Epaxial muscles and ossified tendons in dinosaurs : anatomy, development, histology, and
biomechanics
by Christopher Lee Organ
A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of
Philosophy in Biological Sciences
Montana State University
© Copyright by Christopher Lee Organ (2003)
Abstract:
Intratendinous ossification in dinosaurs (including birds) is a wide spread phenomenon that has
implications for physiology, posture, biomechanics, behavior, and systematics. Four separate studies
were undertaken to elucidate ossified tendon biology in extinct dinosaurs and birds. The first
investigation reconstructs dorsal epaxial musculature in extinct dinosaurs. Crocodilian and avian dorsal
epaxial muscles are homologized. Using the extant phylogenetic bracketing approach, the three-layered
trellis of Hadrosauriformes is homologized to the M. transversospinalis slips in crocodilians and the M.
longus colli dorsalis thoracica in birds. These tendons are ossified in Hadrosauriformes and some birds.
The parallel ossified tendons in other ornithischians are homologized to the M. longissimus dorsi in
Alligator. The second investigation determines how ossified tendons in non-avian dinosaurs developed.
Atrophied muscle was previously thought to be the origin of ossified tendons in non-avian dinosaurs.
Using different age classes of hadrosaurs (Brachylophosaurus and Maiasaura) and turkeys (Meleagris),
the developmental process of ossified tendons in hadrosaurs is shown to be homologous with
intratendinous ossification in birds. But, the degree of ossification is greater in hadrosaurs, whose
tendons also possess primary osteons, an external fundamental system and lines of arrested growth.
The third investigation determines the histological diversity of ossified tendons in Dinosauria. Despite
various anatomical locations and large differences in size, ossified tendons have uniform
microstructure even in specimens that do not normally experience intratendinous ossification (such as
Ceratosaurus and Camarosaurus). The greater degree of ossification noted above occurs in all
non-avian dinosaurs. Also, variation in periosteal bone development occurs along the length of
individual tendons. Ossified tendons from marginocephalians are unique in that they have larger
portions of fibrolamellar bone and radial vascularity. In the fourth investigation, computer finite
element models are constructed for two ornithopods to assess ossified tendon biomechanics. An
Alligator is used to empirically determine joint properties. Ossified tendons probably reduced tail
deflection and played a role in locomotion in all ornithopods. The ossified tendon trellis of
Hadrosauriformes may have increased the skeleton's ability to bear large body mass. Neural spine
height and bone material properties have the greatest impact on spinal rigidity. EPAXIAL MUSCLES AND OSSIFIED TENDONS IN DINOSAURS:
ANATOMY, DEVELOPMENT, HISTOLOGY, AND BIOMECHANICS
by
Christopher Lee Organ
A dissertation submitted in partial fulfillment
of the requirements for the degree
! ■ of
Doctor of Philosophy
in
Biological Sciences
MONTANA STATE UNIVERSITY
Bozeman, Montana
December 2003
© COPYRIGHT
by.
Christopher Lee Organ
2003
All Rights Reserved .
0T 7f
ii
Or 35
APPROVAL
of a dissertation submitted by
Christopher Lee Organ
This dissertation has been read by each member of the dissertation committee and has
been found to be satisfactory regarding content, English usage, format, citations, biblio­
graphic style, and consistency, and is ready for submission to the College of Graduate
Studies.
John R. Homer
Dcite
Kevin M. O'Neill
Date
Approved for the Department of Cell Biology and Neuroscience
Gwen A. Jacobs
re)
Approved for the College of Graduate Studies
Bruce R. McLeod
(Signature)
Date
iii
STATEMENT OF PERMISSION TO USE
In presenting this dissertation in partial fulfillment of the requirements for a doc­
toral degree at Montana State University, I agree that the Library shall make it available
to borrowers under rules of the Library. I further agree that copying of this dissertation
is allowable only for scholarly purposes, consistent with "fair use" as prescribed in the
U.S. Copyright Law. Requests for extensive copying or reproduction of this dissertation
should be referred to Bell & Howell Information and Learning, 300 North Zeeb Road, Ann
Arbor, Michigan 48106, to whom I have granted "the exclusive right to reproduce and
distribute my dissertation in and from microform along with the non-exclusive right to
reproduce and distribute my abstract in any format in whole or in part."
Signature _
D ate
/2 --£ > Q --
iv
ACKNOWLEDGEMENTS
I would like to thank Paul Sereno, William Simpson, Janet Hinshaw, and Jim Gard­
ner for access and discussion over specimens. I would also like to thank Robert Storer for
insightful discussions on intratendinous ossification in birds. Seder Ridge Turkey Farm,
Ruth Elsey at the Rockefeller Wildlife Refuge (Louisiana Department of Wildlife and Fish­
eries), and the Bozeman Raptor Center kindly donated extant specimens.
Jack Horner, Cynthia Marshall, Armand de Ricqles Tobin Hieronymus, Jason Ad­
ams, Nicole Hobbs, Sean Paul, Ellen Lamm, Jim Dent, Rick Blob, Joe Cooley, Joe Beaman,
Jeff LaRock, Martha Middlebrooks, Katie Organ, R. McNeil Alexander, Steve Gatesy, Matt
Carrano, John Hutchinson, Walter Coombs, and others whom I may have forgotten de­
serve thanks for discussions or reviews of my project, though all errors are certainly mine
to bear.
Matt Carrano gave me advice on translating Old German and French papers, for
which I am grateful. Also, I must thank Erich Staudacher for helping me figure out some
inscrutable Old German. Neither are responsible for possible errors in translation.
Funding was provided in part by a 2001 doctoral grant from the International So­
ciety of Biomechanics, the Museum of the Rockies (Department of Paleontology), and the
Department of Cell Biology and Neuroscience at Montana State University. I would par­
ticularly like to thank Gwen Jacobs for the support and opportunities she has graciously
given me.
I am grateful to my committee members Jack Horner, Kevin O'Neill, Matt Lavin,
Gwen Jacobs, and Deb King for their guidance and friendship.
Finally, I am indebted to my parents William Organ and Susan Hadley for inspir­
ing my passion for science and creativity and to my love, Nicole Hobbs, who has stood
with me in good times and bad.
V
TABLE OF CONTENTS
LIST OF FIGURES............................................................................................................. VII
LIST OF TABLES..................................................................
IX
ABSTRACT............................................................................................................................ X
1. DORSAL EPAXIAL MUSCULATURE IN D IN O SA U RS......................................... I
In tro d u c tio n ............................................................................................................................I
M aterials and M e th o d s ..............................................................................................
3
H o m o lo g y .................................................................................................................. 4
Phylogenetic Fram ew ork for Soft Tissue Inferences......................................... 5
R esults...................................................................................................................................... 7
C ro c o d ilia .................................................................................................................. 7
N e o rn ith e s ............................................................................................................... 10
M uscle R econstruction in Extinct D in o sau rs................................................................ 16
D iscu ssio n .............................................................................................................................20
C onclusions...................................................................................................
22
2. DEVELOPMENTAL PALEOBIOLOGY OF OSSIFIED TENDONS
IN H A D RO SA U RS............................................................................................................. 24
In tro d u c tio n ..........................................................................................................................24
M aterials and M e th o d s ......................................................................................................26
H istological D escription.......................:............................................................................28
Turkey Tendon H isto lo g y ..................................................................................... 28
H ad ro sau r Tendon H isto lo g y ..............................................................................29
D iscu ssio n .............................................................................................................................32
C om parative H istologyi........................................................................................ 32
LAGs and the EFS...................................................................................................35
A d a p t a t i o n ...........................................................................................................39
C onclusions....................
40
3. HISTOLOGY OF OSSIFIED TENDONS IN DINOSAURS.................................... 43
In tro d u c tio n ..........................................................................................................................43
vi
TABLE OF CONTENTS - CONTINUED
M aterials and M e th o d s ................................................................................
44
C om parative H istological D escrip tio n ...........................................................................47
C rocodilian Tendon H istology.............................................................................47
O rnithischian Tendon H isto lo g y .........................................................................47
Saurischian Tendon H isto lo g y .............................................................................53
D iscu ssio n .........................................................................................................
60
Ossified Tendon D iversity and E v o lu tio n ......................................................... 60
Tail Rod Tendons.....................................................................................................63
C o n clu sio n ............................................................................................................................64
4. OSSIFIED TENDON BIOMECHANICS IN ORNITHOPOD
D IN O SA U R S....................
67
In tro d u c tio n ..........................................................................................................................67
M aterials and M e th o d s ......................................................................................................71
M aterial P ro p e rtie s................................................................................................ 75
U ltim ate S tresses.....................................................................................................76
A lligator Results and Partially R estrained Joint Properties.......................... 78
R esults.................................................................................................................................... 80
Tenontosaur M odel Results .................................................................................. 80
B rachylophosaur M odel R esults..........................................................................85
D iscu ssio n .........................................................................................................
