ArticlePDF Available

Abstract and Figures

The Late Jurassic sauropod dinosaur, Seismosaurus hallorum, from the Morrison Formation of New Mexico, is reconstructed as a full-scale replica using casts of the holotype. We document the osteology of prepared elements from the thoracic, pelvic and caudal regions in comparison to related diplodocids. The thoracic vertebrae are described pending further preparation. A sauropod femur, collected with this material and believed to have been from the same individual, is also described. Using osteo-morphological comparisons with closely related diplodocids, pre-vious axial length estimates of 39-52 meters for Seismosaurus hallorum are questioned, and the overall length is reinterpreted to have been approximately 33 meters.
Content may be subject to copyright.
Foster, J.R. and Lucas, S. G., eds., 2006, Paleontology and Geology of the Upper Jurassic Morrison Formation. New Mexico Museum of Natural History and Science Bulletin 36.
1School of Physics, University of Western Australia, Crawley, Western Australia 6009,; 2New Mexico Museum of Natural History, 1801
Mountain Road NW, Albuquerque, New Mexico 87104-1375
Abstract—The Late Jurassic sauropod dinosaur, Seismosaurus hallorum, from the Morrison Formation of New
Mexico, is reconstructed as a full-scale replica using casts of the holotype. We document the osteology of prepared
elements from the thoracic, pelvic and caudal regions in comparison to related diplodocids. The thoracic vertebrae are
described pending further preparation. A sauropod femur, collected with this material and believed to have been from
the same individual, is also described. Using osteo-morphological comparisons with closely related diplodocids, pre-
vious axial length estimates of 39-52 meters for Seismosaurus hallorum are questioned, and the overall length is
reinterpreted to have been approximately 33 meters.
The sauropod dinosaur Seismosaurus hallorum is the only en-
demic sauropod taxon described from the Upper Jurassic Morrison
Formation in New Mexico and has been considered by some the long-
est known dinosaur (e.g., Gillette, 1991, 1994). For the purposes of
museum exhibits in Japan and New Mexico, reconstruction of S.
hallorum was undertaken using fossils of the holotype and osteological
information from closely related diplodocids. Mounted replicas (Fig.
1) were produced by the Canadian-based company, Prehistoric Animal
Structures Inc., in conjunction with the New Mexico Museum of Natu-
ral History and Science. The first is on display in the Kitakyushu Mu-
seum of Natural History, Japan and the second at the New Mexico
Museum of Natural History and Science.
Several mid-caudal vertebrae, chevrons, the right pubis and the dis-
tal right ischium of Seismosaurus hallorum were described by Gillette
(1991), however, the thoracic, sacral and proximal caudal vertebrae were
not fully prepared or formally described. Further preparation of these ele-
ments exposed key osteological details for description and allowed recon-
struction to proceed with reasonable accuracy. To date, no cranial material,
cervical vertebrae, terminal caudal vertebrae, scapulae or limbs elements
(other than an incomplete femur) have been recovered. Hence, size and
length estimates of Seismosaurus hallorum have been speculative.
Seismosaurus hallorum is distinctly diplodocid with elements re-
markably similar to the known species of Diplodocus (Paul, 1988; Gillette,
1991, 1994; Lucas, 1993; Curtice, 1996); hence, comparisons of length,
morphological character and osteological proportions used in this recon-
struction centered on Diplodocus. Several workers have questioned the
validity of the genus, Seismosaurus hallorum (Paul, 1988; Lucas, 1993;
Curtice, 1996; Lucas et al., 2004), due to the close similarity of osteo-
morphological characters in the caudal vertebrae with those of Diplodocus.
The ischium of Seismosaurus hallorum, with its dorsally projecting distal
expansion, long appeared to be the only autapomorphic character of the
genus (Gillette, 1991; Upchurch et al., 2004). However, recent complete
preparation of the ischium revealed the hook-like process to be part of a
neural spine, so the validity of Seismosaurus is questionable (see Lucas et
al., this volume).
The relatively high ratio of caudal, sacral and thoracic neural spine
height to centra length for Seismosaurus hallorum, and the upright char-
acter of the caudal neural spines are not necessarily autapomorphic charac-
ters, when compared with other diplodocids (contrary to Gillette, 1991,
1994). Neural spine height and inclination are known to be variable within
archosaur taxa, such as sexually dimorphic crocodilians, and it is not a
reliable character for the separation of genera (Curtice, 1996, p. 26-27).
Similarly, Gilmore (1932) noted that the neural spines of Diplodocus lon-
gus mounted in the United States National Museum are variably erect to
backwardly inclined and thus were not a reliable character for separation of
species of Diplodocus. Further, Curtice (1996, p. 26) notes that erect
neural spines are typical of diplodocids.
Regardless of the generic status of Seismosaurus, the recovered el-
ements are those ofa very large diplodocid. Here, we document in detail
the skeletal reconstruction of Seismosaurus and reassess its axial length.
Institutional Abbreviations-American Museum of Natural His-
tory (AMNH); Carnegie Museum of Natural History (CM); Field Museum
(FM); New Mexico Museum of Natural History and Science (NMMNH);
National Museum of Natural History (NMNH); Prehistoric Animal Struc-
tures (PAST); Science Museum of Minnesota (SMM); Utah Museum of
Natural History (UMNH); Wyoming Geological Museum (UWGM).
Prepared and semi-prepared elements of the type specimen of
Seismosaurus hallorum (NMMNH P-3690) and the femur catalogued
separately (NMMNH P-25079) were photographed, measured and illus-
trated. Other than the elements described by Gillette (1991), recovered
material also included four incomplete dorsal ribs, the partial right ilium in
articulation with the sacrum, proximal caudal vertebrae Ca1 to Ca8 and
thoracic vertebrae T3 to T10. In this work we refer to the detailed synopsis
of Wilson (1999) for nomenclature of vertebral lamina relevant to
For the physical reconstruction (replica construction), all elements
of the holotype were molded and cast in fiberglass where possible. All miss-
ing elements were sculpted using polyurathane foam with plaster texturing
and modeling clay. Fiberglass casts were produced by PAST for all ele-
ments and articulated over a steel structure to produce mounted replicas.
The image of the mount (Fig. 1) was produced using large format slides of
the first mounted replica.
Despite the extensive and sophisticated search methods employed
in the field by Gillette’s team, no cranial or complete cervical vertebrae
were located. The skull of Seismosaurus was modeled using Diplodocus
morphology and the detailed description of Holland (1934). The dorsoven-
tral depth was increased slightly (with artistic license), to reflect the higher
comparative ratio of neural spine height to centrum length evident in the
thoracic and caudal vertebrae of Seismosaurus hallorum.
Cervical Vertebrae
Very fragmented remains of two vertebrae (unprepared) and an in-
tact isolated rib are questionably from the cervical region, but these are
currently thought to be of cranial thoracic location. As no cervical verte-
brae are known, reconstructed cervical vertebrae were modeled completely
on Diplodocus morphology and were proportioned according to the rela-
tive dimensions of Diplodocus thoracic, sacral and caudal vertebrae
(Table 1).
Articulation of the cervical series and skull depicted in Figure 1
follows the gentle “S” curvature often illustrated for sauropods. The
NMMNH mount incorporates a more elongated neck and obtusely-
angled articulation of the skull. Support for this manner of cervical
articulation can be found in Frey and Martin (1997), who discuss the
biomechanical limits of bracing systems, and Stevens and Parrish (1999),
who detail cervical articulation and flexion in both Apatosaurus and
Diplodocus, based on digital reconstructions of the limits imposed on
the cervical articulating surfaces.
Thoracic Vertebrae
Thoracic vertebrae T1 and T2 are in degraded condition. Verte-
bra T3 is mostly prepared, although the caudal surface remains unpre-
pared. Vertebrae T4 to T10 are contained in one large block, which is
partially prepared on the left side and ventrally. The thoracic vertebrae
of Seismosaurus hallorum are strikingly similar in their general mor-
phological character to those described for Diplodocus carnegii (CM
84) by Hatcher (1901, p. 25-30) and Diplodocus longus (USNM 10865)
by Osborn (1899, p. 193-199). In their current state of preparation,
several characters of note can be reported for the thoracic vertebrae
Seismosaurus hallorum, T3 to T10. These include:
1. The neural arches are highly elevated, as they are in Diplodocus.
2. Centra T3, T4 (Figs. 2A-B) and possibly T5 have hemispherical
cranial condyles and concave cotyles (opisthocoelous). These centra are
transversely constricted mid-centra to form keels on their ventral surfaces.
