Segnosaurian Size

["Mickey Mortimer" <Mickey_Mortimer111@msn.com>]

Nick's recent post and Jaime's reply prompted me to examine the size of
segnosaurs in more depth.  These are ordered from smallest to largest.

Beipiaosaurus presents a unique case, due to its primitive morphology.  The
cervicals were about 32 mm, while the posterior dorsals were 44 mm.  The
cervical series is estimated to be 304 mm, and the dorsal series was about
430 mm.  This is based on comparison to Neimongosaurus, compensating for a
vertebral count of twelve cervicals and ten dorsals.  The preserved caudal
is most similar to the fourth of Neimongosaurus in central proportions and
was about 32 mm, suggesting a tail length of 635 mm.  The sacrum may have
been about 220 mm, assuming it had five vertebrae.  The skull length is
difficult to determine precisely because how much of the dentary is
preserved is neither described, nor discernable from the photos.  The
preserved length is said to be 65% of femoral length, which results in 172
mm.  In Erlikosaurus, the dentary is 68% of mandibular length, so
Beipiaosaurus' mandible was at least 253 mm long.  Again, using Erlikosaurus
for a craniomandibular ratio results in a skull at least 265 mm long.  The
total length of Beipiaosaurus was therefore about 1.85 meters long.

Neimongosaurus yangi has an almost complete vertebral column preserved,
assuming 22 presacrals like Nanshiungosaurus.  The fourteen cervicals are
804 mm long, the eight dorsals are 315 mm and the estimated 28 caudals are
about 755 mm (it appears about six are missing).  Though no sacral
measurements are given, both the last dorsal and the first caudal are about
40 mm.  Estimating for Neimongosaurus' six sacral vertebrae leads to a
length of 240 mm.  Assuming the skull is about 95% of humeral length (93% in
Erlikosaurus, 98% in Alxasaurus), that means a 211 mm long skull.  These all
total to 2.33 meters.

Erliansaurus bellamanus has wonderful appendicular preservation, but very
few known vertebrae.  The mid-cervical vertebra is actually shorter than
Neimongosaurus' (~85%), indicating a much shorter neck proportionately, as
all other dimensions are larger (appendicular elements 13-24% larger).  The
cervical series was approximately 680 mm, assuming a similar number of
cervicals.  The anterior dorsal centrum is 13% larger than Neimongosaurus',
giving a dorsal length estimate of 355 mm.  A sacrum about 20% larger than
Neimongosaurus' would be about 290 mm long.  The proximal caudal centrum
resembles the third of Neimongosaurus and is about 1.32 times as long,
suggesting a tail length of 990 mm.  Finally, the skull was probably about
255 mm long, judging by the humeral length.  Thus, Erliansaurus was
approximately 2.57 meters.

Erlikosaurus andrewsi is known from a skull, fragmentary cervical vertebrae,
humerus
and pes.  Unfortunately, the vertebral dimensions are undescribed, so the
length can be only loosely estimated.  Based on the humeral, cranial and
metatarsal lengths, Erlikosaurus was somewhere around 3.4 meters long.

Alxasaurus elesitaiensis is known from five individuals, but only three have
published measurements.  The holotype (IVPP 88402a) is largest, while IVPP
88501 is about 75% as large, and IVPP 88402b is said to be about 66% as
large.  The skull was about 370 mm long, judging by the dentary.  Using
Neimongosaurus, but estimating for a more plesiomorphic vertebral count,
generates lengths of 904 mm and 605 mm for the cervical and dorsal series
respectively.  The cervicals were comparatively shorter in Alxasaurus.
Finally, the sacrum was 345 mm long, and the tail was 1554 mm long,
accounting for the several missing vertebrae (which also total around 28).
This totals to 3.78 meters, remarkably close to Russell and Dong's 3.8 meter
estimate based on a probably inaccurate presacral formula (10 cervicals, 13
dorsals).  IVPP 88501 would then be about 2.83 meters, and IVPP 88402b about
2.49 meters.