89
Evaluating the Fem oral and H ogging H ypotheses......................................... 91
C onclusions.................................................................................. .....................;................. 94
APPENDIX: M ORPHOLOGIC DATA FOR CHAPTER 4 MODELS ...................... 97
REFERENCES CITED.......................................................................................................103
vii
LIST OF FIGURES
Figure
Page
1. Phylogenetic fram ew ork used in chapter 1...........................................................6
2. D iagram of dorsal epaxial m uscles in Alligator................................................. 8
3. D iagram of m ajor dorsal epxial m uscles in Meleagris...................................... 11
4. D iagram of the M. Iongus colli dorsalis thoracica slips in b ird s.................... 12
5. O ssified epaxial tendons in dinosaurs................................................................. 17
6. Phylogenetic fram ew ork u sed in chapter 2.........................................................27
7. H istological photographs o f Meleagris tendons................................................. 29
8. H istological photographs of h ad ro sau r ten d o n s............................................... 31
9. Juxtaposed histological im ages of Meleagris tendons and
Brachylophosaurus te n d o n s .......................................................................................... 34
10. A d u lt h ad ro sau r tendons w ith LAGs..... ........................................................... 37
11. Phylogenetic fram ew ork u sed in chapter 3 ...................................................... 45
12. H istological photographs of Alligator te n d o n s.................................................48
13. H istological photographs of H adrosauriform es te n d o n s ............................. 49
14. H istological photographs of Tenontosaurus tendons....................................... 50
15. H istological photographs of M arginocephalia te n d o n s................................. 52
16. H istological photographs of Euplocephalus tendons........................................ 53
viii
LIST OF FIGURES - CONTINUED
17. H istological photographs of Camarosaurus ten d o n s........................................54
18. H istological photographs of non-avian theropod te n d o n s........................... 55
19. H istological photographs of th ero p o d "tail ro d s " .......................................... 57
20. H istological photographs of N eornithes ossified ten d o n s....................... .....58
21. R econstruction of Brachylophosaurus and Tenontosaurus.................................68
22. Trellis of ossified tendons from Brachylophosaurus.......................................... 69
23. Phylogenetic fram ew ork used in chapter 4.......................................................70
24. D iagram of a m odeled vertebra...........................................................................73
25. D iagram of the Alligator m odel w ith endzones.........-...................................... 80
26. Tenontosaur m odel deflection by lo ad ...............................................................82
27. Tenontosaur m odel bending stress b y lo ad ...................................................... 82
28. Tenontosaur m odel axial stress b y lo a d .............................................................83
29. B rachylophosaur m odel deflection b y lo ad...................................................... 87
30. B rachylophosaur m odel bending stress by lo ad ..............................................87
31. B rachylophosaur m odel axial stress b y lo ad .................................................... 88
ix
LIST OF TABLES
Table
Page
1. M uscle hom ologies betw een crocodilians and b ird s........................................ 15
2. Specim ens u sed in chapter 2 ..................................................................................46
3. U ltim ate stress values u sed in chapter 4.............................................................7 8
4. C alculated ultim ate tensile strain values.............................................................79
5. Sum m ary of tenontosaur m odel re s u lts ..............................................................81
6. U ltim ate lo ad s............................................................................................................ 84
7. Sum m ary of brachylophosaur m odel results......................................................86
8. Percentage changes in the m o d e ls............................................ ............................92
9. M orphologic data from MOR 794..........................................................................98
10. M orphologic data from MOR 682......................................................................100
X
ABSTRACT
Intratendinous ossification in dinosaurs (including birds) is a w ide sp read
phenom enon th a t has im plications for physiology, posture, biom echanics,
behavior, and systematics. Four separate studies w ere u n d ertak en to elucidate
ossified ten d o n biology in extinct dinosaurs and birds. The first investigation ■
reconstructs dorsal epaxial m usculature in extinct dinosaurs. C rocodilian and
avian dorsal epaxial m uscles are hom ologized. U sing the extant phylogenetic
bracketing approach, the three-layered trellis of H adrosauriform es is
hom ologized to the M. transversospinalis slips in crocodilians and the M. Iongus
colli dorsalis thoracica in birds. These tendons are ossified in H adrosauriform es
and som e birds. The parallel ossified tendons in other ornithischians are
hom ologized to the M. Iongissim us dorsi in Alligator. The second investigation
determ ines h ow ossified tendons in non-avian dinosaurs developed. A trophied
m uscle w as previously th o u g h t to be the origin of ossified tendons in non-avian
dinosaurs. U sing different age classes of h ad ro sau rs (Brachylophosaurus and
Maiasaura) and turkeys [Meleagris), the developm ental process of ossified tendons
in h ad ro sau rs is show n to be hom ologous w ith intratendinous ossification in
birds. But, the degree of ossification is greater in hadrosaurs, w hose tendons
also possess prim ary osteons, an external fundam ental system an d lines of
arrested grow th. The th ird investigation determ ines the histological diversity
of ossified tendons in D inosauria. D espite various anatom ical locations and
large differences in size, ossified tendons have uniform m icrostructure even
in specim ens th a t do n o t norm ally experience intratendinous ossification
(such as Ceratosaurus and Camarosaurus). The greater degree of ossification
noted above occurs in all non-avian dinosaurs. Also, variation in periosteal
bone developm ent occurs along the length of individual tendons. Ossified
tendons from m arginocephalians are u nique in th at they have larger portions of
fibrolam ellar bone and radial vascularity. In the fourth investigation, com puter
finite elem ent m odels are constructed for tw o ornithopods to assess ossified
ten d o n biom echanics. A n Alligator is u sed to em pirically determ ine joint
properties. Ossified tendons probably reduced tail deflection an d played a role
in locom otion in all ornithopods. The ossified tendon trellis of H adrosauriform es
m ay have increased the skeleton's ability to bear large b o d y m ass. N eu ral spine
heig h t and bone m aterial properties have the greatest im pact o n spinal rigidity.
I
CHAPTER I
DORSAL EPAXIAL MUSCULATURE IN DINOSAURS
Introduction
A rchosaur epaxial m usculature is generally considered conserved
com pared w ith the diversity of m uscles in the appendicular skeleton.
C onsequently interest in m uscular reconstruction in extinct archosaurs has
focused on the shoulder girdle, pelvis, or lim bs (Romer 1923a; R om er 1923b;
R om er 1927; G alton 1969; W alker 1977; G atesy 1994; Dilkes 2000; H utchinson
2001a; H utchinson 2001b; C arrano and H utchinson 2002). Some of these studies
deal w ith epaxial m uscles in a peripheral m anner by grouping them into the
"dorsalis trunci". But, the dorsal vertebral colum n has diverged significantly in
living archosaurs. For exam ple, birds possess a synsacrum and m an y develop
a notarium and "vertebral struts" (Storer 1982). The dorsal epaxial m usculature
associated w ith avian vertebral fusion is reduced in size and divisions
com pared w ith crocodilians. In m any birds the epaxial m uscles are pinched off
by the anterom edial un io n of the iliac blades. In contrast, crocodilians retain
plesiom orphic epaxial m uscles, except for the synapom orphic M. tendinoarticularis th a t contains anteriorly directed cones of m yosepta (Case 1981).
The only attem pts at detailed reconstruction of dorsal epaxial m usculature
in extinct dinosaurs are concerned w ith ossified epaxial tendons in M inm i
(Molriar and Frey 1987) and Iguanodon (Dollo 1886). M olnar an d Frey (idem)
u sed crocodilian w hile Dollo (idem) u sed avian anatom y com parisons for m uscle
reconstruction. But recent studies (H utchinson 2001a; H utchinson 2001b) have
2
show n the resolution pow er and robustness th a t an explicit phylogenetic context
using m any specim ens can provide for m uscle reconstruction an d the evolution
of soft tissues. M uscle reconstruction th a t uses a phylogenetic context relies
on osteological correlates (hard tissue indicators of soft tissue, such as bum ps
or scars) for inferring the condition of soft tissues (W itmer 1995). But ossified
tendons are m ore th a n ju st an osteological correlate consisting of bone surface
rugosity or projections caused by m uscle or ten d o n attachm ent. They are literally
p a rt of a soft tissue th a t becam e ossified and th en fossilized. Therefore, ossified
tendons are ideal for inferring soft tissues. U nfortunately, they only, regularly
occur in ornithischian dinosaurs, w here they are used as a diagnostic character
(synapom orphy).
D orsal epaxial m uscles are associated w ith breathing in b ird s (Baumel
et al. 1990) and terrestrial locom otion in crocodilians (Frey 1985). The epaxial
ten d o n trellis tho u g h t to be unique to h adrosaurs, lam beosaurs, an d iguanodons
(H adrosauriform es) are always interpreted as rigidifying structures that
affect b ipedal posture (O strom 1964) and locom otion (Dollo 1886). Also, the
existence of the trellis is u sed to argue for claims of sem i-aquatic (Brown 1916;
C olbert 1951) or terrestrial (O strom 1964) behavior. These argum ents assum e
th a t the ossified trellis is an anatom ical structure unique to H adrosauriform es.
In addition^ because ossified tendons are also u sed in system atics, a better
u n d erstan d in g of their m uscular reconstruction is im portant for assessing
characters com m only used to support m ultiple dinosaurian clades (e .g ..
O rnithischia and H adrosauriform es). This study's purpose is to reconstruct the
dorsal epaxial m usculature in extinct dinosaurs using a phylogenetic approach.
The m orphology and function of dorsal epaxial m uscles are discussed in this
context. This reconstruction w ill provide a basis for interpreting the biology of
3
e p axial m uscles and ossified tendons in all dinosaurs. Furtherm ore, because the
epaxial m usculature in birds is poorly resolved (Baumel et al. 1993), this stu d y
w ill h elp elucidate the structure and function of the dorsal region in birds.
M aterials and M ethods
Osteological d ata w ere collected from extinct dinosaurs at the M useum
of the Rockies (MOR), Royal Tyrrell M useum of Paleontology (RTMP), Field
M useum of N atu ral H istory (FMNH), an d U niversity of Chicago D epartm ent
of O rganism al Biology an d Anatomy. M ost of the structural d ata from ossified
tendons w as b ased on an articulated Brachylophosaurus canadensis (MOR-794) th at
possesses a three-layered trellis of ossified tendons in situ.
O steological data from birds w ere collected at the M OR an d M ontana
State U niversity (MSU), b u t m ostly at the U niversity of M ichigan M useum of
Zoology (UMMZ). Specimens w ere chosen based on availability an d spanned
seven orders: A nseriform es (Chuana torquata), C haradriiform es (Limnodromus
scolopaceus, and Ptychoramphus aleuticus), FalcOniformes (Accipiter cooperi),
G ruiform es (Porphyrio mantelli), Podicipediform es (5 specim ens of Podilymbus
gigas, 8 specim ens of Podilymbus podiceps, and Aechmorphus occidentalis-, and
4 specim ens of Podilymbus major), Psittaciform es (Amazona amazonica), and
Struthioniform es (Apteryx australis).
M yological d ata w ere collected from donated specim ens dissected on
the cam pus of M ontana State University, Bozeman. Avian specim ens spanned
five orders: Ciconiiformes (Ardea herodias), Colum biform es (Columba livia),
Falconiform es (Buteo jamaicensis), Galliform es (7 specimens of Meleagris gallopavo
an d Gallus domesticus), and Podicipediform es (Podilymbus podiceps). Three
4
crocodilians (Alligator mississippiensis) and tw o lizards (Tupinambis and Anolis
carolinensis) w ere also dissected.
The nom enclature u sed for crocodilian anatom y follows Gasc (1981) and
Frey et al. (1989), th o u g h som e hom ologies and synonym s from Gasc (idem) are
preferred. Avian anatom y nom enclature follows Baumel et al. (1993).