The medioventral constriction is greatest in T4 and corresponds with the
description for Diplodocus carnegii by Hatcher (1901, p. 28). The overall
transverse width of centrum T4 is small, in comparison with Diplodocus
carnegii (Table 1) and other diplodocids. Similarly, the centrum T5 ap-
pears limited in transverse size, although only the cranial portion is pre-
served. The remarkably reduced size of centra T4 and T5 within the tho-
racic series seems at odds with the large overall proportions of Seismosaurus
3. The middle and caudal thoracic vertebrae, from T6 to T10, are
convexly shallow at their cranial articulating surfaces and are concave on
their caudal surfaces (amphyplatyan). The centra are gently convex in mid-
transverse section, with the centrum size of the thoracics caudal to T5
markedly increasing in overall transverse width and depth. The cranio-cau-
dal lengths of these vertebral centra do not increase and become shortest in
T9 and T10 (Table 1). The widths of the caudal thoracic vertebrae from T6
exceed their lengths (Table 1), which is a characteristic of Diplodocus and
4. Pleurocentral openings (pc [pleurocoels]) are seen in all centra,
T3 to T10, where prepared. These can be seen in Figure 2 for T3, T4 and
T8, which show varying architecture in the openings. Pleurocoels of T3
appear as fossae, or shallow excavations (Fig. 2A). The pleurocoels of
T4 are far more developed as two deep lateral foramina or openings
into the centrum (Fig. 2B). The anteroventral foramina of T4 extend
into the rim of the vertebral condyle (Fig. 2B). Although T5 is partially
preserved, the cranial end is attached to T4 and shows a single small
pleurocentral opening in the same lateral line as the ventral-most
pleurocoel in the centrum of T4. The left pleurocoel of T8 is indicative
of the caudal thoracics; it is ovular at the parasagittal surface, sub-
triangular at the lateral surface, which extends onto the neural arch,
and is displaced slightly cranially.
5. Open cancellous structure is exposed within the condyles of cen-
tra T3, T4 and T5 (Fig. 2A-B).
6. The caudal centroparapophyseal lamina (pcpl) (the oblique
lamina in Hatcher, 1901), which project caudoventrally from the
parapophyses (pp) to the caudal neurocentral junction, are well developed
in the middle to caudal thoracics from T5, T7, T8 (Fig. 2C) and T9. This is
at odds with Wilson (1999), who reports that the pcpl are not generally
well developed in sauropods. In these vertebrae the pcpl converges with
the caudal centrodiapophyseal lamina (pcdl) (the inferior diapophyseal
lamina in Hatcher, 1901), in a manner similar to that seen in Diplodocus
(Hatcher, 1901, p. 29), on the midlateral surface of the neural arch be-
tween the diapophyses (dp) and the neurocentral junction. In Apatosaurus
louisae, pcpl and pcdl of T6 cross (Gilmore, 1936, pl. 25), with the pcdl
terminating at the cranial dorsolateral junction of the centrum. The pcdl of
T6 in Seismosaurus hallorum appears to continue to the caudal neurocentral
junction, similar to Diplodocus, although further preparation is needed to
verify this condition.
7. Similar to Diplodocus longus, the transverse processes and di-
apophyses of T5 to T10 and especially T8 (Fig. 2B) and T9 are more highly
elevated above the level of the zygapophyses than in Diplodocus carnegii
and Apatosaurus. The transverse processes of T3 are lower than the
prezygapophyses (prz) and postzygapophyses (poz) and are more robust
than the caudal thoracics. The diapophyses of T3 face ventrolaterally. In
shape, the diapophyseal articulating surfaces of T3 are triangular and dor-
soventrally elongated (Fig. 2A-C). The transverse processes of T4 are tran-
sitional between the condition of T3 and the caudal thoracics (Fig. 2B) The
articulating surfaces of the middle to caudal diapophyses are small,
subtriangular in shape and project slightly ventrolaterally. The parapophyses
are similarly small and are connected to the prezygapophyses by the
paradiapophyseal lamina (prpl). These characters are seen in T8 (Fig. 2C)
8. The postzygodiapophyseal lamina (podl), which extends between
the diapophyses and the post-zygapophyses of T8 (Fig. 2C) and T9, are
paired with the superior (upper) lamina emanating from the transversely
oriented, sheet-like spinodiapophyseal lamina (spdl). A similar condition is
seen in Diplodocus carnegii.
9. The prezygodiapophyseal lamina (prdl) that extend between the
diapophyses and the prezygapophyses are well developed wing-like sheets
of bone in T5 to T10. The paradiapophyseal laminae (ppdl) that extend
FIGURE 1. Seismosaurus hallorum, NMMNH P-3690, right lateral view of the mounted replica. Total axial length equals 33 meters; vertical height to dorsal extremity
of the sacral neural spines equals 4.95 meters.
between the parapophyses and the diapophyses (or from the pcdl in
some caudal thoracics) in Diplodocus and Apatosaurus are not well
developed in Seismosaurus hallorum. Paradiapophyseal laminae are
evident in T8 (Fig. 2C) and T9, however, these structures are uncertain
and are as likely to be small accessory lamina.
10.Hypantra are not evident on the prezygapophyses of T3, with
the paired intraprezygapophyseal laminae (tprl) forming a broad concave
sheet of bone between the left and right prezygapophyses. The caudal sur-
face of T3 is unprepared and existence of a hyposphene cannot be deter-
mined at this stage. Figure 2A illustrates the probable layout of lamina on
the caudolateral surface of the neural arch. Unfortunately, the
prezygapophyseal region of T4 is eroded, and the presence of the hypantra
cannot be determined. The postzygapophyses of T8 (Fig. 2C) are ventro-
laterally and caudally oriented and are transversely concave in the manner
reported for Diplodocus and Barosaurus by Upchurch et al. (2004). In
Seismosaurus hallorum, the left and right post-zygapophyses of T8 coa-
lesce at their medioventral margins and appear to form a hyposphene
(see Fig 2C), although this needs further examination.
11.The centropostzygapophyseal lamina (tpol) extends from the
medioventral base of the postzygapophyses to the caudolateral neurocentral
junction. This can be seen on T8 (Fig. 2C). The centropostzygapophyseal
lamina (cpol) is also seen on T8, extending from the cranial dorsolateral
edge of the postzygapophyses ventrally to the the podl, and projects
cranioventrally to the midlateral neural arch. The cpol appears to continue
cranioventrally as a poorly developed lamina that extends to the cranial
neurocentral junction (Fig. 2C). The prezygapophyses of T9 (not illustrated)
appear to extend to a hypantrum. A similar condition is likely for the other
middle to caudal thoracics, but further preparation will be required to de-
termine where the hypantrum-hyposphene articulation starts in the tho-
racic series.
12.In Apatosaurus louisae, the parapophyses migrate dorsally onto
the neural arch at T3, and this continues to T10. With the migration of the
parapophyses onto the neural arch, the cranial centrodiapophyseal lamina
(acdl), which extends between the diapophyses and the neurocentral junc-
tion above the parapophyseal facets of the cranial thoracic vertebrae, di-
vides into two separate lamina: the cranial centroparapophyseal lamina
(acpl), which projects ventrally from the parapophyses to the cranial
neurocentral junction; and the ppdl, which extends dorsally or
dorsoposteriorly from the dorsal parapophyses (Gilmore, 1936, pl. 25; Wil-
son, 1999). In Seismosaurus hallorum, however, and similar to Diplodocus
(Hatcher, 1901), the parapophyses are located on the centrum of T3 and
possibly at the neurocentral junction of T4 (Figs. 2A-B), thus maintaining
the singular acdl condition at T4.
13.Neural spines from T3 to T5 are bifurcated, with the left and
right spinoprezygapophyseal lamina (sprl), spinopostzygapophyseal lamina
(spol) and the transversely-oriented spdl forming each spine. The middle to
caudal thoracic neural spines, from T6, become decreasingly less bifur-
cated with the increased fusion of the left and right sprl, spol and spdl along
the medial prespinal lamina (prsl) and postspinal lamina (posl). The dorso-
lateral summits of these vertebrae form slightly outwardly-projecting horns,
as can be seen at the neural spine summit of T8 (Fig. 2C). In the caudal
thoracic vertebrae of Seismosaurus hallorum, the spol are apparent as two
rami (Fig. 2C), one lateral (lat. spol) and the other medial (med. spol), and
this is a condition common in the middle and caudal thoracics of sauropods
(Wilson, 1999). In Seismosaurus hallorum the lateral spol of the cranial
thoracic vertebrae are well developed, as they are in other diplodocids (Wil-
son, 1999). These are seen in T3 and especially T4 (Figs. 2A-B), where the
lat. spol and spdl form deep fossae. In the middle and caudal thoracics, the
fused lat. spol and spdl form a pronounced, laterally-projecting bump. On
the cranial surface, a small accessory lamina is formed laterally between
the inner edge of the bump and the prsl. These laterally-projecting bumps
of the mid-lateral neural spines are seen in Diplodocus from T5 to T9
(Hatcher, 1901) and Barosaurus (Lull, 1919), however, they are not ap-
parent in Apatosaurus (Gilmore, 1936, pl. 1V).
