Correlated with the above character and also present in oviraptorids.
- reduced hypopubic cup. Plesiomorphic and present in nearly all dinosaurs anyway....
- ball-shaped femoral head. Plesiomorphy only absent in Archaeopteryx and Rahonavis.
- reduced fibula. False (Ji et al., 1998), but present in troodontids, etc. anyway.
- reduced calcaneum. Less reduced than troodontids :-)
- greatly shortened tail with evidence of pygostyle formation. See above.
In conclusion, "shortened tail" is the only character presented by critics that is a potential Caudipteryx-pygostylian synapomorphy. Several new discoveries surprisingly add evidence to this hypothesis, including the possible low dorsal vertebral count and the reduced third manual digit. Still, even when these are added to a phylogenetic analysis, the non-avian hypothesis is much more parsimonious.
Originally thought of as a non-avian avialan by Ji, Currie, Norell and Ji (1998), recently it has been allied with oviraptorids. Ji et al. made Caudipteryx sister group to avians based on two characters- unserrated teeth; premaxilla reaches to anterior border of antorbital fossa. As only the premaxilla is toothed in Caudipteryx, and premaxillary teeth are often unserrated in coelurosaurs (juvenile tyrannosaurids, ornithomimosaurs, compsognathids, Byronosaurus, Sinornithosaurus, avians, etc.), this character is suspect. The second character is also seen in oviraptorids and more bird-like taxa (Bambiraptor, Sinornithosaurus), so can also be dismissed. Even Currie (1999 Ostrom Symposium) now places it as an oviraptorosaur.
Sereno (1999) groups it as an oviraptorosaur basal to caenagnathids and oviraptorids. This was based on twelve characters.
- ventral margin of external naris dorsal to maxilla. Not true in Caudipteryx, as can be clearly seen in BPM 0001.
- premaxilla participates in antorbital fossa. There does not appear to be much of an antorbital fossa in Caudipteryx. The promaxillary fossa is clearly separated from the premaxilla by a raised anterior maxillary rim.
- nasal shorter than frontal. The frontal length is difficult to determine in Caudipteryx. The nasals are shorter in NGMC 97-9-A, but longer in BPM 0001. The frontals are impossibly long in IVPP V 12430, as the orbit would be intersected by the jugal dorsal process. This suggests the long anterior portion of the frontals was overlapped by the nasals. This would also explain why the nasals are longer in BPM 0001, which is articulted, while the disarticulated NGMC 97-9-A shows the opposite condition. Thus, when articulated, the nasals would appear longer than the frontals. I use the character "nasals subequal in length to frontals", which is diagnostic of a more inclusive group, as it is present in the oviraptorosaur outgroup Erlikosaurus.
- ventral margin of dentary inset medially for dentary and surangular. This would be very difficult to determine from crushed two-dimensional specimens such as are preserved at Liaoning. Moreover, in BPM 0001 and IVPP V 12430, the maxilla ventrally overlaps the dentary and surangular on their lateral sides, suggesting the opposite condition. Thus, I do not consider the character an oviraptorosaur synapomorphy.
- anterior process of jugal rod-shaped in cross section. While the cross section of any element is nearly impossible to determine from the literature, the jugal of Caudipteryx is comparable to compsognathids, Ornitholestes, Scipionyx, ornithomimids, segnosaurs and many other coelurosaurs in slenderness. On the other hand, oviraptorids have rod-like jugals comparable to Avimimus, Shuvuuia and pygostylians. Because of this, I do not consider this an oviraptorosaurian character.
- palate mostly ventral to maxilla and jugal. This is also not easy to determine in Caudipteryx. However, I consider it unlikely based on a few observations. First, there is no indication of a ventrally projecting maxillary palate. Secondly, the vomers may be incredibly short and posterodorsally projected, in no position to reinforce maxillary "teeth". Finally, the C-shaped ectopterygoid would not be able to function like a strut between the maxilla and pterygoid. These features make it improbable that Caudipteryx had a ventrally projecting palate and the character is rejected as an oviraptorosaur synapomorphy.
- dentary dorsal margin convex. This is also present in therizinosauroids, so was a more inclusive character.
- external mandibular fenestra 40% of lower jaw length. Only confuciusornithids have such long mandibular fenestrae. Oviraptorids vary from 26-36%, while Chirostenotes has a 27% ratio and Caudipteryx varies between 33-39%. Change the ratio to 25% and it becomes a defensible oviraptorosaur character.
- caudal centra twice as wide as tall. The disarticulated caudal vertebrae of NGMC 97-9-A (visible in ventral view) show this not to be the case in Caudipteryx.
- anterior caudal transverse processes (1-8) twice length of neural spines. This is hard to determine when the vertebrae are preserved in lateral view, but the eighth caudal of NGMC 97-9-A is preserved in anterior (or posterior) view and seems to have transverse processes subequal in length to the neural spine.
- acromion prong projected anteriorly or anterodorsally. This is valid, although paralleled in most eumaniraptorans.
So most of Sereno's evidence for referring Caudipteryx to the Oviraptorosauria is not valid. An external mandibular fenestra over 25% of mandibular length and an anteriorly or anterodorsally projected acromial prong do support this assignment.