H om ology
The term "H om ology", as u sed in this study, refers specifically to
supraspecihc hom ology (Roth 1994), w hich is a correspondence am ong
characters in different taxa th a t share a recent com m on ancestor. T hat is,
hom ologies are synapom orphies (Patterson 1982; de Pinna 1991; R oth 1994;
H aw kins et al. 1997), w hich define m onophyletic groups (clades). Hom ologies
identified in this stu d y are also taxic (characters). Transform ational (change in
character state) hom ology is considered by the inclusion of extinct forms.
U sing these definitions, there are three w ays to test hypotheses of
hom ology: similarity, conjunction, and congruence. The first test is similarity,
w hich is the sim plest b u t w eakest tool (de Pinna 1991). It consists of identifying
hom ologies based on appearance. As Patterson (1982) notes, sim ilarity is no t
a test, b u t provides supporting data for proposing a hypothesis of homology.
D espite these caveats, it is pow erful because any kind of sim ilarity m ay be used,
such as com position, ontogeny, position, an d so on.
The second test for hom ology hypotheses is conjunction, w hereby the
proposed hom ologous structures are searched for in the sam e organism . If such
structures are discovered, hypothesis falsification follows (Patterson 1982).
For exam ple, w ings in a b ird cannot be hom ologous w ith the h in d lim bs of
m am m als, because birds possess h in d lim bs as well. D espite th e conclusiveness
5
of this test, it is of lim ited utility for transform ation (character state) hypotheses
of homology.
The th ird test, w hich is considered the strongest, is congruence. A
hom ologous character should only arise once for a group (it sh o u ld be a
synapom orphy). This uses com parisons of other hom ologies in a phylogenetic
fram ew ork to strengthen the proposed hypothesis of homology. The greater
the n u m b er of structures found to be hom ologous betw een tw o taxa, the m ore
-
su p p o rt the proposed hypothesis has. Like m any other historical sciences,
this test uses a "convergence of indep en d en t evidence" approach w here the
hypothesis becom es m ore robust according to the degree of su p p o rt (num ber of
characters found to be congruent).
Phylogenetic Fram ew ork for Soft Tissue Inferences
R econstruction of unpreserved soft tissues in extinct taxa is problem atic
because of variables such as m uscle attachm ent type, ontogeny, shifts in function,
and so forth (Bryant and Seym our 1990; Bryant and Russell 1992). The best
w ay to address these problem s is by using an explicit phylogenetic fram ew ork
(W itmer 1995) to infer unpreserved soft tissues. This m ethod u ses living sister
(bracket) taxa to provide phylogenetic su p p o rt for hypothesizing unpreserved
features in extinct organism s. Inferences are ranked according to the degree of
phylogenetic su p p o rt they provide. A level I inference is m ade w h e n b o th extant
bracketing taxa possess a soft tissue and associated osteological correlates. This is
unequivocal su p p o rt for inferring an u n p reserv ed soft tissue. A level I' inference
is m ade w h en b oth bracketing taxa possess a soft tissue th at lacks osteological
correlates. Level T inferences are less robust th a n a level I inference, b u t m ore
supportive th a n a level II inference. A level II inference is m ad e w h en one of
6
Figure I . Phylogenetic fram ew ork used in this study. Stem-based g roups are labeled
at the branches and node-based groups are labeled at nodes. I, A rchosauria; 2,
D inosauria; 3, Ornithischia; 4, Saurischia; 5, N eom ithes; 6, Galliform es. Bold tax a
are extant. M odified from G authier (1986), Cracraft (1988), and Sereno (1999).
the bracketing taxa possesses a soft tissue and associated osteological correlates
of the soft tissue in question. This provides lim ited su p p o rt for inferring an
unpreserved soft tissue. A level IT inference is m ade w hen one of the bracketing
taxa possess a soft tissue that lacks osteological correlates. Level IT inferences are
less robust than a level II inference, b u t m ore supportive than a level III inference.
A level III inference is m ade w hen neither of the bracketing taxa possesses a soft
tissue. This provides no phylogenetic su p p o rt for inferring an unpreserved soft
tissue. O nly level I , T, II, and IT inferences are used in this study. Lim itations on
level T and IT are discussed on an individual basis.
7
The phylogenetic fram ew ork u sed in this stu d y (Fig. I) is constructed
from several sources. G authier (1986) is u sed for the base of the tree; Sereno
(1999) is u sed to expand D inosauria, and Cracraft (1988) is u sed to expand
N eornithes.
Results
Crocodilia
Epaxial m usculature in crocodilians is straightforw ard an d does n o t differ
dram atically from the condition encountered in squam ates. H ow ever, epaxial
m uscles are less diverse in crocodilians th an those in squam ates because of
osteoderm s and the p ropulsive function of the tail (Case 1981). The follow ing
anatom y is based on the dissections of Alligator and the literature (Vallois 1922;
C ase 1981; Frey 1982; M olnar and Frey 1987; Frey et al. 1989).
Intrinsic Vertebral M uscles Mm. interneuralis (also called M. interspinales)
an d M m. interarcuales connect neural spines together at their anterior and
posterior edges. Their fibers ru n anteroposterior and are indistinguishable
from one another. Therefore, they w ill together be referred to together as the
M m. interneuralis. M edial to this m uscle lies the interspinal ligam ent. Mm.
interarticularis superiores connects successive p re and postzygapophyses. The
anterior border of the Mm. interarticularis superiores inserts on th e posterior
aspect of the zygapophyseal joint by a short fan shaped tendon. Because the
epaxial m uscles are fibrous, the preceding m uscle divisions w ere n o t easily
distinguishable. Lateral to this m uscle is the Mm. intertransversarii, w hich
connects successive transverse processes.
8
Figure 2. D iagram of dorsal epaxial m uscles in Alligator. A and B, dorsal and lateral
view s of m ajor muscles. C and D, slips of the M. transversospinalis excluding the
tendino-articularis. A nterior is to the right.
9
M. Transversospinalis Lateral to the intrinsic vertebral m uscles are a series
of slips th a t form the b ulk of the M. transversOspinalis (Fig. 2). The M. m ultifidus,
M. neurospinalis of Frey (1982) and Vallios (1922), is the m ost m edial slip and lies
next to the lateral surface of the neural spines. The tendon from the M. m ultifidus
originates as a sm all m ediolaterally fattened sheet in front of the n eu ral spine.
This passes anterodorsal over tw o vertebrae and inserts on the posterior sum m it
of the th ird n eural spine just u n d e r the attachm ent of the m edial tendinous head
of the M. sem ispinalis (M. articulo-spinalis).
Im m ediately lateral to this tendon, separated by only a th in layer of
m uscle fibers, is the M. spinalis, M. spino-articularis of Frey (1982) an d Vallios
(1922). This m uscle slip originates n ear the p rezygapophysis as a m ediolaterally
flattened sheet and runs posterodorsally to insert on the anterior sum m it of the
neu ral spine five to seven vertebrae posterior (Fig. 2C and D). In the lum bar
region, this ten d o n bifurcates twice, anteriorly receiving another ten d o n from
the m uscle belly and posteriorly a tendon from the anterior aspect of the neural
spine.
The M. sem ispinalis lies lateral to the M. spinalis and is sep arated from
it b y m uscle tissue (Fig. 2C and D). It is divisible into different sections, the
M. articulo-spinalis and M. tendino-articularis. The M. articulo-spinalis is the
m edial division, w hich originates at the base of the n eural spine an d passes
anterodorsally past four vertebrae. H ere it flattens out and splits into three
tendinous heads. The m edial head inserts on th e posterior aspect of the sum m it
of the next neural spine. The lateral h ead connects to the fascia of th e M. tendinoarticularis.
The lateral division of the sem ispinalis is the M. tendino-articularis.
A nteriorly pointed cones of m yosepta form successive slips in this m uscle. It is
10
interconnected to the M. articulo-spinalis b y a lateral ten d o n branch.
M. Longissim us D orsi Lateral to the M. transversospinalis and separated
from it by m yosepta lies the larger M. Iongissim us dorsi (Fig. 2A and B).
Posteriorly pointed cones of m yosepta form the slips of this m uscle. The internal
m argin connects to the lateral aspect of the M. tendino-articularis w hile the
external m argin connects w ith the M. iliocostalis. Each conical slip extends over
three vertebrae and inserts on the distal p a rt of the transverse processes. Fibers
from the M. intertransversarii interw eave w ith the fibers of the M. Iongissim us
dorsi. The m uscle originates on the anterom edial aspect of the ilium an d inserts
on the dorsal aspect of the transverse processes.
N eornithes
The dorsal epaxial m usculature in b ird s is poorly u n d ersto o d w ith the M.
Iongissim us dorsi com pared w ith m am m alian spinal m uscles b y Baum elz et al.
(1993). Therefore, aside from several references th at describe th e dorsal epaxial
m uscles as a w hole (H arvey et al. 1968; Z usi an d Bentz 1984; Baum el et al. 1993),
epaxial m uscle divisions are based on the dissections of this study.
Intrinsic Vertebral M uscles The intrinsic vertebral m uscles of the dorsal
region are reduced because of dorsal vertebral fusion into the notarium . The
M m. interneuralis (also called Mm. interspinales) are absent. Even at the notarialsynsacral articulation an d free dorsal vertebrae, the interspinus ligam ent is m ore
developed th an the Mm. interneuralis. The M m. intercristales in b ird s m ight
be hom ologous to M m. interarcuales based on its lateral position to the Mm.
interneuralis.
11
A
M. Iongus colli
thoracicus
B
Figure 3. D iagram of major dorsal epxial m uscles in Meleagris. A , dorsal view. B7
lateral view. A nterior is to the right.
Mm. interarticularis superiores is also disru p ted by the notarium , w hich
generally lacks laterally projecting zygapophyses associated w ith this muscle.
H ow ever, on the posterior free vertebra of the notarium and free dorsal vertebra,
a fan shaped tendon attaches to the posterior aspect of the zygapophyseal joint.
This tendon passes posteriorly to blend in w ith the fibers of the M. Iongissim us
dorsi. This is the only candidate for the hom ologue to the crocodilian Mm.
interarticularis superiores based on tendon insertion.