14. On the cranial surfaces of the middle to caudal neural spines, the
sprl and the prespinal lamina (prsl) fuse to become broad medial bony plates
that are capped by a rugose, bony thickening formed between the lateral
“horns.” The left and right med. spol in these vertebrae coalesce in the
lower part of the neural spine with the posl, to form a narrow ridge to the
summit of the spine and are similarly capped by a rugose bony thickening.
The ratio of vertebral height to centrum length (Table 1) indicates
that Seismosaurus hallorum and Diplodocus carnegii are comparable in
the preserved cranial thoracics, T3 and T4. The middle to caudal thoracics
(T6 to T10) of S. hallorum are, however, approximately 20 percent longer
than those of D. carnegii (Table 1). Total vertebral heights are comparable
in S. hallorum and Apatosaurus (Table 2), although the middle to caudal
thoracics of both Apatosaurus louisae and Apatosaurus excelsius are mar-
ginally higher for the vertebra represented. The ratio of vertebral height to
centrum length (Table 1) indicates that Seismosaurus hallorum is gener-
ally similar to Apatosaurus, although possibly less in the middle to caudal
thoracics T6 to T10.
Vertebrae P-3690 CM 84 CM 3018 FM 7163
T3 3.37 2.21 2.85 1.76
T4 3.00 3.21 3.83 2.48
T5 3.60 3.45 4.08 3.03
T6 3.59 3.11 4.72 3.79
T7 3.32 5.15 4.62
T8 4.20 3.08 4.91 4.30
T9 4.38 4.20 5.24
T10 4.64 3.69 5.36
CA1 6.30 5.73 5.21 5.89
CA2 5.81 5.46 4.59
CA3 4.74 4.00
CA4 4.58 3.32 4.84 3.78
CA5 3.64 3.11 4.37 3.45
CA6 3.85 3.14 4.37 3.24
CA7 3.44 2.91 3.55 3.15
CA8 2.74 3.38 2.65
CA9 2.41 3.11 2.65
CA10 2.27 2.79 2.50
CA11 2.27
CA12 2.23 1.95 2.63 2.74
CA13 2.38 2.35
CA14 2.39 2.41 2.35
CA15 1.99 2.00
CA16 1.88 1.86
CA17 1.89 1.73
CA18 1.86 1.64
CA19 1.41 1.00
TABLE 1. Seismosaurus hallorum NMMNH P 3690, Diplodocus
carnegii CM 84, Apatosaurus louisae CM 3018 and A. excelsus FM
7163, ratios of total vertebral height to centrum total lengths.
Ratios of vertebral height to centrum length
FIGURE 2. Thoracic vertebrae of Seismosaurus hallorum, NMMNH P-3690. A, T3 (right to left) cranial and left lateral views; B, T4 (right to left) cranial and left lateral
views; C, T8 (left to right) cranial, left lateral and caudal views. For abbreviations see text; scale bar equals 300 mm.
Thoracic Ribs
Two rib heads and two partial ribs were recovered by Gillette’s team.
These elements are not described formally, but during reconstruction were
positioned according to their appropriate morphological shape and size.
These positions correspond to left thoracic 1, right thoracics 3 and 4 and
left thoracic 8. Left thoracic 1 was positioned because of the small overall
size compared with the other recovered ribs, with the shaft straight and
triangular in cross section (description matching that of Diplodocus
carnegii: Hatcher, 1901). The tuberculum is aligned linearly with the shaft,
indicating a relatively perpendicular orientation visceral to the glenoid. The
two rib heads, thoracic right 3 and 4, are partially prepared on one block of
sandstone. Casts of these rib heads were used to locate their correct posi-
tion on the thoracic vertebrae. These positions were matched by their sec-
tional size, capitulum and tuberculum spacing and alignment on the verte-
bral apophyses and shaft alignment. Left thoracic rib 8 is proximally com-
plete and was recovered overlying the sacral spines. This rib was similarly
matched to T8. Left thoracic rib 1 has the appearance of an ossified
tendonous mass, which is likely to be caused by exposure and erosion of
the internal cancellous bone structure.
The sacrum of Seismosaurus hallorum is robust and has under-
gone little compression or distortion. Some oblique offset is evident sagit-
tally between the left and right sides. The sacrum is prepared dorsally and
clearly shows a radiating pattern of five sacral ribs in articulation with the
right ilium and the neural spines; the third and broadest sacral rib measures
190 mm in parasagittal section. Estimated transverse widths between the
sacricostal yoke are cranially ~1250 mm and caudally ~1100 mm (Table
2). Ventrally, the sacral centra, ribs and sacricostal yoke are unprepared.
Due to erosion, there is some ambiguity as to the exact morphological char-
acter of the cranial and caudal surfaces of sacral ribs and centra 1 and 5,
however, the estimated sacral axial length measures ~1100 mm.
In agreement with Gillette (1994), the sacrum of Seismosaurus
hallorum is comparable in size and robustness to that of Apatosaurus
louisae, although the former is broader transversely and craniocaudally
shorter (Table 2). The first and fifth sacral spines are mostly unfused, al-
though these may be fused at their summits. The fused sacral spines 2-3-4
measure approximately 680 mm, from the dorsalmost edge of the right
ilium. The heights of the sacral spines are not available for Apatosaurus
louisae (CM 3018), as they are not preserved (Gilmore, 1936). However,
reconstruction of this height using Apatosaurus excelsus (CM 563: Gilmore,
1936, fig. 19) as a model indicates that the neural spine height for
Apatosaurus louisae would have been approximately 600 mm and is there-
fore comparable with Seismosaurus hallorum. The maximum sacral spine
height for Seismosaurus hallorum is 50% greater than Diplodocus carnegii
(CM 84: Table 3).
Caudal Vertebrae
The cranial caudal series Ca1 to Ca7 of Seismosaurus hallorum is
exposed in two sandstone blocks in a semi-prepared state. Preparation of
the cranial caudal vertebrae shows: the zygapophyses in articulation; the
neural arches and attachment surfaces for the separated, recovered neural
spines of Ca2 and Ca4 to Ca7; and well-developed laminae of the laterally-
projecting transverse processes and left, upper lateral surfaces of the centra
in articulation. The ventral surfaces of these caudal centra remain unpre-
pared. As the centrum of Ca1 is mostly eroded, it was reconstructed using
caudal Ca2 and Ca3 as guides.
An isolated caudal vertebra was established as Ca8 (confirming the
tentative placement of Gillette, 1991) due to the perfect articulation of the
prezygapophyses with the postzygapophyses of Ca7. Centrum size, shape,
the development of deep pleurocentral fossae and the degree of develop-
ment of the transverse processes are consistent with direct articulation of
the preceding vertebrae. Similar to Diplodocus, a dorsal cleft is present at
the summits of the neural spines of the cranial caudal vertebrae, up to Ca7.
Although a cleft could not be established for the transversely-oriented
spines of the more terminal vertebrae, it was included in the recon-
structed spine of Ca8.
Some distal sections of the transverse processes are missing from
cranial Ca2 to Ca8. However, intact sections show that these processes
were pronounced and elaborately developed, as they are in Diplodocus.
The transverse processes of the cranial caudals, up to Ca8, extend
dorsolaterally and form pronounced shoulders. The rugose parasagittal
margins of these transverse processes extend cranially beyond their centra
from Ca1 to at least Ca4. Terminal to Ca4, the transverse processes do not
appear to extend ventrally below the middle centrum and reduce rapidly in
their lateral extent from Ca8. The transverse processes are reduced to bumps
at Ca19. The cranial centrodiapophyseal laminae are well-developed in the
first 15 caudal vertebrae, as noted for Diplodocus by Wilson (1999). In
Seismosaurus hallorum, this lamina is weakly developed at Ca17 and ab-
sent from Ca18.
Since the publications of Gillette (1991, 1994), recent preparation
of cranial caudal vertebrae Ca1, Ca2 and Ca4 to Ca7, with their intact
neural spines, gives an accurate indication of spine height above the
zygodiapophyseal lamina (the horizontal lamina that forms the dorsal mar-
gins of the transverse processes), as well as total vertebral height at these
vertebral positions. Total vertebral heights are summarized in Table 2, to-
gether with comparative heights for closely-related diplodocids.
The missing neural spines for cranial vertebrae 3 and caudal 8 were
reconstructed using the intact vertebrae with their sacral spines as a guide.