Holtz (1999 Ostrom Symposium) found it to either be a basal maniraptoran or a basal member of the segnosaur-oviraptorosaur group. His more recent analysis from SVP 2000 places it as an oviraptorosaur- sister group to Nomingia, which are in turn the sister group to Microvenator and caenagnathoids.
Although they did not perform a phylogenetic analysis, Zhou and Wang (2000) came to a similar conclusion, that it is either a basal oviraptorosaur or a basal paravian. Xu, Wang and Wu (1999) and Xu, Zhou and Wang (2000) (with near-identical analyses) also find it to be a basal paravian. Norell, Makovicky and Clark (1999 Ostrom Symposium) place Caudipteryx in a troodontid + segnosaur + oviraptorosaur group that is sister group to the Eumaniraptora. Zhou et al. (2000) admit that although cladistic analysis places Caudipteryx as a non-avian dinosaur, this is based on the untestable principle of parsimony and hypothetically no amount of characters could not have been reversed in a basal avian. This could just as easily be true if reversed (an infinite amount of bird-like characters could have evolved in parallel) and therefore is as good as saying "we can never know for sure what anything is related to". While this may be true, there are ways to test the probability one hypothesis is superior, which in this case is parsimony.
My analysis of 322 characters and 47 coelurosaur taxa* independently agree with Holtz's 2000 analysis. Caudipteryx is oviraptorosaurian based on- toothless maxilla; pneumatic lacrimal; prefrontal absent; frontal subequal in length to parietal; two long caudal dentary processes; dentary toothless; external mandibular fenestra more than 25% of lower jaw in length; retroarticular process long and tapering; no interdental plates; thin, rod-like prearticular?. It is the sister group of Nomingia based on- less than eleven dorsal vertebrae; tail less than 2.5 times femoral length; less than twenty-five caudal vertebrae; some distal caudal prezygopophyses elongate. It is more basal than Microvenator and caenagnathoids based on- distal caudals lack transverse processes; distal prezygopophyses elongate. When both groups of Caudipteryx specimens are included as separate OTU's, they are sister groups. This leaves open both possibilities- that the differences are sexual or taxonomic. I think the variation is too great to be sexual, but that is simply my undefendable personal preference. I therefore recommend Caudipteryx dongi be provisionally synonymized with C. zoui and that BPM 0001 and IVPP V 12430 be referred to as Caudipteryx sp. nov. until a more in depth study is published. Those who want to place all specimens in C. zoui are welcome to do so, as this is largely a subjective matter until a better study is completed or more specimens are discovered.
* Adding four characters and separating the Caudipteryx groups changed the results somewhat. Siamotyrannus has been kicked out of the Tyrannosauroidea and is now next to the outgroup (smile Nick :-) ). Ornitholestes is not maniraptoriform, but other examined "basal coelurosaurs" are maniraptoran. Bagaraatan is now in a polytomy with "basal coelurosaurs", segnosaur-oviraptorosaurs and paravians. Troodontids, Avimimus+alvarezsaurids and Achillobator+eumaniraptorans are in a polytomy. The Dromaeosauridae has standardized, with Dromaeosaurus and Adasaurus in the Dromaeosaurinae and Velociraptor, Deinonychus and Saurornitholestes in the Velociraptorinae. Bambiraptor, Sinornithosaurus, Microraptor, Unenlagia, Rahonavis and Archaeopteryx are successive outgroups to Yandangornis and pygostylians (sorry, no more secondarily flightless dromaeosaurs).
There's a lot of scans I could send people. The skull of NGMC 97-9-A and skeletons of NGMC 97-4-A and 97-9-A are available. Additionally, high quality photos of these specimens I took at the RTMP can be seen at- http://dinosauricon.com/artists/mm.html . The skeleton, pelvis and pes of IVPP V 12344 are available, as are plates of the torso and arm, ilia, pedes and feathers. The skull and pectoral girdle of BPM 0001are available. The skull, distal caudals and hand of IVPP V 12430 are also available. Also, plates of the skeleton, pectoral girdle and forelimbs, pelvis and pedes are available from BPM 0001; plates of the skeleton, pectoral girdle, distal caudals and ilia and tibiae and fibulae of IVPP V 12430 can be requested. If you want any of these, just ask.
Appendix A: ratios of skeletal elements in Caudipteryx.
1 2 3 4 5 6 7 8 9
IVPP V 12430 55 48 39 19 77 50 126 77
BPM 0001 55 21 50 41 19 79 50 130 78
NGMC 97-9-A 47 42 122 79
NGMC 97-4-A 24 47 42 69 52 128 78
IVPP V 12344 17 49 39 19 77 49 132 83
2. sternal plate/femur
1 1 1 1 1 1 1 1 1 1
1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
BPM 0001 10 10 86 6 82 63 39 89 89 57 61 86 54 10 100 14 90 95 27
IVPP V 12430 12 10 88 7 81 63 39 82 93 57 68 86 54 11 96 13 91 95 28
IVPP V 12344 12 8 84 7 79 61 45 93 86 52 64 86 62 11 92 15 90 94 29
1. humerus width/length
2. ulna width/length
4. radius width/length
14. femur width/length
2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0
BPM 0001 12 11 20 14 17 21 17 13 16 12 7 5 4 12
IVPP V 12430 11 11 20 13 15 21 15 12 16 11 7 5 4 13
IVPP V 12344 10 9 20 13 15 22 16 14 16 12 6 6 6 13