Mm. intertransversarii are present in the cervical region, b u t are absent
over the notarium due to the osseous developm ent betw een transverse processes,
though the M. Iongissim us dorsi sends fibers ventrad th ro u g h the transverse
foram ina. The anterior free dorsal vertebrae contains Mm. intertransversarii
12
M. spinalis
-M. multifidus
Figure 4. D iagram of the M. Iongus colli dorsalis thoracica slips in birds. A and
B, dorsal and lateral view of the slips in Meleagris. C and D, dorsal and lateral
view of the slips in Podilymbus. A and B dem onstrate one condition in birds w ith
an external layer hom ologous w ith the M. sem ispinalis of crocodilians. C and D
dem onstrate another condition in birds w ith an interal layer hom ologous w ith the
M. m ultifidus of crocodilians. A nterior is to the right.
13
connecting it to the transverse processes of the neck and the notarium . A thick
ten d o n of the lateral epaxial m usculature laterally b ounds this muscle.
M. Longus Colli D orsalis Thoracica This m uscle continues u n in terru p ted
into the M. Iongus colli dorsalis of the neck and lies im m ediately lateral to the
n eural spines. A nd, it has obliquely ru n n in g tendons located on its m edial
surface. These insert on the posterior or anterior sum m its of the n eu ral spines, a
com plex m orphology also seen in M. transversospinals of crocodilians. Given its
location adjacent to the neu ral spines and obliquely angled tendons, this m uscle
is hypothesized to be the hom ologue of the M. transversospinals of crocodilians.
Confusingly, this m uscle is also called the M. Iongissim us dorsi in b ird s (H arvey
et al. 1968), b u t M. Iongus colli dorsalis thoracica is preferred to avoid confusion
w ith the M. Iongissim us in crocodilians, w hich it is no t hom ologous with.
Like the hom ologous M. transversospinalis of other reptiles, this m uscle
com plex is com posed of various slips defined b y their long ten d o n s (Fig. 4).
U nlike those in crocodilians, the fibers of this m uscle are b len d ed together and
n o t easily divisible. Consequently, this complex m uscle is better u n d ersto o d
in a phylogenetic context. Birds possess tw o m orphologies of the slips. In
each m orphology, the M. spinalis is present. This m uscle slip is identified and
hom ologized based on its origin at the base of the neural spine an d tendons
th a t ru n posterodorsally attaching to the anterior aspect of the n eu ral spines or
corresponding locations on the crest of the n o tariu m or synsacrum . The num ber
of tendons is variable and appears to be d ep en d en t on synsacral an d notarial
developm ent. For exam ple, in Galliformes only three tendons are present. The
first ten d o n attaches onto the neural spine sum m it of the free vertebra at the
notarial-synsacral articulation. The other tw o tendons attach to th e roof of the
14
canal form ed by the m edial connection of the anterior iliac blades. In other birds,
such as Buteo, Podilymbus, and Ardea, these tendons ru n along the entire dorsal
region.
The tw o m orphologies m entioned above are distinguishable based on
w h eth er the second oblique ten d o n layer is m edial or lateral to the M. spinalis.
If a m edial layer is present, the lateral layer is absent and vice versa. The layer
m edial to the M. spinalis is hypothesized to be hom ologous to the M. m ultifidus
based its m edial anatom ical location in relation to the M. spinalis. A nd, its origin
at the base of the neural spine and tendons th a t ru n anterodorsally attaching
on the posterior aspect of the n eural spines also suggest th a t it is hom ologous
to the M. m ultifidus. It is n o t interrup ted by the notarium or synsacrum .
The M. m ultifidus is present in C haradriiform es, Podicipediform es, and
Struthioniform es.
The other m orphology is characterized b y a slip lateral to the M. spinalis,
w hich it is separated from by a thin sheet of m uscle tissue. This is hypothesized
to be hom ologous to the crocodilian the M. sem ispinalis based o n its lateral
anatom ical location in relation to the M. spinalis. U nlike the condition seen
in crocodilians, it does n o t appear to be divisible into separate m uscle slips
anteriorly. This m uscle originates at the base of the neural spine an d passes
anterodorsally to insert on the posterior sum m its of the n eu ral spines or
corresponding locations on the notarial or synsacral crests. A lth o u g h it often
passes along the dorsal region u n interru p ted , in Galliformes th e M. sem ispinalis
inserts on the neural spine sum m it of the first free dorsal vertebra b y one long
ten d o n an d on the posterior region of the n o tariu m by tw o long tendons. Thus,
like the M. spinalis, tendonous insertions over the notarium m ay be absent. The
M. sem ispinalis occurs in Ciconiiformes, Galliformes, and Falconiformes.
15
Table I. M uscle hom ologies betw een crocodilians and birds. A sterisks indicate
m u tually exclsuive presence of muscles.
Squamata
M. intemeuralis
M. interarticularis superiores
M. intertransversarius
M. Transversospinalis
M. multifidus
M. spinalis
M. semispinalis
M. Iongissimus dorsi
Crocodilia
Intrinsic Vertebral Muscles
M. intemeuralis
M. interarticularis
superiores
M. intertransversarius
Dorsal Epaxial Muscles
M. Transversospinalis
M. multifidus
M. spinalis
M. articulo-spinalis
M. tendino-articularis
M. Iongissimus dorsi
Neomithes
M. intemeuralis
M. interarticularis superiores
M. intertransversarius
M. Iongus colli dorsalis
thoracica
M. multifidus*
M. spinalis
M. semispinalis*
M. ascendens thoracicus
In birds, the M. m ultifidus and M. sem ispinalis do n o t occur together. In
either m orphology, the posterior tendons ru n w ith muscle tissue th ro u g h the
canal form ed the m edial fusion of the preacetabular iliac blades. The M. Iongus
colli dorsalis thoracica ends posteriorly at the posterior m argin of the canal or
outside it on the dorsal surface of the synsacrum m edial to th e anterior p a rt of
the M. levator caudae.
M. A scendens Thoracicus This is lateral to the M. Iongus colli dorsalis
thoracica. (Fig. 3). It originates on the anterior m argin of the iliac blades and the
dorsal surface of the transverse processes. A fter inserting on the dorsal vertebrae
(notarium ) it passes forw ard to insert on cervical vertebrae. It m ay have tw o
slips separated by an aponeurosis, as is the situation for Meleagris. In either case,
longitudinally oriented tendons are often p resen t deep w ith in this m uscle and
insert on the transverse processes of first free dorsal vertebra. The position of this
16
m uscle lateral to the M. Iongus colli dorsalis thoracica (hom ologized w ith the
crocodilian M. transversospinalis), its attachm ent onto the ilium and transverse
processes suggest th a t this m uscle is hom ologous w ith the M. Iongissim us
dorsi in crocodilians. N ishi (1938) suggests th a t this m uscle is hom ologous w ith
the crocodilian M. tendino-articularis. H ow ever, given its u n iq u e m yosepta in
crocodilians and the presence of the M. sem ispinalis in some birds, it seem s m ore
parsim onious to hypothesize hom ology betw een the M. ascendens thoracicus
w ith the crocodilian M. Iongissim us dorsi. H ow ever, the evidence supporting
either view point is rath er weak.
M uscle Reconstruction in Extinct D inosaurs .
A side from the proposed m uscle hom ologies (Table I), th e muscleten d o n system s described for crocodilians an d birds share another feature, an
architectural one, th at m ight not be obvious from the preceding descriptions.
The tendons of the M. m u ltih d u s and M. spinalis in crocodilians form a two
layered rhom boidal trellis th at lies vertically along the lateral aspects of the
neu ral spines. M ore laterally, a third layer (M. semi-spinalis) is obvious u p o n
dissection in Alligator. This three layered rhom boidal trellis is form ed entirely by
the M. transversospinalis tendon slips. The m edial m ost layer ru n s anterodorsal,
the interm ediate layer posterodorsal, and the lateral layer anterodorsal. In birds,
a tw o layered trellis is form ed in the M. Iongus colli dorsalis thoracica by either
the M. m u ltih d u s and M. spinalis or b y the M. spinalis and M. sem ispinalis.
G iven the current phylogeny of Aves (Cracraft 1988; Sibley 1994) there does no t
appear to be a phylogenetic pattern for these m orphologies. A nother feature of
the dorsal epaxial m usculature th at appears to have evolved in d ep en d en tly is
17
Figure 5. Ossified epaxial tendons in dinosaurs. The tw o layered trellis of the
M. Iongus colli dorsalis thoracica in A, Ptychoramphus aleuticus (Cassin's auklet,
Charadriiform es); B, Podilymbus podiceps (pied-billed grebe, Podicipediform es);
and C, Chuana toryuata (screamer, Anseriformes). A three layered trellis in a
h ad ro sau r (Brachylophosaurus, MOR-794). A nterior is to the rig h t in all panels.
an ossifying trellis. It is ossified in C haradriiform es, Podicipediform es and in
Meleagris. In all birds that possess them, the ossified tendons of the trellis are
fused to the neural spines. These are the "vertebral struts" of Storer (1982) and
their length varies from short vertebral projections (e.g. Chuana) to long rods
from the entire tendon (e.g. Ptychoramphus).
U nfortunately, the vertebral strut m orphology only occurs in som e birds.
Since epaxial m uscles do not appear to leave traces on the surfaces of vertebral
bone, the struts (ossified tendons) provide the only osteological correlate from
18
vertebrae. H ow ever, ossified tendons are n o t only osteological correlates,
b u t are fossilized parts of the m uscle-tendon systems. Ossified tendons in
H adrosauriform es form a regular rhom boidal trellis structure along the neural
spines of the dorsal, sacral, and caudal vertebrae. They begin in th e latter
cervicals, as is the case o f Iguanodon (N orm an 1980), or the anterior dorsals, in
Brachylophosaurus (Prieto-M arquez 2001) and extend into the distal th ird of the
tail. The lattice consists of three layers w ith a w eakly developed m edial layer,
th o u g h all three m ight n o t be preserved (Dollo 1886; N orm an 1986; PrietoM arquez 2001).
In crocodilians an d birds, the M. spinalis is easily identifiable because
it is the only slip w ith a ventral anterior end th a t runs posterodorsally. In
H adrosauriform es, the lateral m ost layer also runs this direction. Therefore, it
is a level I inference to reconstruct the lateral layer as the M. spinalis. The M.
m u ltih d u s an d M. sem ispinalis ru n anterodorsally b u t only the M. m ultihdus
is m edial to the M. spinalis. Because the M. m u ltih d u s is p resent in crocodilians
an d in som e birds, it is a level I inference to reconstruct the interm ediate layer
as the m ultihdus. The m edial m ost layer of the H adrosauriform es trellis runs
parallel to the outer layer (M. spinalis) b u t has no apparent hom ologue in living
archosaurs. U ndoubtedly, this layer is the rem ains of another set of slips in the
M. transversospinalis, and is probably u n ique (apom orphic). A lth o u g h tendons
of the M. sem ispinalis are n o t ossihed, it is a level T inference to reconstruct this
m uscle slip in all non-avian dinosaurs.