A diagrammatically-imposed dorsal ridgeline following the progression of
decrease in neural spine height from the sacrum across the recovered cra-
nial caudal vertebrae Ca1 to Ca8 gave an indication of the likely placement
of the preceding recovered vertebrae. This line indicated that Gillette’s
(1991) placement of the articulated series of eight caudal vertebrae, num-
bered Ca20 to Ca27, was inconsistent with the progression of decrease in
the height of their neural spines. Thus, it was apparent these spines were
untenably high at these caudal assignments. The curve of the dorsal ridgeline
became consistent when this caudal series was assigned to positions Ca12
to Ca19.
Comparisons of caudal morphology with Diplodocus further indi-
cate that more cranial caudal assignments for the above-mentioned articu-
lated series of eight vertebrae is consistent with known diplodocid mor-
phology and supports the views of Curtice (1996, p. 24-26), who states
that in known diplodocids, caudal ribs are not known beyond caudal verte-
brae Ca20, and as the second to the last caudal possesses caudal ribs, the
caudal assigned position Ca26 by Gillette (1991), should be moved for-
ward to approximately Ca18. Pneumatic fossae are evident in all recovered
caudal vertebrae in the caudal series, including the second to the last caudal
vertebra, which Gillette (1991) positioned as C26. As the caudal-most oc-
currence of pneumatic fossae in any previously described sauropod is on
the nineteenth caudal, this vertebra is unlikely to have been located at
assignment 26, and moving the caudals forward five or six positions places
Seismosaurus hallorum in line with all known diplodocids.
In our reconstruction, the longest caudal, numbered by Gillette
(1991) as Ca24, with a centrum length of 362 mm, is assigned position
Ca16. The longest caudal in well known diplodocids is at Ca18, or adja-
cent to this number (Hatcher, 1901; Gilmore, 1932, 1936; Curtice, 1996).
Our placement at caudal Ca16, although only marginally longer than the
more cranially and terminally articulating caudals (Table 2), is congruent
with neural spine height at this location.
Comparisons of the percentage decrease in succeeding caudal cen-
tra widths, using measurements from Diplodocus longus (from Osborn,
1899; Gilmore, 1932) and Diplodocus carnegii (from Hatcher, 1901),
indicate that the widths of proximal to mid-caudal centra decrease at vary-
ing rates of between 2 and 12 percent, with an average of about 3 to 4
percent. Our suggested placements for the caudal vertebrae of Seismosaurus
hallorum fall within this range.
The incomplete caudal described as Ca13 by Gillette (1991) was
assigned position Ca11 in our reconstruction. Although the neural spine is
T1 *525 310 370 255 355 400 *680 845 490
T2 416 315 360 233 355 410 691 850 530
T3 300 326 310 330 300 311 350 410 1010 722 885 580
T4 340 *240 260 290 250 *330 315 400 1020 *770 995 720
T5 300 *245 260 290 300 300 335 410 1080 *845 1060 880
T6 320 255 270 280 380 280 335 400 1150 793 1275 1060
T7 340 260 260 390 335 410 1130 1340 1200
T8 300 275 275 300 400 309 350 410 1260 847 1350 1290
T9 290 *245 255 250 400 *310 315 410 1270 *1028 1310
T10 C. 280 *290 240 250 400 *325 365 400 C. 1300 *1070 1340
CA1 C. 200 183 240 192 C. 410 334 300 390 1260 1049 1250 1130
CA2 C. 220 205 220 290 380 1220 995 1120 1010
CA3 215 230 332 280 360 897 1020 920
CA4 C. 240 250 190 230 C. 380 330 280 350 1100 830 920 870
CA5 C. 280 250 190 220 C. 380 325 330 330 1020 777 830 760
CA6 C. 260 237 175 210 C. 360 309 250 310 1000 744 765 680
CA7 285 237 190 200 C. 360 317 260 300 C. 980 690 675 630
CA8 240 246 185 200 309 245 290 675 625 530
CA9 270 185 200 300 240 290 651 575 530
CA10 269 190 200 295 240 290 610 530 500
CA11 335 269 **332 285 610
CA12 330 295 190 190 **285 272 235 280 790 576 500 520
CA13 200 200 220 260 475 470
CA14 310 195 200 215 250 740 470 470
CA15 360 200 210 190 220 715 420
CA16 365 200 220 **220 190 220 685 410
CA17 **350 220 210 21 0 660 380
CA18 **338 205 220 200 170 210 **630 360
CA19 320 200 330 200 170 200 C. 450 330
Centrum greatest length
Centrum greatest width at caudal end Vertebrae total height
TABLE 2. Seismosaurus hallorum NMMNH P 3690, Diplodocus carnegii CM 84, Apatosaurus louisae CM 3018 and A. excelsus FM 7163,
dimensions of T, thoracic and Ca, caudal vertebrae in mm.
1CM 84 2CM
missing, morphological characters, such as centrum width, pleurocoel
development and the height, inclination and projection of the
prezygapophyes and postzygzpophyses, are consistent with Diplodocus
longus (Osborn, 1899). The oblique, ventrally projecting, cranial ar-
ticulating surface of the centrum Ca11, illustrated by Gillette (1991,
fig. 4), was not discernable by us on the fossil. Hence, the tail in our
reconstruction (Fig. 1) was not kinked ventrally at this caudal position.
The number of caudal vertebrae is variable within and between sau-
ropod species and may reach more than 80 for diplodocids (McIntosh, 1990).
The recorded number of caudal vertebrae for Apatosaurus louisae is 83,
with approximately 40 forming a so-called whiplash (Gilmore, 1936; McIn-
tosh, 1990). In our reconstruction of Seismosaurus hallorum, the number
of caudal vertebrae was kept at 75, with approximately 30 rod-like termi-
nal vertebrae forming the characteristic diplodocid whiplash, in parity with
the number assigned to Diplodocus composites (from Holland, 1906). If
the number were 80, the tail length would not be significantly increased.
From the caudal series Ca1 to Ca19, the only vertebrae not accounted
for in Seismosaurus hallorum are Ca9 and Ca10, a view supporting
Curtice (1996). Missing or partially missing vertebrae were sculpted
with morphological reference to Diplodocus carnegii, while maintain-
ing the upright neural spine condition evident in the preserved caudal
vertebrae of Seismosaurus hallorum. A similar, upright neural spine
condition is reported by Gilmore (1936) to be apparent in some caudal
vertebrae of Diplodocus longus (USNM 10865).
Gillette (1991, figs. 9-11) identified three chevrons of
Seismosaurus hallorum. Two of these are dorsoventrally elongated, with
paddle-shaped ends. These chevrons were positioned as Ca2/Ca3 and
Ca4/Ca5 and match the positions described in Gillette (1991). The other
chevron is represented by one half of the pair, which is caudoterminally
elongated and asymmetrical. We believe this bone is part of a rib, not a
chevron. Reconstruction of the missing chevrons used the recovered
types as a guide, together with those of Diplodocus. The last sculpted
chevron was at Ca24/Ca25.
C. estimated; 1dimensions from Hatcher (1901); 2dimensions from Gilmore (1936); *Peterson’s dimensions from Lull (1919); **dimensions from Gillette (1991)
Ilium, Pubis and Ischium
Only a partial right ilium was recovered during the NMMNH
excavations. It is attached to the sacrum and is missing the cranial
portion of the iliac blade and the distal extremity of the post-acetabular
process. The lateral dimensions of the ilium are strikingly similar to
those of Apatosaurus louisae CM 3018 (Gilmore, 1936, p. 229; for
comparisons see Table 2).
The cranial transverse width between the left and right ilia of
Seismosaurus hallorum is likely to have been broader than for A. louisae
by approximately 28 percent (using estimates from A. excelsus, Gilmore,
1936, fig. 31; Table 2). The equivalent dimension for Diplodocus carnegii
CM 84 is comparable to Apatosaurus louisae (based on transverse dimen-
sions of CM 94 in Hatcher, 1901, fig. 9 and Gilmore, 1932, table 6). How-
ever, the sacral length and ilium lengths for D. carnegii are shorter than
those of A. louisae. The sacroiliac proportions of S. hallorum are, there-
fore, more like Diplodocus than Apatosaurus, with broad transverse to
short axial proportions (Table 2).
At 1010 mm, the pubic (dorsoventral) length for Seismosaurus
hallorum is less than Apatosaurus louisae, which has a length of 1190
mm (Gilmore, 1936). The pubic length of S. hallorum is comparable with
Diplodocus carnegii, which has a length of 1000 mm (Hatcher, 1901, and
in agreement with the description of Gillette, 1991). Hence, the relative
length of the pubis of S. hallorum against the proportions of the ilium and
sacrum are comparatively less than the same proportions in Apatosaurus
and Diplodocus. Accordingly, to articulate the relatively short pubes (the
real right and the mirror-sculpted left) of S. hallorum at their medial sym-
physes, the pubes were acutely inclined medioventrally. Although the dis-
tal pubic peduncle of the recovered right ilium escaped preservation, the
proximal portion does indicate this pronounced medioventral inclination.