Nanshiungosaurus brevispinus preserves the complete presacral column, sacrum
and pelvis.  It has twelve cervical vertebrae, ten dorsal vertebrae and five
sacral vertebrae.  Unfortunately, the description (Dong, 1979) is in Chinese
and few illustrations are given.
One potential problem with size estimates for this species is that both the
line drawings and the plates are said to be at 1/4 size, but the plates are
1.7 times larger.  The line drawing scale is correct, as shown by two pieces
of evidence.  First, Russell and Dong (1994) estimate Nanshiungosaurus' mass
as 1.5 tons, compared to Segnosaurus' 1.9 tons.  The pelves of Segnosaurus
(GM 100/80) and Nanshiungosaurus are similar in size assuming the smaller
size estimate, with two other Segnosaurus specimens being slightly larger.
Russell and Dong's mass estimates make sense in this case.  If
Nanshiungosaurus were 1.7 times larger though, it would also be 1.7 times
larger than Segnosaurus' holotype.  It would then be almost 5 times heavier,
putting the mass estimate at over 9 tons.  Second, Dong and Yu (1997)
diagnosed Nanshiungosaurus? bohlini as being larger than any known
segnosaur.  The first dorsal vertebra in this taxon is 124 mm long.  If the
smaller size estimate is correct, Nanshiungosaurus brevispinus' twelfth
cervical (topologically equivalent to N? bohlini's first dorsal, as the
latter has two less cervicals) is 76 mm long, but if the larger estimate is
correct, it is 129 mm long.  This would make N. brevispinus equivalent to N?
bohlini in size, clearly not what Dong thought was true when describing the
latter species.  Note both of the reasons involve Dong, who described
Nanshiungosaurus and should therefore know its size.
Subtracting for spaces between centra in the mounted specimen gives a
cervical length of 1400 mm, a dorsal length of 730 mm and a sacral length of
680 mm.  Adding 375 mm for a Segnosaurus-proportioned skull and 1200 mm for
a Neimongosaurus-proportioned tail gives a total length of 4.40 meters.

"Chilantaisaurus" zheziangensis was described in the same paper as
Nanshiungosaurus brevispinus, and was thought by Dong (pers. comm. to
Molnar, 1984) to possibly be part of the same specimen.  It is based on a
manual ungual and digits II and III of the pes.  They are quite similar to
the Therizinosaurus(?) pes, confirming the assignment to the Segnosauria.
They are also about 77% as large as that specimen, giving the taxon an
approximate length of 5 meters.  This is close to Nanshiungosaurus' length,
adding evidence to Dong's hypothesis of synonymy.

Enigmosaurus mongoliensis is only known from an incomplete pelvis, which is
about 95% the size of Segnosaurus' holotype.  This leads to a tentative
length estimate of 5 meters.

Nothronychus mckinleyi is known from a partial skeleton which includes many
appendicular elements, but only a few vertebrae.  The longest cervicals are
about 2.6 times longer than Neimongosaurus, leading to a cervical length of
1780 mm, assuming only twelve vertebrae.  The width and height of the
anterior dorsal vertebra (which is one of the first three because it has a
hypapophysis) averages 118% (110-132%) of the equivalents in
Nanshiungosaurus brevispinus' twelfth dorsal and third dorsal.  A dorsal
column 1.18 times larger than Nanshiungosaurus' would be 860 mm.  The
proximal caudal vertebra is most similar to the third in Neimongosaurus,
which would indicate a tail 1380 mm long when scaled up.  Finally, scaling
Segnosaurus' skull and Nanshiungosaurus' sacrum adds 440 mm and 800 mm
respectively.  Combining these leads to a total length estimate of 5.26
meters.

Nanshiungosaurus? bohlini is based on most of a presacral column.  The
cervical series is 1515 mm long, though it only contains ten vertebrae
according to Dong and Yu (1997).  The anterior dorsals are 23% larger than
N. brevispinus', suggesting a dorsal length of 1035 if two more vertebrae
are added to make up for N. brevispinus' high number of cervicals.  Scaling
up cranial, sacral and caudal lengths from N. brevispinus adds 460 mm, 835
mm and 1480 mm respectively.  Nanshiungosaurus? bohlini was thus about 5.3
meters long.