There are rare fossil specim ens th a t possess abnorm al ossihed
epaxial tendons. For exam ple, M olnar and Frey (1987) reconstruct the M.
transversospinalis in M ihm i from abnorm al ossihed tendons th a t form a lattice
typical of the M. transversospinalis. Also, the ossihed tendons in Dryosaurus
19
(G allon 1981) w ere possibly p a rt of the M. transversospinalis b ased o n their
distribution along the neural spines.
Given these level I and I' inferences, the trellis is probably the fossilized
rem ains of the M. transversospinalis, no t the M. Iongissim us dorsi. This
reconstruction places the trellis close to the vertebral colum n, separated by only
th in layers of muscle. Therefore, the H adrosauriform es trellis w as likely oriented
in the p arasagittal plane like those in birds and crocodilians. They u n d o u b ted ly
attached to neural spines by unm ineralized ten d o n entheses an d to m uscle by
m yotendinous junctions. This m orphology of a bony core and flexible term ini is
com m on in ossified tendons in birds (Vanden Berge and Storer 1995; Landis and
Silver 2002). M ost ossified tendons have a tap ered and flared term inus. Based
com parisons w ith birds, the tapered end likely continued on as unm ineralized
tendon. The flared end either m ineralized into the m yotendinous junction or onto
the fascia, giving it a !im bricated appearance. Ossified tendons are n o t usually
attached to the vertebral colum n in non-avian dinosaurs. But som e exceptions do
exist. For exam ple, it has been noted th at som e sacral tendons in Igmnodon are
attached to the n eural spines of th at region (N orm an 1986).
R econstruction of the M. transversospinalis is h ard er to su p p o rt in other
groups of non-avian dinosaurs th at lack a trellis of ossified tendons. Still, it is a
level T inference to reconstruct the rest of th e non-avian dinosaurs w ith the M.
transversospinalis (including M. m ultihdus, M. spinalis, and M. semispinalis).
R econstruction of the outer M. tendino-articularis is a level IT inference. This
m uscle slip, w hich has uniquely developed m yosepta in crocodilians and
apparently absence in birds, w ill not be considered further d u e to the lack of
su p p o rt inherent in a level IT inference.
O ther dinosaurs, such as some H adrosauriform es (Dollo 1886; Prieto-
20
M arquez 2001), Heterdontosaurus (Santa Luca 1980), Tenontosaurus (Forster
1990; W inkler et al. 1997), and Pachycephalosaurus (Sues and G alton 1987)
display parallel bundles of ossified tendons. N orm an (1980) suggests th at the
ligam entum apicum dorsalis and the ligam entum transversarii in Iguanodon
w ere displayed tendons from the lattice. But, the presence of these tendons
in other o rnithopods and Pachycephalosaurus suggest that they w ere no t
displaced. The parallel bundles of ossified tendons could be correlates to the
M m. interarticularis superiores or the M. Iongissim us dorsi. They could also be
ossified tendons from both m uscles, th o u g h the Mm. interarticularis superiores
u sually gives rise to short fan-like tendons in living archosaurs. In b ird s the M.
ascendens thoracicus usually gives rise to longitudinally arrayed tendons, w hich
m ay ossify (as in grebes). Therefore, it is a level II inference to reconstruct the
parallel bundles of tendons at the base of the n eu ral spines as th e M. Iongissim us
dorsi.
In Aves, the dorsal epaxial m usculature is reduced and sim plified
com pared w ith crocodilians. It also d isru p ted along the n o tariu m an d the
synsacrum . The evolution of the dorsal epaxial m usculature (such as the
disappearance of the M. m ultihdus) is hypothesized to coincide w ith changes in
the n o tariu m and the synsacrum . Evolution of the dorsal epaxial m usculature
probably follow ed changes in progressive inclusion of sacral vertebrae in
N eotheropoda, M aniraptora, C onfuciusom ithidae, O rnithothoraces, O rnithurae,
and N eornithes (H utchinson 2001b). A nterior expansion and m edial m igration of
the p reacetabular ilium correlates w ith vertebral fusion (H utchinson 2001b) and
notarial developm ent, w hich began in Archaeopteryx (Baumel et al. 1993).
21
D iscussion
Dollo (1886) supposed th a t the H adrosauriform es ossified ten d o n trellis
consisted of the M. spinalis dorsi, M. m ultifidus, and M. obliquu-spinales.
H ow ever, he hom ologized the m uscle tissue of living birds w ith the ossified
tendons in Iguanodon. Thus, he proposed th a t the H adrosauriform e trellis form ed
by atro p h y of m uscle tissue into ligam ent, w hich then ossified. Like Dollo,
N o rm an (1980; 1986) suggests th at e p axial m uscles are red u ced to ossified rods
(tendons) for reasons of metabolic efficiency. O sfrom (1964) also agreed w ith
Dollo's assessm ent, w hile A lexander (1985) suggests th at the trellis represented
the M. m u ltih d u s and M. spinalis based on Case's w ork on crocodilians (1981).
D espite this recognition, epaxial m usculature has consistently b een reconstructed
sim ply as the M. Iongissim us dorsi (Romer 1923b; Romer 1927; Lull and W right
1942; N orm an 1986; H utchinson 2001b). Parks (1924) reconstructs the epaxial
m usculature based on slips defined b y the ossified tendons, b u t th en m akes no
reference to the m usculature of living archosaurs.
This lack of epaxial m uscle and ten d o n resolution has m eaningful
consequences for hypotheses about the paleobiology of extinct dinosaurs.
Indeed, the trellis in H adrosauriform es has long been considered a unique
biological structure an d is often used as a synapom orphy (Sereno 1986; Sereno
1999). In these analyses, Sereno coded the "tw o-layered lattice" as a character
state of "ossified tendons". Given the presence of the trellis in living archosaurs
an d an ossifying trellis in C haradriiform es, Podicipediform es, an d some
G alliform es (Meleagris), it is plesiom orphic for archosaurs and convergently
ossifies at least four tim es w ithin D inosauria.
O ssified tendons in ornithischian dinosaurs have historically been
22-
in terpreted in functional term s of rigidity (Dollo 1886; Brow n 1916; Broili 1922;
C olbert 1951; O strom 1964; N orm an 1980; N o rm an 1986; R othschild 1987;
Bultynck 1992; Coombs 1995). These argum ents are used to justify current
paleobiological interpretations of H adrosauriform es. For exam ple, various
authors (Brown 1916; Colbert 1951) use the trellis in argum ents for the semiaquatic habits of hadrosaurs. The trellis of ossified tendons is in terp reted as an
adap tatio n th a t p rovided additional su p p o rt and pow er to lateral m ovem ents
of the tail w hile sw im m ing. Terrestrial interpretations of the trellis include
Dollo's (1886) suggestion th a t it stiffened the b o d y about the sacrum and tail. The
resulting rigidity supposedly countered dorsoventral loading im p arted b y the
M. caudofem oralis longus d u ring locomotion. A second terrestrial interpretation
suggests th a t the trellis of ossified tendons in H adrosauriform es resisted sagging
of the b o d y about the hips (Ostrom 1964). The ossified ten d o n trellis is also
suggested to have stiffened the tail m ediolaterally thus p ro v id in g an additional
argum ent against sem i-aquatic hadrosaurs and iguanodons (N orm an 1986). Since
the trellis occurs th ro u g h o u t D inosauria, these functional and behavioral claims
m u st be reevaluated.
Conclusions
The M. m ultifidus and M. sem ispinalis are found to occur exclusively
to one another in N eornithes as p art of the M. longus colli dorsalis thoracica
(hom ologous to M. transversospinalis in Crocodilia). The M. transversospinalis
is reconstructed in H adrosauriform es su p p o rted b y a level I inference and in
the rest of non-avian dinosaurs by a level T inference. R econstruction of the M.
Iongissim us dorsi is less clear (level II phylogenetic support), b u t the parallel
23
bun d les of tendons in som e ornithischians are hypothesized to b e th e M.
Iongissim us dorsi.
The slips of the M. transversospinalis form a tw o-layered rhom boidal
trellis in N eornithes and a three layered rhom boidal trellis in Crocodilia.
M oreover, these tendons are ossified in som e birds: C haradriiform es
(Limnodromus scolopaceus and Ptychoramphus aleuticus), Podicipediform es
(Podilymbus gigas, Podilymbus podiceps, Podilymbus major, and Aechmorphus
occidentalis), Chuana torquata, and Meleagris. Thus, various birds possess
an ossified ten d o n trellis. This structure w as th o u g h t to be u n iq u e to
H adrosauriform es. A nd, it has always been considered a key innovation for
h ad ro sau rs and iguanodons. The trellis is plesiom orphic for archosaurs and
ossification of the trellis occurs convergently w ith in dinosauria. Given the
w id esp read phylogenetic occurrence of the epaxial ten d o n trellis in dinosaurs,
functional and behavioral interpretations as well as system atic utility, m u st be
reevaluated.
24
CHAPTER 2
DEVELOPMENTAL PALEOBIOLOGY OF OSSIFIED TENDONS IN
HADROSAURS
Introduction
Birds are the only living group of dinosaurs (G authier 1986), m an y of
w hich possess tendons th at m ineralize and ossify d u rin g developm ent. These
avian groups include tinam ous (Tinamiformes), penguins (Sphenisciformes),
chickens and turkeys (Galliformes), cranes, herons, storks, ibises, an d flam ingos
(Ciconiiformes), gulls and shorebirds (Charadriiform es), rails (Rallidae), owls
(Strigiformes), hum m ingbirds (Trochilidae), w oodcreepers (D endrocolaptinae),
an d grebes (Podicipediform es) (Vanden Berge and Storer 1995). Ossification
occurs in various tendons th ro ughou t the body, b u t m ost often in the shank,
vertebral colum n, and w ing. The process of avian ossified ten d o n developm ent is
w ell un d ersto o d (Johnson 1960; Likins et al. 1960; N ylen et al. 1960; A bdalla 1979;
Landis et al. 1995; Landis and Silver 2002).