The iliac pubic peduncles of Supersaurus vivianae appear similarly in-
The distal portion of the right ischium and partial shaft were de-
scribed by Gillette (1991). The missing proximal portion of the right is-
chium and complete left ischium were sculpted to correctly articulate with
the pubes and the partially eroded ischiac peduncle of the right ilium. After
reconstruction, the appearance of the proximal ischium was similar to
Supersaurus vivianae (Jensen, 1985), with the iliac process of the ischium
quite elongated. However, unlike the broad, uniform and upwardly curved
ischial shaft of Seismosaurus hallorum, the shaft of S. vivianae is straight
and narrows in midsection. The dorsolaterally projecting process of the
distal ischium of S. hallorum is also evident on the distal ischium of S.
vivianae (Jensen, 1985). However this process is more pronounced in S.
Limbs, Pes, Manus and Pectoral Girdle
During our reconstruction, no elements of the limbs, pes, manus or
scapulocoracoid had been recovered or attributed to the holotype, hence,
from the recovered material, the pelvic elements provided the most usable
dimensions for determining the hind limb proportions. From these mea-
surements (Table 3) femur length was estimated first. The lengths of all
other limb elements were calculated from femur length using ratios re-
ported in McIntosh (1990); relative measurements from Diplodocus
carnegii (Hatcher, 1901) and Apatosaurus louisae (Gilmore, 1936); and
photographs of SMM P 84.15.8 (Table 3).
Compared with Diplodocus carnegii, the relative overall heights of
Seismosaurus hallorum thoracic vertebrae are no greater than 30 percent
taller; the sacral spines and ilia are approximately 50 percent taller, and the
cranial caudal vertebrae about 35 percent taller; however, the pubis is com-
parable in length. Upon these dimensions, femur length was proportioned
at about 24 percent greater than the femur length of Diplodocus carnegii
CM 84 (Table 3), or 1900 mm long. This estimate of femoral length is
comparable with lengths recorded for Apatosarus (Gilmore, 1936) (Table
3). The total distance from the dorsal extremity of the sacral neural
spines, along the hind limb bones to the ground at the pes, measures
4.94 meters (Fig. 1)
During reconstruction of the mounted replicas, the partial femur
NMMNH P-25079 (Fig. 3A-B) was not considered to have been an ele-
ment of the holotype, NMMNH P-3906. This incomplete sauropod fe-
mur, found within 6 meters of the vertebral elements of Seismosaurus
hallorum, was the only other dinosaur bone from the site not included in
the holotype. Gillette (1991), however, does make mention of the recov-
ery of this bone, and it is likely that a shortfall of time and resources sent
this specimen into obscurity.
The femur is in three pieces, has eroded articular ends and is of the
same preservation as the remaining bones of the holotype of Seismosaurus
hallorum. We believe this femur is likely to belong to the same individual
as the holotype specimen of S. hallorum due to the taphonomic nature of
the site, close proximity, proportions when compared with the reconstructed
estimate and morphological characteristics.
The fourth trochanter of the femur is eroded from the mediocaudal
surface (Fig. 3A), although a prominent swelling is evident in lateral view
just above midway on the shaft (Fig. 3B). The location of the fourth tro-
chanter, the intercondylar groove and slight curvature of the shaft, indicate
it is a right femur. The proximal and distal maximum widths are at least
380 mm and length is at least 1680 mm. The shape in caudal view (Fig.
3A), although generally similar to Diplodocus, appears more expanded
about the medial deflection (Upchurch et al., 2004) of the femoral lateral
edge (previously called the third trochanter), not unlike Barosaurus lentus
(Lull, 1919, p. 36-37, fig. 9; McIntosh, 1990, p. 372, fig 16.16).
The reconstructed length of this femur (Fig. 2C), with the articular
ends added, is estimated at approximately 1800 mm and is compatible with
the size range of the other elements of the holotype. The overall dimen-
sions of the preserved shaft with reconstruction to the articulating ends
(Fig. 3C) are similar to those of Apatosaurus louisae and A. excelsus (Table
3) and, congruently, the dimensions of the ilia, acetabular length and dors-
oventral sacral dimensions of Seismosaurus hallorum and A. louisae are
Scapulae, coracoids, sternal plates, pes and manus were sculpted
using Diplodocus data for reference. Dimensions of the recovered right
scapulocoracoid of the diplodocid, Supersaurus vivianae (Jensen, 1985),
which measures 2.7 meters in length, were similarly used as reference, as
the size range of other S. vivianae elements, such as the pelves and caudal
vertebrae, appear comparable. From these dimensions, reconstruction of
the scapulocoracoid for Seismosaurus hallorum was estimated at 2.4 meters
in length and locates the distal end of the scapula over the fifth thoracic rib
(Fig. 1).
Estimates of length for Seismosaurus hallorum initially came from
an article and the book “Seismosaurus hallorum The Earth Shaker” by
Gillette (1991, 1994). In these publications, Gillette postulated the length
of Seismosaurus hallorum to have been between 25 and 52 meters. How-
ever, Gillette favored the higher estimates of length, of between 39 to 52
meters, using comparative allometric inference from length-to-height data
of vertebral and pelvic elements from Diplodocus.
In his estimates of the overall length of Seismosaurus hallorum,
Gillette used three criteria in particular, relative to Diplodocus:
1. An expectation of protracted overall length, reasoning that the
disproportionately extended heights of the caudal and sacral neural spines,
relative to their centra lengths, the lengths of the cranial chevrons and the
massive proportions of the pelvis, were indicative of an axial allometric
increase in the tail and neck and thus in overall length.
2. An articulated series of eight mid-caudal vertebrae, positions
Gillette assigned as 20 to 27, appeared disproportionately developed in their
neural spine height and gross morphology (at relatively terminal positions)
compared to known diplodocids and indicated a comparatively protracted
tail length.
3. The pubis and elongated chevrons appeared relatively longer in
axial lengths (presumably meaning anterio-posterior lengths), which
indicated, through allometric inference, that the tail and overall lengths
were likely protracted.
The first and third criteria are essentially based on data of the dors-
oventral lengths of elements, such as the neural spines, chevrons and pu-
bis. These data were then extrapolated to give the anteroposterior estimate
of lengths of vertebral elements, using allometric inference. In our opinion,
using dorsoventral lengths of pelvic and caudal vertebral elements to give
vertebral axial lengths, via allometric inference, is not viable.
With regard to the second criterion, our placement of the articulated
series of eight mid-caudal vertebrae is consistent with more cranial assign-
ments of these vertebrae (in agreement with Curtice, 1996) and at odds
with the more terminal position assignments suggested by Gillette (1991,
1994). More cranial placement of these vertebrae effectively reduces the
estimates of tail length. Paul (1988), Lucas (1993) and Curtice (1996)
questioned the extreme length estimates suggested by Gillette (1991,
1994). Lucas (1993) offered a more conservative estimate of around 30
Estimating the overall axial length for Seismosaurus hallorum,
or any incomplete vertebral series, will always be constrained by the
lack of necessary vertebral lengths, vertebral numbers and the unknown
intervertebral distance between artculating centra. This intervertebral
distance, which constitutes the cartilaginous disc or synovium forming
the actual articulation, can be variable between procoelous,
amphicoelous and opisthocoelous forms in extant reptiles (Hoffstetter
and Gasc, 1969) and as likely, extinct reptiles. Research in this area is
needed for sauropods. Over the approximately 105 articulated verte-
brae for diplodocids, excluding the sacrum, every 10 mm error in the
calculated intervertebral joint can change the overall length estimate
by about a meter, which for discussion of the length of extremely long
dinosaurs is not very significant. The intervertebral length for
amphicoelous extant crocodiles, measured by Hoffstetter and Gasc
(1969), is around 11.5 percent. This percentage translated to sauropod-
sized vertebrae, indicates that for a caudal vertebrae of 300 mm, the
intervertebral length is approximately 33 mm and for a cervical of 700
mm length, about 77 mm. In our reconstruction, the longer vertebrae were
given an intervertebral spacing of about 40 mm to 70 mm and less for
smaller (axially shorter) vertebrae.
Gillette (1994) suggested that relative to Diplodocus, the apparent
allometry in the lengths of the sacral and caudal neural spines and chev-
rons, against axial measures of their centra lengths, indicated a protracted
length in the tail and hence, a relative increase in axial length overall. This,
of course, cannot be determined with certainty; Curtice (1996), however,
argued that a relative increase in allometry in caudal neural spines could
imply functions other than an increase in axial allometry. We argue further
that the apparent allometric lengthening of dorsoventrally elongated ele-
ments, such as the neural spines and chevrons, are more likely to have a
biomechanical basis related to the disproportionate increases in mass as the
tail lengthens during growth.