The holotype of Segnosaurus galbinensis (IGM 100/80) includes a mandible,
pelvis and forelimb and vertebral elements.  Unfortunately, only the
mandibular, manual and pelvic dimensions are described.  The pelvis is of
virtually identical size to Nothronychus, suggesting this specimen was also
close to 5.3 meters.  A referred specimen (IGM 100/83) is a cervical
fragment, pectoral girdle and forelimb, the latter missing all manual
material save an ungual.  The forelimb elements are about 20% larger than
Nothronychus', so the specimen may have been about 6.3 meters long.  Another
possibility is that this was similar in size to the holotype, but
Segnosaurus has proportionately longer arms than Nothronychus.  The holotype
includes comparable elements (humerus, radius, ulna), so the problem is
soluble granted someone examines the remains.  The second referred specimen
(IGM 100/82) includes rib fragments, partial pelvis and virtually complete
hindlimb.  There is confusion regarding the limb element length in this
specimen.  Barsbold and Perle (1980) state the tibia is shorter than the
femur, but cite a femoral length of 840 mm and a tibial length of 860 mm.
Barsbold and Maryanska (1990) stated the tibia was "more than 80%" of
femoral length.  The illustration in Barsbold and Perle suggests the tibia
is about 830 mm long.  This is very close to the cited measurement,
suggesting the latter is accurate.  The femur would then be slightly less
than 1075 mm in length.  Comparing the hindlimb elements to other segnosaurs
indicates it was approximately 7.0 meters long.  Again, it's possible this
was similar in size to the type, but had proportionately longer hindlimbs
than Nothronychus.  And again, examination of the partial pelvis the two
share in common could resolve the problem.

Finally, the largest segnosaur- Therizinosaurus cheloniformis.  The holotype
(PIN 551-483) consists of three incomplete manual unguals.  I believe two
belong to the first digits, with the smaller one belonging to the second
digit.  This is supported by the resemblence of the smaller one to an ungual
 associated with the second digit in a referred specimen, the fact
Erliansaurus' largest ungual is the first, the strong resemblence between
the two large unguals, and the theropod rule that manual ungual III is
always smallest (the exception being confuciusornithids).  The referred
specimen IGM 100/15 includes the pectoral girdle, forelimb (missing digits I
and III), a tooth, dorsal ribs and gastralia.  The forelimb is proportioned
very much like Nothronychus, and is about 1.8 times larger, leading to a
length of 9.6 meters.  The holotype is about 85% as large, so 8.2 meters
long.  Another proximal end of manual ungual I (IGM 100/17) was about 66% as
long as the holotype, so was 5.4 meters long.  Another referred manual
ungual lacks the proximal portion, so cannot be measured accurately.  The
hindlimb specimen (IGM 100/45) is not directly comparable, so is referred
only tentatively.  It is actually only 94% as large as Segnosaurus specimen
IGM 100/82, so would be about 6.5 meters long.

Segnosauriam masses can also be estimated by using Russell and Dong's (1994)
value for Alxasaurus, though the varying proportions make any such numbers
tentative.
Beipiaosaurus- 1.85 m (45 kg)
Neimongosaurus- 2.33 m (90 kg)
Alxasaurus referred IVPP 88402b- 2.49 m (110 kg)
Erliansaurus- 2.57 m (120 kg)
Alxasaurus referred IVPP 88501- 2.83 m (160 kg)
Erlikosaurus- 3.4 m (~275 kg)
Alxasaurus holotype- 3.78 m (380 kg)
Nanshiungosaurus brevispinus- 4.40 m (600 kg)
"Chilantaisaurus" zheziangensis- ~5 m (~900 kg)
Enigmosaurus- ~5 m (~900 kg)
Nothronychus- 5.26 m (1.0 ton)
Nanshiungosaurus? bohlini- 5.3 m (~1 ton)
Segnosaurus holotype- 5.3 m (~1 ton)
Therizinosaurus referred IGM 100/17- 5.4 m (~1.1 tons)
Segnosaurus referred IGM 100/83- 6.3 m (~1.8 tons)
Therizinosaurus? hindlimb- 6.5 m (~1.9 tons)
Segnosaurus referred IGM 100/82- 7.0 m (~2.4 tons)
Therizinosaurus holotype- 8.2 m (~3.9 tons)
Therizinosaurus referred IGM 100/15- 9.6 m (~6.2 tons)

Mickey Mortimer