Ossified tendons associated w ith the axial skeleton also occur in all
O rnithischian dinosaurs, except for some thyreophorans. The m ost w ell know n
arrangem ent of these structures is found in had ro sau rs and iguanodons as
a rhom boidal trellis several layers thick (Fig. I). Dollo (1886) suggested that
ossified tendons w ere ossified ligam ents from atrophied epaxial muscle.
M bodie's (1928) histological w ork agreed w ith Dollo's, th o u g h h e term ed the
structures ossified tendons as they w ere th o u g h t to be derived from the tendons
of the M. sacrolum balis (Dollo 1886). In another histological analysis, Broili
25
(1922) suggested th at the high vascular content indicates developm ent directly
from m uscle tissue, skipping a ligam entous stage. M oodie (idem) and Broili (idem)
m ention histological w ork done on ossified tendons in birds (L ieberkuhn 1860).
But, they d id n o t use the developm ental process of intratendinous ossification
in b ird s to infer the developm ent in the non-avian dinosaurs. This lack of
com parison has led som e authors, like Reid (1996), to conclude: "But, how ever
their grow th w as initiated, the 'ossified ten d o n ' described from Iguanodon is
clearly n o t an ossified tendon in the literal sense of the term ." Reid's "literal
sense" m ay be in terpreted as a tendon from a m uscle-tendon u n it th at becam e
ossified leaving no trace of the original tissue.
The question then arises: w h at are these bony structures in om ithischians?
This question is im portant n o t only for physiological reasons^ b u t because
ossified tendons in ornithischian dinosaurs have historically b een interpreted in
functional term s of spinal rigidity (Dollo 1886; Brow n 1916; Broili 1922; Colbert
1951; O strom 1964; N o rm an 1980; N orm an 1986; Rothschild 1987; Bultynck 1992;
Coom bs 1995). These adaptationist interpretations have im plications for posture,
locom otion, and general biological un d erstan d in g of extinct om ithischians. Since
such interpretations date to Dollo (1886), w ho argued th at these b o n y rods w ere
developed from atrophied muscle, a developm ental u n d erstan d in g of ossified
tendons is essential to evaluate adaptational claims.
The aim of this stu d y is to discover the developm ental process of
intratendin ous ossification in ornithischian dinosaurs. This is accom plished by
characterizing the histology of fossil tendons across three age classes (nestling,
juvenile, and adult) of closely related hadrosaurs. These d ata are com pared
w ith an ontogenetic sequence of histological d ata from ossified tendons in
birds. This com parison form s the basis for inferring intratendinous ossification
26
developm ental patterns in hadrosaurs and other ornithischians.
M aterials and M ethods
Eighteen fossilized epaxial ossified tendons w ere sam pled for histological
sectioning from a nestling Maiasaura peeblesorum (H orner et al. 2000), tw o from a
juvenile Brachylophosaurus canadensis and fifteen from an ad u lt Brachylophosaurus
(Prieto-M arquez 2001). Relative body size, su tu re fusion, an d histology (e.g.
lines of arrested grow th or LAGs) of long bones w ere u sed to infer grow th
stages (H orner et al. 2000). The use of Maiasaura and Brachylophosaurus to infer
developm ental m echanism s in intratendinous ossification in h ad ro sau rs is
justified by their status as sister groups (Fig. 6) and postcranial conservation in
hadro sau rs (W eishampel and H orner 1990; Prieto-M arquez 2001).
Fossil specim ens u sed in this stu d y w ere collected from th e Late
Cretaceous Two M edicine (mid-late C am panian) and Judith River (late
C am panian) Form ations of M ontana b y M ontana State U niversity and Princeton
U niversity field crews. Taphonomically, all fossil tendons w ere articulated along
the axial skeleton except the juvenile tendons, w hich came ou t of a monospecific
bone be d for juvenile hadrosaurs. The fifteen tendons from the a d u lt hadrosaur,
a com plete and articulated Brachylophosaurus (MOR-794, H IP 1999-20), w ere
sam pled from the dorsal, sacral, and caudal regions along the axial skeleton.
W here possible, sections from MOR-794 tendons w ere taken from th e term ini
and shafts of each ten d o n across the three laterally stacked layers. Juvenile
Brachylophosaurus (M O R 1071, H IP 2002-14) ten d o n sam ples w ere tak en from
the shaft. N estling Maiasaura (YPM-PU 22400, H IP 1990-18) sam ples w ere taken
across the w hole vertebral column, capturing vertebrae as w ell as tendons.
27
Figure 6. Phylogenetic fram ew ork used in this study. Stem -based groups are
labeled at the branches and node-based groups are labeled at nodes. I, A rchosauria;
2, D inosauria; 3, Ornithischia; 4, N eornithischia; 5, H adrosaurinae; 6, Saurischia; 7,
Theropoda. M odified from Sereno (1999).
H istological procedure followed standard paleohistological techniques (Wilson
1994).
Turkey (Meleagris) tendons w ere donated by the Seder Ridge Turkey
Farm. Tendons from the M. gastrocnem ius, M. fibularis longus, and M. Iongus
colli dorsalis thoracica w ere sam pled from tw o groups of turkeys: 16 week-old
hens and 23 w eek-old toms. U sing both sexes is justified because intratendinous
ossification does not vary betw eens sexes (V anden Berge and Storer 1995). The
tendons w ere fixed in 10% neutral buffered form alin for storage. M ineralized
tendons from Meleagris w ere not decalcified before sectioning. Meleagris
tendons w ere sectioned in different areas along their length to increase the
ap p aren t ontogenetic sam pling range - avian tendons have a central ossification
28
center (Johnson 1960; Landis and Silver 2002). H em atoxylin an d eosin (H&E)
stains w ere u sed on the Meleagris sam ples. Terminology follows established
nom enclature for bone (de Ricqles 1980; Fancillon-Vieillot et al. 1990; Reid 1996;
C urrey 2002) and tendon m icroanatom y (Kannus 2000).
H istological Description
Turkey Tendon H istology
Figure 7A show s unm ineralized tendon. It is ordinary vertebrate
tendon, w hich is highly conserved across the group (Summers an d Koob 2002).
M ineralized tendons from the 16 w eek-old h e n show a variety of histologic
changes th at are well docum ented in the literature (Lieberkuhn 1860; Johnson
1960; A bdalla 1979). These changes include enlarging fibroblasts w ithin
developing lacunae (Fig. 7B) th at appear sim ilar to h y p ertro p h ied chondrocytes
or fibroblasts in ten d o n hbrocartilage (Rooney 1994; Benjamin an d Ralphs 1998;
Felisbino and C arvalho 1999). A ssociated w ith this change, crim p (waviness)
is lost an d collagen fibers becom e straight (Fig. TB). C rim p loss is associated
w ith a n ew m atrix consisting of apatite th a t pervades the tendon, expanding
as a m ineralization front. Following these changes, resorption cavities appear,
som e of w hich possess scalloped edges indicating osteoclastic activity (Fig. TC).
Subsequently, in the 23 w eek-old turkeys, H aversian canals are p resen t in the
center of the tendon (Fig. TC and D). These canals appear identical w ith those
found in bone tissue. Prim ary osteons are absent, because the tissue is ossifying
from the center. Vascularity is uniform ly longitudinal.
The zone of transform ation (m ineralization) progresses from the m iddle
of the ten d o n outw ard, centrifugally and longitudinally (Johnson 1960). This
29
Figure 7. Histological photographs of Meleagris tendons. A, unm ineralized tendon
from a 16 w eek old hen w ith longitudinal fibroblasts (PB) lying along w avy
collagen fibers (longitudinal section, 400x). B, hypertrophied fibroblasts (HFB)
associated w ith the onset of ossification (longitudinal section, 400x). Also note the
m ineralized collagen fibers (MF) that have lost their characteristic w aviness (UMF).
C, tendon from a 23 w eek old tom that contains young H aversian canals (HG) and
scalloped edged (SE) resorption cavities (cross section, 400x). D, a H aversian canal
(HG) in an ossified tendon a 23 week old tom (cross nicols, cross section, 400x).
produces an osseous core present in m ost Meleagris tendons. Even in the 23 weekold tom s, the original tendon paratenon su rro u n d s an osseous core.
H ad ro sau r Tendon H istology
N estling Maiasaura tendons are characterized by a high degree of
30
vascularity, as is vertebral bone (Fig. 8A). M ost of the cavities in the tendon
are scallop-edged resorption cavities (erosion rooms). The vascular structure
is uniform ly longitudinal in orientation. M ineralized fibrils b o u n d together in
bun d les constitute the base m aterial. These are longitudinal like the orientation
of the vascular canals. N ascent H aversian canals are present. Fibroblasts are
large, ro u n d and reside betw een fiber bundles (Fig. SB). Thick canaliculi project
from fibroblasts.
The larger of the tw o juvenile sam ples (Brachylophosaurus) is identical
w ith sam ples found in the ad u lt tendons and contains three rest lines. These
could be biom ineralization tide m arks or lines of arrested g ro w th (LAGs). The
rest lines are probably LAGs based on appearance, w hich im plies periosteal
accretion pa st the original bo u n d ary of the tendon. The. sm aller sam ple contains
a core of dense H aversian bone in the center of the ten d o n (Fig. SC). M any
have cem ent lines around them indicating the previous existence of resorption
room s. Osteocyte lacunae and canaliculi are visible, th o u g h the latter are no t
alw ays easily identified. The center of the ten d o n also contains longitudinal fiber
b u ndles visible betw een osteons (Fig. SD). In the periphery, the ten d o n possesses
a different appearance, w here collagen fibers are m ore abundant, as are prim ary
osteons.