Associated with a strictly proportional or isometric increase in the
size of an object, such as the sauropod tail, there is a disproportionate
relationship between increasing size to the resulting mass (Alexander,
2CM 3018
(2CM 563)
((2FM 7163))
1CM 84
(1CM 94)
Ilium overall length
Width across ilia cranially
Width across ilia caudally
IIium superior border height
above acetabulum
Acetabulum length
Sacral length
Sacral width cranially
Sacral width caudally
Sacral spine height above
superior iliac border
Sacral spine height above
superior acetabular border
Greatest sacral vertebral height
Femur length
Femur shaft narrowest
transverse width
Femur transverse width
at 3rd trochanter
Femur proximal transverse
Femur distal transverse width
C. 1400
c. 1600
c. 1500
c. 1100
c. 1400
c. 1200
c. 1300
c. 1800
c. 320
c. 413
c. 555
c. 515
TABLE 3. Seismosaurus hallorum NMMNH P 3690, Diplodocus
carnegii CM 84, CM 94, Apatosaurus louisae CM 3018, CM 563 and
A. excelsus FM 7163, dimensions of ilia, sacra and femora in mm.
c. 1300
c. 570
c. 460
c. 1080
c. 970
(c. 1000)
c. estimated; * estimated size of CM 3018, based on a 14 percent size increase from
CM 563; 1dimensions from Hatcher (1901); 2dimensions from Gilmore (1936)
FIGURE 3. Right sauropod femur, NMMNH P-25079. A, caudal view; B, medial
view; C, caudal view with the reconstructed outline of the articular ends. IG,
intercondylae groove; 3T, 3rd trochanter; 4T, 4th trochanter; GT, greater trochanter;
TC, tibial condyle; FC, fibular condyle; FH, femur head. Scale bar equals 0.5 meters.
1985; Wedel et al., 2000). Our calculations of volume for a typically
shaped sauropod tail indicate that an isometric increase in tail size of
20 percent is accompanied by a disproportionate increase in mass of
about 60 percent. With further increased axial allometry, the disparity
between tail size and mass would be even greater. Due to the biome-
chanical constraints imposed by the relationship between ligament forces
and ligament diameters, a disproportionate increase in tail mass to axial
length would require a relative increase in the length of the leverage
structure, such as the neural spines, which affect ligament suspension
and bracing. Lever calculations for ligament forces of a suspended ver-
tebral series show that these forces are reduced as neural spine height
increases (Alexander, 1985, 1997). As such, an increase in the height
of the sacral and caudal neural spines would be an adaptive response to
counter increased tail mass during growth, in both the ontogenetic sense
and phylogenetically.
In Seismosaurus hallorum, an increase in relative dorsoventral al-
lometry of the sacral and caudal bones, with respect to an axial increase in
tail length, supports this conclusion. Individual caudal centra of S. hallorum
show axial length increases of between 10 to 25 percent, when compared
with Diplodocus carnegii (CM 84: Gilmore, 1932). The heights of the
sacral, cranial caudal and mid-caudal neural spines show relative increases
of approximately 50 percent at the coalesced sacral spines, 32 percent at
caudal vertebra Ca5, Ca6 and Ca7 and less than 32 percent for the intact
mid-caudal vertebrae from Ca11 to Ca19 (Table 2). The ratio of total verte-
bral height to centrum length (Table 1, up to Ca12) better illustrates the
decreasing disparity, cranially to terminally, in relative caudal dimensions
between S. hallorum and D. carnegii. Similarly, the ratio of total vertebral
height to centrum length appears greater in S. hallorum than in Apatosaurus
for the recovered caudal vertebrae, although the disparity is greater in the
cranial caudals and increasingly less in caudals terminally. Increased axial
allometry in the tail of S. hallorum may have occurred relative to
Diplodocus, however, such allometry cannot be determined from the known
material and is unlikely to have been to the degree suggested by Gillette
(1991, 1994).
For Seismosaurus hallorum, the estimated length of the recovered
caudal series from caudal 1 to 19 is approximately 5.76 m, with the com-
plete caudal series of 75 vertebrae measuring 18.85 m; the total length of
the five fused sacral vertebrae is estimated at 1.10 m; the total thoracic
series of 10 vertebrae, including estimates for the length of 2 missing cra-
nial thoracic vertebrae, measures 3.78 m; the reconstruction of the 15 cer-
vical vertebrae measures 8.32 m and the skull 780 mm. The overall length
is estimated at 32.83 (33) meters. With an added 10 percent error on the
sculpted bones the total length would be 35 m and with negative 10 per-
cent error the length would be 30.5 m.
Although only 30 percent of the skeleton of Seismosaurus
hallorum was recovered by the NMMNH during the 1985 to 1990
excavations, the thoracic, pelvic and caudal elements gave adequate
osteological information and dimensions for comparisons with related
diplodocids. The initial descriptions of caudal and ventral pelvic bones
were published in Gillette (1991, 1994). However, subsequent prepa-
ration allowed further descriptive comparisons to be made and some
re-interpretation. Production of a three-dimensional mount gave a rare
opportunity to assess critically the reconstruction at full-scale, which,
in so doing, has allowed reinterpretation of the overall length of the
The undescribed femur, NMMNH P-25079, which was recov-
ered during the excavations of Seismosaurus hallorum within the quarry
site, was formerly disregarded as an element of the holotype. Com-
parative dimensions of this femur with the femur of Apatosaurus, which
also shares similar sized pelvic and thoracic dimensions, indicates that
this femur is congruent with the holotype. Femur size estimates calcu-
lated during the reconstruction of Seismosaurus hallorum further sup-
port our proposed inclusion of the femur in the holotype.
The morphological similarities in the thoracic vertebrae of
Seismosaurus hallorum and Diplodocus, such as their general proportions,
development of their pleurocoels and overall layout of their vertebral lamina,
reflects the undeniable taxonomic closeness of these two genera and begs
the question: If the thoracic vertebrae had been found in isolation would
assignment to the genus Diplodocus have been more appropriate?
The positions we assigned the middle caudal vertebrae were based
first on the natural taper formed by the medial dorsal ridge of the recovered
neural spines. Morphological similarities between comparative caudal ver-
tebrae, although important, were considered secondarily. We considered
placement of the intact mid-caudal vertebrae at the more terminal positions
proposed by Gillette (1991, 1994) to be disproportionate.
With regard to axial length, we support the views of Paul (1988),
Lucas (1993) and Curtice (1996) that the higher overall axial lengths pro-
posed by Gillette (1991, 1994) of greater than 39 meters are an overesti-
mation. Nevertheless, at 33 meters long, Seismosaurus hallorum remains
among the longest terrestrial animals to have existed.
From PAST we thank Keith Russell, Chris Robinson and Nigel Yez
for the construction of mounts; Janet Rosine and Wayne Marshall for as-
sistance with sculpting; Joe Barry and Don Jefferies for molding and cast-
ing; Colin Cooke for preparation; Gilles Danis and Connie Russell for project
management. We thank Roland Bouten for photography used in Fig. 1,
David Herne for computer support and Aurora Oval for their diverse sup-
port. From NMMNH we thank Adrian Hunt (Director), Andrew Heckert
and Peter Reser. We especially acknowledge David Gillette for his work
and his team of excavators and preparators. Thanks to Karen Cloward,
Scott Sampson and Ken Stadtman for access to collections in Utah. We
also thank Don Brinkman, Kenneth Carpenter, John McIntosh, Dan Miller,
Larry Rinehart, Justin Spielmann and Jeffrey Wilson for their useful input.
Brian Curtice and Jack McIntosh provided helpful reviews of the manu-
Alexander, R. McN., 1985, Mechanics of posture and gait of some large dinosaurs:
Zoological Journal of the Linnean Society of London, v. 83, p. 1-25.
Alexander, R. McN., 1997, Biomechanics; in Currie, P. J. and Padian, K., eds., En-
cyclopedia of Dinosaurs: San Diego, Academic Press, p. 57-59.
Curtice, B., 1996, Codex of diplodocid caudal vertebrae from the Dry Mesa Dino-
saur Quarry [M. S. thesis]: Provo, Brigham Young University, 250 p.
Frey, E. and Martin, J., 1997, Long necks of sauropods; in Currie, P. J. and Padian,
K., eds., Encyclopedia of Dinosaurs: San Diego, Academic Press, p. 407-409.