Ossified tendons from the adu lt h ad ro sau r (Brachylophosaurus) have m ore
H aversian tissue and prim ary osteons th an the nestling or juvenile. Figure SE is
representative of the general structure observed in mid-section. The perim eter
consists of a th in layer of lightly vascularized tissue w ith longitudinally arrayed
fibers. V ascularity in this region is alm ost solely com prised of p rim ary osteons
oriented longitudinally. The occurrence of m ultiple LAGs in every section is
notable (Fig. IOA and B) com pared w ith their absence in the turkey, nestling,
31
Figure 8. H istological photographs of h ad ro sau r tendons. A, tendon from a
Maiasaura nestling (cross section, 40x). N ote the abundance of collagen fibers
(CF) and young H aversian canals (HG). B, detail im age of fibroblasts and thick
and stunted canaliculi (CA) (cross section, 400x). C, the juvenile Brachylophosaurus
tendon contains a core of H aversian canals su rro u n d ed prim ary osteons (cross
section, 40x). D, detail im age of collagen fibers (CF) betw een H aversian canals
(cross section, 400x). E, tendon from an ad u lt Brachylophosaurus (cross section
IOOx). N ote the dense H aversian bone. F, collagen fibers (CF) present betw een
H aversian canals in another adult Brachylophosaurus tendon (cross nicols, cross
section, 400x).
32
an d one of the juvenile tendons. Lines of arrested grow th are com pressed at the
surface form ing an external fundam ental system (EFS) in m ost sam ples. Prim ary
osteons are p resent in the periphery w ith a h ig h concentration of H aversian
canals at the center of the tendon. W hile H aversian replacem ent is extensive in
the shafts of the tendons, longitudinal collagen fibers are still p resen t betw een
som e H aversian canals (Fig. 8F).
Sections of the proxim al and distal ends of the ad u lt tendons show
dram atically less H aversian replacem ent and lack the LAGs seen in the shafts.
Also, the ratio of fibrous tissue w ith respect to osteons is greater in the term ini of
the tendons com pared w ith the shafts. H aversian canals are confined to the core
of the ten d o n term ini w hile the periphery contains m ostly longitudinally arrayed
fibers (Fig. 9B). R esorption cavities are concentrated centrally in the term ini.
Some sam ples also contain fields of thin d ark filam ents th a t ru n at oblique angles
to the transverse sections of the term ini. These are m ost likely Sharpey's fibers,
suggesting th a t the term ini interfaced w ith m uscle or unm ineralized tendon.
D iscussion
C om parative H istology
Turkey ten d o n m ineralization begins aro u n d 12 w eeks w ith noticeable
histological changes around 16 weeks of age, th o u g h the degree of m ineralization
varies am ong tendons w ithin a particular anim al (Johnson 1960; A bdalla 1979).
This transform ation can be thought of as a series of six stages. The first stage
involves proliferation and hypertrop h y of fibroblasts (tenocytes) (Johnson
1960; A bdalla 1979). Secondly, m ineralization begins in extracellular vesicles,
th en separately in the sm all (40 nm) hole zone channels am ong sequentially
33
arrayed collagen molecules. These apatite crystals act as nuclei for crystal
plaques th a t grow in the direction of the long axis of the ten d o n (Landis and
Silver 2002). C ollagen m ineralization begins aro u n d the exterior of th e bundles
an d proceeds into the fibrils (Johnson 1960). Increased vascularity also occurs
at this tim e w ith the developm ent of resorption cavities (A bdalla 1979). These
features are interesting because tendon is norm ally no t vascularized. Third, the
collagen bundles enlarge to diam eters of 15 pm, becom e dense, an d lose crim p
(waviness), w hich characterizes the so called transform ation zone (Johnson
1960). This zone grow s as apatite crystals form along the len g th of collagen fibrils
(Traub et al. 1989; Landis and Silver 2002). D uring the fourth stage, fibroblasts
develop canaliculi and becom e encased in lacunae th ro u g h a n ew m atrix (osteoid
containing m ostly collagen w ith lesser polysaccharide am ounts an d traces of
lipid) th a t is secreted into the interstitial space am ong collagen b u n d les (Johnson
1960; A bdalla 1979). The fifth stage is m arked b y the onset of ossification.
Osteoclasts further increase the size of the resorption cavities. Progressive
H aversian system developm ent is the final stage of intratendinous ossification
(Johnson 1960). Vascularity is oriented longitudinally an architecture parallel to
the collagen fibers of the tendon.
As n oted above, in avian species, m ineralization begins in the central
p a rt of the ten d o n and proceeds axially and centrifugally (Johnson 1960; Landis
an d Silver 2002). The turkey tendons used in this study confirm these general
findings, w ith the ratio of prim ary tendon tissue to osteonal bone hig h er in the
16 w eek-old turkeys th an in the 23 w eek-old turkeys (Fig. 2). The histology of the
h ad ro sau rian ossified tendons strongly suggests th at this sam e developm ental
p attern w as occurring. For example, the degree of osteonal developm ent is
low est in the Maiasaura nestling, w hich contains young H aversian canals
34
Figure 9. Juxtaposed histological im ages of Meleagris tendons on the left w ith
Brachylophosaurus tendons on the right. A, core tissue su rro u n d ed by peripheral
tissue (both longitudinal section, IOOx). In the case otMeleagris, the core is com posed
of m ineralized collagen fibers and the peripheral tissue unm ineralized collagen
fibers. The core of the fossilized Brachylophosaurus bone tissue is su rro u n d ed by
straightened collagen fibers in a m ineral matrix. B, cross sections arranged as
above (both IOOx). M ineralized core of a Meleagris tendon show s resorption cavities
intersperced am ong collagen fibers. The tip of a Brachylophosaurus tendon show s
a lesser degree of bone developm ent than in the m id-shaft and contains scallop
shaped collagen fibers sim ilar to those seen in the Meleagris m ineralized core.
35
an d n o p rim ary osteons. This condition is characteristic of all 18 samples.
O steonal developm ent is greater in the juvenile samples. The center is filled
w ith H aversian bone dem arcated by cem ent lines indicative of rem odeling.
The highest degree of H aversian developm ent occurs in the shafts of the adult
h a d ro sau r tendons. H ow ever, the term ini of the ad u lt h ad ro sau r tendons are
filled w ith collagen fibers and few H aversian canals and p rim ary osteons.
Progressive bone developm ent tow ards the center of the ten d o n is identical to
the condition in intratendinous ossification in birds. Also, a core of dense tissue is
su rro u n d e d by parallel arrayed fibers in the periphery. This suggests centrifugal
ossification and is com parable to the m orphology seen in lateral sections of from
Meleagris tendons (Fig. 9A).
The base tissue in the nestling H ad ro sau r tendons is com posed of bundles
of fibrils approxim ately 6 pm diameter. Fiber bundles are approxim ately 13 pm in
diam eter in the ad u lt H adrosaur. A t higher m agnification, these bundles consist
of fibrils approxim ately 2 pm in diam eter. In tu rk ey tendons, I pm fibrils are
packed together into bundles 15 pm in diam eter (Johnson 1960). These fibril sizes
are large com pared w ith those found in non-m ineralizing tendons, w hich range
in diam eter from 150-200 n m in the patellar ten d o n and 20-50 run in th e Achilles
ten d o n in hum ans (Rooney 1994). Thus, sim ilar size increases in fibril and b u n d le
diam eter due to intratendinous ossification in turkeys and h ad ro sau rs suggest a
sim ilar m ode of developm ent.
LAGs and the EFS
Lines of arrested grow th are absent in the h ad ro sau r nestling (Fig. 8A
an d B). The larger sam ple of the juvenile ten d o n contains three LAGs w hile the
sm aller juvenile ten d o n lacks them . A long w ith fibrolam ellar bone, m ultiple
36
LAGs characterize ad u lt h a d ro sau r tendons (Fig. 10). These features are absent
in the tu rk ey tendons. Lines of arrested g row th in ossified ten d o n s of the
juvenile and a d u lt hadrosaurs is un u su al in regards to their absence in the
tu rk ey specim en given their sim ilarity on nearly all other histological features
(save the degree of H aversian developm ent). The lines could be interpreted
as biom ineralization tide m arks, b u t these are n o t ap p aren t in cross section of
Meleagris tendons. Lines of arrested grow th are pro d u ced b y periodic cessation
of periosteal grow th, a plesiom orphic feature in tetrapods (H orner et al. 1999).
They have been observed in dinosaur bone in m ost taxa an d have been u sed in
discussions of developm ental paleobiology in dinosaurs and other archosaurs
(de Ricqles et al. 2000; H orner et al. 2000; de Ricqles et al. 2001; Erickson et al.
2001; H orner et al. 2001; Stark and C hinsam y 2002). The lines in H ad ro sau r .
ossified tendons w ere interpreted as LAGs b y H orner and de Ricqles et al
(1999), w here their num ber was found to correspond to LAG n u m b ers in other
skeletal elem ents across age classes. In this study, LAG num bers vary w ithin
an d am ong tendons, a phenom enon discovered for LAGs in long bones in
H adrosaurs (H orner et al. 2000). The n u m b er of LAGs in ossified tendons at
each grow th stage (0 for nestling, 0-3 in the juvenile, and 2-5 in th e adult) also
corresponds to LAG developm ent in h a d ro sau r long bones at th e sam e grow th
stages (H orner et al. 2000). M oreover, the a d u lt had ro sau r's te n d o n shafts ap p ear
to have an external fundam ental system (EFS) form ed by the com pression of
LAGs at the external surface. This indicates th a t new bone w as being laid dow n
in a m anner sim ilar to long bone appositional grow th. Indeed, a n u trien t vessel
w as discovered entering a tendon shaft in th e ad u lt hadrosaur. This feature is
com m on in bone b u t absent in tendons, w hich are fed th ro u g h th e enthesis,
m yotendinous junction (Rooney 1994) or a vascular p aratenon sh eath (K irkendall
37
Figure 10. A dult had ro sau r tendons sectioned through the shaft. N ote the arrow s,
w hich point to lines of arrested grow th (LAGs). These features are norm al seen in
bone b u t are absent in the turkey tendons used in this study.
and G arrett 1997). Thus, this feature dem onstrates the fully transform ed
condition present in the hadrosaur tendons.
W hat m akes this interpretation novel is that intratendinous ossification is
seen as a unique process of tissue transform ation, so called m etaplasia (Haines
and M ohuiddin 1968). By definition, m etaplasia is term inal tissue transform ation
into another type, nam ely bone. Thickened, stunted, or reversing canaliculi
and h y pertrophied fibroblasts are notable features for m etaplastic bone at
cartilage, ligam ent, and tendon entheses (Haines and M ohuiddin 1968; Reid
1996; Benjamin et al. 2002). A bnorm al canaliculi are also k now n in new bone
d u ring repair (K usuzaki et al. 2000). However, it has been noted (Reid 1997)
th at the lacunae of ossified tendons in extinct non-avian dinosaurs em it norm al
canaliculi. H e also describes highly vascular osteoblastic tissue th a t could have
only form ed from an endosteum and a periosteum . These facts are then used
to argue that dinosaurian ossified tendons w ere not of m etaplastic origin. But,
abnorm al canaliculi are only found in the m ineralized prim ary tissue not in
subsequent bone deposition. And, ossified tendons in birds develop H aversian
38
system s w ith norm al canaliculi (Lieberkuhn 1860; A bdalla 1979), so abnorm al
canaliculi should n o t be expected in the osteonal bone of ossified tendons in
ornithischians.