Gillette, D. D., 1991, Seismosaurus halli, gen. et sp. nov., a new sauropod dinosaur
from the Morrison Formation (Upper Jurassic/Lower Cretaceous) of New Mexico,
USA: Journal of Vertebrate Paleontology, v. 11, p. 417-433.
Gillette, D. D., 1994, Seismosaurus: The Earth Shaker. New York, Columbia
University Press, 205 p.
Gilmore, C. W., 1932, On a newly mounted skeleton of Diplodocus in the United
States National Museum: Proceedings of the U.S. National Museum, v. 81, p. 1-2.
Gilmore, C. W., 1936, Osteology of Apatosaurus with special reference to speci-
mens in the Carnegie Museum: Memoirs of the Carnegie Museum, v. 11, p.
Hatcher, J. B., 1901, Diplodocus Marsh, its osteology, taxonomy, and probable
habits, with a restoration of the skeleton: Memoirs of the Carnegie Museum, v.
1, p. 1-63.
Hoffstetter, R. and Gasc, J-P., 1969, Vertebrae and ribs of modern reptiles; in Gans,
C., d’A. Bellairs, A. and Parsons, T. S., eds., Biology of the reptilia, volume 1:
London and New York, Academic Press, p. 201-310
Holland, W. J., 1906, The osteology of Diplodocus Marsh: Memoirs of the Carnegie
Museum, v. 2, p. 225-264.
Holland, W. J., 1934, The skull of Diplodocus: Memoirs of the Carnegie Museum,
v. 9, p. 379-403.
Jensen, J. A., 1985, Three new sauropod dinosaurs from the Upper Jurassic of Colo-
rado: Great Basin Naturalist, v. 45, p. 697-709.
Lucas, S. G., 1993, Dinosaurs of New Mexico. Albuquerque, New Mexico Acad-
emy of Science, 130 p.
Lucas, S. G., Herne, M. C., Heckert, A. B., Hunt, A. P. and Sullivan, R. M., 2004,
Reappraisal of Seismosaurus, a Late Jurassic sauropod dinosaur from New
Mexico: Geological Society of America, Abstracts with Programs, v. 36, no. 5,
p. 422.
Lull, R. S., 1919, The sauropod dinosaur Barosaurus Marsh: Memoirs of the Con-
necticut Academy of Arts and Science, v. 6, p. 1-42.
McIntosh, J. S., 1981, Annotated catalogue of the dinosaurs (Reptilia, Archosauria)
in the collections of the Carnegie Museum of Natural History: Bulletin of the
Carnegie Museum of Natural History, v. 18, p. 1-67.
McIntosh, J. S., 1990, Sauropoda; in Weishampel, D. B., Dodson, P. and Osmólska,
H., eds., The Dinosauria: Berkeley, University of California Press, p. 345-401.
Osborn, H. F., 1899, A skeleton of Diplodocus: Memoirs of the American Museum
of Natural History, v. 1, p. 191-214.
Paul, G. S., 1988, The brachiosaur giants of the Morrison and Tendaguru with a
description of a new subgenus, Giraffatitan, and a comparison of the world’s
largest dinosaurs: Hunteria, v. 2, no. 3, p. 1-14.
Stevens, K. A. and Parrish, J. M.,1999, Neck posture and feeding habits of two
Jurassic sauropod dinosaurs: Science, v. 284, p. 798-800.
Upchurch, P., Barrett, P. M. and Dodson, P. 2004. Sauropoda; in Weishampel, D.
B., Dodson, P. and Osmólska, H., eds., The Dinosauria: Berkeley, University of
California Press, p. 259-322.
Wilson, J. A., 1999, A nomenclature for vertebral laminae in sauropods and other
saurischian dinosaurs: Journal of Vertebrate Paleontology, v. 19, p. 639-653.
Wedel, M. J., Cifelli, R. L. and Sanders, R. K., 2000, Osteology, paleobiology, and
relationships of the sauropod dinosaur Sauroposeidon: Acta Palaeontologica
Polonica, v. 45, p. 343-388.
... Although sauropods were large animals in general, it is important to point out that extreme sizes (close to or in excess of 40 t) were reached independently by several different lineages of sauropods at different times throughout the later Mesozoic (Fig. 6). Specific cases are the Late Jurassic (Kimmeridgian) basal eusauropod Turiasaurus (Royo-Torres et al., 2006), possibly the basal diplodocoid Amphicoelias (Carpenter, 2006), the Late Jurassic (Tithonian) Diplodocus (Seismosaurus) hallorum (Gillette, 1991(Gillette, , 1994Herne & Lucas, 2006) and 'Supersaurus' (Upchurch et al., 2004) among the Diplodocoidea, the Early Cretaceous (Aptian) brachiosaurid Sauroposeidon (Wedel et al., 2000a, b), and several titanosaurs. The latter include Paralititan from the early Late Cretaceous (Cenomanian) of Egypt (Smith et al., 2001), as well as Argentinosaurus (Bonaparte & Coria, 1993;Mazzetta et al., 2004), Puertasaurus (Novas et al., 2005), Antarctosaurus giganteus (Van Valen, 1969;Mazzetta et al., 2004), and Futalognkosaurus (Calvo et al., 2007) from the Late Cretaceous of Argentina. ...
... Although sauropods were large animals in general, it is important to point out that extreme sizes (close to or in excess of 40 t) were reached independently by several different lineages of sauropods at different times throughout the later Mesozoic (Fig. 6). Specific cases are the Late Jurassic (Kimmeridgian) basal eusauropod Turiasaurus (Royo-Torres et al., 2006), possibly the basal diplodocoid Amphicoelias (Carpenter, 2006), the Late Jurassic (Tithonian) Diplodocus (Seismosaurus) hallorum (Gillette, 1991(Gillette, , 1994Herne & Lucas, 2006) and 'Supersaurus' (Upchurch et al., 2004) among the Diplodocoidea, the Early Cretaceous (Aptian) brachiosaurid Sauroposeidon (Wedel et al., 2000a, b), and several titanosaurs. The latter include Paralititan from the early Late Cretaceous (Cenomanian) of Egypt (Smith et al., 2001), as well as Argentinosaurus (Bonaparte & Coria, 1993;Mazzetta et al., 2004), Puertasaurus (Novas et al., 2005), Antarctosaurus giganteus (Van Valen, 1969;Mazzetta et al., 2004), and Futalognkosaurus (Calvo et al., 2007) from the Late Cretaceous of Argentina. ...
... Although sauropods were large animals in general, it is important to point out that extreme sizes (close to or in excess of 40 t) were reached independently by several different lineages of sauropods at different times throughout the later Mesozoic (Fig. 6). Specific cases are the Late Jurassic (Kimmeridgian) basal eusauropod Turiasaurus (RoyoTorres et al., 2006), possibly the basal diplodocoid Amphicoelias (Carpenter, 2006), the Late Jurassic (Tithonian) Diplodocus (Seismosaurus) hallorum (Gillette, 1991Gillette, , 1994 Herne & Lucas, 2006) and 'Supersaurus' (Upchurch et al., 2004) among the Diplodocoidea, the Early Cretaceous (Aptian) brachiosaurid Sauroposeidon (Wedel et al., 2000a, b), and several titanosaurs. The latter include Paralititan from the early Late Cretaceous (Cenomanian) of Egypt (Smith et al., 2001), as well as Argentinosaurus (Bonaparte & Coria, 1993; Mazzetta et al., 2004), Puertasaurus (Novas et al., 2005), Antarctosaurus giganteus (Van Valen, 1969; Mazzetta et al., 2004), and Futalognkosaurus (Calvo et al., 2007) from the Late Cretaceous of Argentina. ...
... Although sauropods were large animals in general, it is important to point out that extreme sizes (close to or in excess of 40 t) were reached independently by several different lineages of sauropods at different times throughout the later Mesozoic (Fig. 6). Specific cases are the Late Jurassic (Kimmeridgian) basal eusauropod Turiasaurus (RoyoTorres et al., 2006), possibly the basal diplodocoid Amphicoelias (Carpenter, 2006), the Late Jurassic (Tithonian) Diplodocus (Seismosaurus) hallorum (Gillette, 1991Gillette, , 1994 Herne & Lucas, 2006) and 'Supersaurus' (Upchurch et al., 2004) among the Diplodocoidea, the Early Cretaceous (Aptian) brachiosaurid Sauroposeidon (Wedel et al., 2000a, b), and several titanosaurs. The latter include Paralititan from the early Late Cretaceous (Cenomanian) of Egypt (Smith et al., 2001), as well as Argentinosaurus (Bonaparte & Coria, 1993; Mazzetta et al., 2004), Puertasaurus (Novas et al., 2005), Antarctosaurus giganteus (Van Valen, 1969; Mazzetta et al., 2004), and Futalognkosaurus (Calvo et al., 2007) from the Late Cretaceous of Argentina. ...
Full-text available
The herbivorous sauropod dinosaurs of the Jurassic and Cretaceous periods were the largest terrestrial animals ever, surpassing the largest herbivorous mammals by an order of magnitude in body mass. Several evolutionary lineages among Sauropoda produced giants with body masses in excess of 50 metric tonnes by conservative estimates. With body mass increase driven by the selective advantages of large body size, animal lineages will increase in body size until they reach the limit determined by the interplay of bauplan, biology, and resource availability. There is no evidence, however, that resource availability and global physicochemical parameters were different enough in the Mesozoic to have led to sauropod gigantism.
Full-text available
The fusion of the sacrum occurs in the major dinosaur lineages, i.e. ornithischians, theropods, and sauropodomorphs, but it is unclear if this trait is a common ancestral condition, or if it evolved independently in each lineage, or even how or if it is related to ontogeny. In addition, the order in which the different structures of the sacrum are fused, s well as the causes that lead to this co-ossification, are poorly understood. Herein, we escribed the oldest record of fused sacral vertebrae within dinosaurs, based on two primordial sacral vertebrae from the Late Triassic of Candelária Sequence, southern Brazil. We used computed microtomography (micro-CT) to analyze the extent of vertebral fusion, which revealed that it occurred only between the centra. We also assessed the occurrence of sacral fusion in Dinosauria and close relatives. The degree of fusion observed in representatives of the major dinosaur lineages suggested that there may be a sequential pattern of fusion of the elements of the sacrum, more clearly observed in Sauropodomorpha. Our analyses suggest that primordial sacral vertebrae fuse earlier in the lineage (as seen in Norian sauropodomorphs). Intervertebral fusion is observed to encompass progressively more vertebral units as sauropodomorphs evolve, reaching up to five or more fully fused sacrals in Neosauropoda. Furthermore, the new specimen described here indicates that the fusion of sacral elements occurred early in the evolution of dinosaurs. Factors such as ontogeny and the increase in body size, combined with the incorporation of vertebrae to the sacrum may have a significant role in the process and in the variation of sacral fusion observed.
Full-text available
We provide an annotated listing and bibliography of the type (holotype and paratype) specimens of fossil vertebrates in the New Mexico Museum of Natural History paleontology collection as of 2015. The collection includes 129 primary type specimens—81 holotypes and 48 paratypes.
Full-text available
New Mexico has a relatively sparse Jurassic record of fossil vertebrates, much less than is known from either the Triassic or the Cretaceous strata in the state. The oldest Jurassic vertebrates from New Mexico are the osteichthyans Hulettia americana, Todiltia schoewei and Caturus dartoni from the Middle Jurassic (Callovian) Luciano Mesa Member of the Todilto Formation. The overlying Callovian-Oxfordian? Summerville Formation has yielded fragmentary sauropod dinosaur bones and teeth assigned to Camarasaurus, and theropod footprints identi ed as Megalosauripus and cf. Therangospodus. Most of New Mexico’s Jurassic vertebrate fossils are from the Upper Jurassic (Kimmeridgian-Tithonian?) Brushy Basin Member of the Morrison Formation and include the turtle Glyptops, the theropod dinosaurs Allosaurus and Saurophaganax, the ornithischian Stegosaurus and (mostly) sauropod dinosaurs identi ed as Apatosaurus, Brachiosaurus and Diplodocus (= “Seismosaurus”). The sparse record of Jurassic vertebrate fossils in New Mexico is partly due to the extensive eolian and evaporitic facies in parts of the Jurassic section, but mostly to a relative lack of effort to explore the Jurassic strata in New Mexico for vertebrate fossils.
Full-text available
Diplodocidae are among the best known sauropod dinosaurs. Several species were described in the late 1800s or early 1900s from the Morrison Formation of North America. Since then, numerous additional specimens were recovered in the USA, Tanzania, Portugal, and Argentina, as well as possibly Spain, England, Georgia, Zimbabwe, and Asia. To date, the clade includes about 12 to 15 nominal species, some of them with questionable taxonomic status (e.g., ‘Diplodocus’ hayi or Dyslocosaurus polyonychius), and ranging in age from Late Jurassic to Early Cretaceous. However, intrageneric relationships of the iconic, multi-species genera Apatosaurus and Diplodocus are still poorly known. The way to resolve this issue is a specimen-based phylogenetic analysis, which has been previously implemented for Apatosaurus, but is here performed for the first time for the entire clade of Diplodocidae. The analysis includes 81 operational taxonomic units, 49 of which belong to Diplodocidae. The set of OTUs includes all name-bearing type specimens previously proposed to belong to Diplodocidae, alongside a set of relatively complete referred specimens, which increase the amount of anatomically overlapping material. Non-diplodocid outgroups were selected to test the affinities of potential diplodocid specimens that have subsequently been suggested to belong outside the clade. The specimens were scored for 477 morphological characters, representing one of the most extensive phylogenetic analyses of sauropod dinosaurs. Character states were figured and tables given in the case of numerical characters. The resulting cladogram recovers the classical arrangement of diplodocid relationships. Two numerical approaches were used to increase reproducibility in our taxonomic delimitation of species and genera. This resulted in the proposal that some species previously included in well-known genera like Apatosaurus and Diplodocus are generically distinct. Of particular note is that the famous genus Brontosaurus is considered valid by our quantitative approach. Furthermore, “Diplodocus” hayi represents a unique genus, which will herein be called Galeamopus gen. nov. On the other hand, these numerical approaches imply synonymization of “Dinheirosaurus” from the Late Jurassic of Portugal with the Morrison Formation genus Supersaurus. Our use of a specimen-, rather than species-based approach increases knowledge of intraspecific and intrageneric variation in diplodocids, and the study demonstrates how specimen-based phylogenetic analysis is a valuable tool in sauropod taxonomy, and potentially in paleontology and taxonomy as a whole.
Full-text available
In birds, diverticula of the lungs and air sacs pneumatize specific regions of the postcranial skeleton. Relationships among pulmonary components and skeletal regions they pneumatize allow inferences about pulmonary anatomy in non-avian dinosaurs. Fossae, foramina and chambers in the postcranial skeletons of pterosaurs and saurischian dinosaurs are diagnostic for pneumaticity. In basal saurischians only the cervical skeleton is pneumatized, by cervical air sacs. In more derived saurischians pneumatization of posterior dorsal, sacral, and caudal vertebrae indicates abdominal air sacs. Abdominal air sacs in sauropods are also indicated by a pneumatic hiatus (a gap in vertebral pneumatization) in Haplocanthosaurus. Minimally, saurischians had dorsally attached diverticular lungs plus anterior and posterior air sacs, and all the pulmonary prerequisites for flow-through lung ventilation like that of birds. Pneumaticity reduced skeletal mass in saurischians. I propose the Air Space Proportion (ASP) as a measure of proportional volume of air in pneumatic bones. The mean ASP of a sample of sauropod and theropod vertebrae is 0.61, so on average, air occupied more than half the volume of these vertebrae. In Diplodocus, pneumatization lightened the living animal by 7-10 percent, and that does not include extraskeletal diverticula, air sacs, lungs, or trachea. If all these air reservoirs included, the specific gravity of Diplodocus is 0.80, higher than published values for birds but lower than those for squamates and crocodilians. Pneumatization of cervical vertebrae facilitated evolution of long necks in sauropods. Necks longer than nine meters evolved at least four times, in mamenchisaurs, diplodocids, brachiosaurids, and titanosaurs. Increases in the number of cervical vertebrae, their proportional lengths, and their internal complexity occurred in parallel in most of these lineages.
Full-text available
Sauroposeidon proteles is a large brachiosaurid sauropod recently described from the Antlers Formation (Aptian-Albian) of southeastern Oklahoma. Sauroposeidon represents the culmination of brachiosaurid trends toward lengthening and lightening the neck, and its cervical vertebrae are characterized by extensive pneumatic structures. The elaboration of vertebral air sacs during sauropod evolution produced a variety of internal structure types. We propose a new classification system for this array of vertebral characters, using computed tomography (CT) of pneumatic internal structures. Comparisons with birds suggest that the vertebrae of sauropods were pneumatized by a complex system of air sacs in the thorax and abdomen. The presence of a thoraco-abdominal air sac system in sauropods would dramatically affect current estimates of mass, food intake, and respiratory requirements. Sauroposeidon was one of the last sauropods in the Early Cretaceous of North America; sauropods disappeared from the continent by the early Cenomanian. The demise of sauropods in the Early Cretaceous of North America predates significant radiations of angiosperms, so the decline and extinction of this dinosaur group cannot be linked to changes in flora.