E ndochondral bone develops from a cartilage precursor. C hondrocytes
hypertrophy, the precursor tissue m ineralizes, and the bone grow s centrally
by m arro w tube form ation and peripherally b y periosteal apposition (Castanet
et al. 1993). Intratendinous ossification in b ird s occurs by the sam e general
pattern, except the precursor tissue is ten d o n and it occurs late in developm ent
com pared w ith endochondral bone ossification. But, the h y p ertro p h ied lacunae
of ossifying ten d o n fibroblasts are n o t chondrocytes (Retterer an d Lelievre 1911).
Some authors (Suzuki et al. 2002) still refer to the h y p ertro p h ied tenocytes as
chondrocytes at the hbrocartilage b o rd er of ten d o n entheses. H ow ever, given
the association of h ypertrophied fibroblasts w ith biom ineralization in bone
developm ent, intratendinous ossification, an d entheses, these cells are m ore
appropriately identified as tenocytes n o t chondrocytes.
Ossified ten d o n developm ent in h ad ro sau rs appear to have been tim ed
w ith the developm ent of their long bones, w hich indicates an earlier onset
of ossification com pared w ith birds. Also because hadrosaurs d id n o t reach
full size w ith in a year, unlike living birds w ho reach adulthood w ith in a year
(H orner et al. 1999; H orner et al. 2000), the developm ent of intratendinous
ossification w as prolonged com pared w ith birds. This late developm ental offset
in ten d o n ossification was likely responsible for producing ten d o n s com posed of
rem odeled bone w ith a periosteum . Subsequent periosteal bone g ro w th outside
the perim eter of the original tendon best explains the occurrence of LAGS and an
EPS.
C ollagen fibers are p resent in ad u lt hadrosaurs, b u t distinguishing am ong
39
fibers laid d o w n by a periosteum and those of the original ten d o n is im possible
based on appearance. Given the w ay tendons ossify in birds, th e collagen fibers
betw een H aversian system s at the tendon center m ay be from the original ten d o n
tissue w hile those outside of the interior m ost LAG w ere probably deposited b y a
periosteum or a paratenon taking on a periosteal function.
A daptation
D evelopm ental evidence p resented here suggests th a t caution should be
exercised w h en form ing adaptational or functional claims for ossified tendons
in m echanical term s (e.g. increasing stiffness). There is no histological evidence
suggesting th a t h a d ro sau r ossified tendons w ere derived from m uscle tissue.
In fact, the turkeys sam pled for this stu d y possessed ossified ten d o n s in the M.
Iongus colli dorsalis thoracica th at did n o t appear atrophied at all, suggesting
th a t D ollorS (1886) an d BroilirS (1922) developm ental hypotheses are incorrect.
DollorS original claim th a t ossified tendons stiffened the vertebral colum n w as
based on the loss of epaxial muscles. Since ossified tendons d id n o t develop from
atrophied m uscle, the basis for functional claims lacks support.
It has been claim ed th a t ossification in tendons occurs m echanically
th ro u g h su p p o rt of heavy loads and th at ossified tendons in ornithischians
resem ble long rods of bone th at develop in long tendons subjected to large
continuous loads (M oodie 1927). A nd, pathological ossification (diffuse
idiopathic skeletal hyperostosis or DISH) can occur at stress p oints in tendon
and. ligam ent (Rothschild an d M artin 1993). A lthough ossified ten d o n s are no t
pathologic, stress related DISH has been suggested as the m ode of ossification .
in ornithischians (Rothschild 1987). Stress induced ossification has also been
suggested for Meleagris (Landis and Silver 2002). However, it is unlikely th at
40
forces generated in the hatchling physiologically caused these tendons to ossify
given their sm all size and early onset of m etaplasia before hatching. Em bryonic
Hypacrosaurus ossified tendons su p p o rt this conclusion (Rothschild an d Tanke
1992).
Conclusions
O ssified tendons develop by transform ation of ten d o n into bone. Several
characteristics distinguish intratendinous ossification from skeletal bone
developm ent. Skeletal bone and ossified tendons originate from precursor
avascular connective tissue. The ability of fibroblasts to m ineralize and
vascularize this precursor tissue is shared b y skeletal bone an d ossified tendons.
The shape of tendons constrains collagen fiber orientation into longitudinal
orientation. Therefore, original tendon fibers and bone fibers laid d o w n b y a
periosteum cannot be distinguished in heavily ossified h ad ro sau rian tendons.
A nd, since they lack a periosteum , ossified tendons in b ird s cannot develop
prim ary osteons, w hich are entrapped.by periosteal grow th. All osteons in avian
ossified tendons are secondary.
H ad ro sau r ossified tendons appear nearly identical w ith tu rk ey ossified
tendons at m ore advanced grow th stages. Both change histologically along
the length of the ten d o n becom ing m ore ossified tow ard th e center. Also, the
increase in collagen fiber size through ossification is sim ilar in b o th b ird s and
hadrosaurs. These d ata suggest th at the developm ental process of intratendinous
ossification is identical in bo th taxa, differing only in the onset, an d offset of the
process. W hile noticeable histological changes in turkey ten d o n ap p ear around
16 w eeks of age (Johnson 1960; A bdalla 1979), the hatchling Maiasaura already
41
has y o ung H aversian canals developed. Since intratendinous ossification has
also been reported in em bryonic had ro sau rs (Rothschild and Tanke 1992),
biom ineralization m u st have begun embryonically. Thus, intratendinous
ossification w as initiated d u rin g a m uch earlier developm ental stage in
h ad ro sau rs th a n in birds. The degree of ossification in the ad u lt tu rk ey tendon
slightly m ore developed th a n the h ad ro sau r nestling, b u t m u ch less so th an the
juvenile h ad ro sau r tendon.
H ad ro sau rian ossified tendons regularly develop fibrolam ellar bone and
secondary osteons, as well as LAGs and a periosteum . Early o nset of m etaplasia
allow s h ad ro sau rian ossified tendons to m atch the developm ental patterns and
tim ing of their long bones (H orner et al. 2000). Late offset of m etaplasia produces
ossified tendons th a t developm entally correspond w ith the rest of th e skeleton in term s of LAG num bers and presence of H aversian bone.
In his w ork on the histology of d inosaur bones, Reid (1996) states "But,
how ever their grow th w as initiated, the 'ossified ten d o n ' described from
Iguanodon is clearly n o t an ossified ten d o n in the literal sense of the term ." The
histological data in this rep o rt argue against Reid's interpretation and suggest
th a t the long bony rods found in h ad ro sau r dinosaurs are in d eed tendons
nearly identical w ith those found in their close living relatives, birds. Therefore,
ornithischian ossified tendons should be interpreted as m etaplastic tendons,
n o t atrophied ligam entous m uscle tissue or as a biologic structure of Unknown
origin.
This stu d y revealed several areas concerning ossified te n d o n developm ent
in dinosaurs (including birds) th at rem ain unansw ered. For exam ple, does this
developm ental process occur in all ornithischians? If so, is early onset and late
offset of intratendinous ossification characteristic for this group? D id they evolve
42
by peram orphosis? Do birds of advanced age develop rest lines or LAGs? Will
birds developm entally ossify their tendons, even w hen norm ally m etaplastic
tendons are relieved of all stress? The answ ers to these developm ental questions
rem ain to be investigated.
43
CHAPTERS
HISTOLOGY OF OSSIFIED TENDONS IN DINOSAURS
Introduction
Tendon m icrostructure is conserved in vertebrates (Sum m ers and Koob
2002). But a strange phenom enon regularly occurs in m any din o sau rian clades;
tendons m etaplastically transform into bone. This occurs in various avian species
(Vanden Berge and Storer 1995) and in nearly all ornithischian dinosaurs, w here
they are synapom orphic (Sereno 1999). O ssified tendons are fo u n d in m any
b o d y p arts in birds such as the leg (Landis and Silver 2002), back (Rydzewski
1935), neck (Boas 1929), and w ing (Vanden Berge and Storer 1995). O rnithischian
tendons are m orphologically sim ilar to (though m uch larger than) ossified
tendons fo u n d th ro u g h o u t N eornithes. They also develop b y sim ilar processes
(see chapter 2). H ow ever, w here in N eornithes they occur th ro u g h o u t the body,
ossified tendons are found only along the vertebral colum n in ornithischian
dinosaurs. Small ornithopod dinosaurs have ten d o n bundles th a t ru n along the
dorsal aspect of the spinal colum n (e.g. Forster 1990). H ypaxial tendons are also
present in the tail of these animals. Iguanodons, hadrosaurs, an d lam beosaurs
have an ossified ten d o n rhom boidal trellis along the dorsal aspect of their spinal
colum n (Brown 1933; Lull an d W right 1942; N o rm an 1980; N o rm an 1986).
A nkylosaurs have nested V -shaped tendons in their tails th a t extend anteriorly
from the club (Coombs 1995). A nd enlarged bundles of tendons are often found
at the base of pachycephalosaur spines (Sues an d G alton 1987). There are also
fossilized ossified tendons from taxa n o t k n o w n to experience intratendinous
44
ossification, such as M inm i (M olnar and Frey 1987). These isolated cases are m ore
com m on th a n m ight be expected in non-avian theropods (MOR 693 and M N N
TIG6) and a sauropods (DNM28).
Thus, ossified tendons occur in a w id e phylogenetic range w ith in
D inosauria. Since they are also located in m an y different areas in the body,
they m ight be expected to develop by a variety of physiologic m echanism s
and have disparate functions. Historically, histological descriptions have only
been perform ed on isolated tendons and there are no data on w h eth er ossified
te n d o n histology varies according to anatom ical region w ith in an in dividual or
across taxa. It is also u n k n o w n w hether ossified tendon m icrostructure changes
in different parts of the body in birds, the only living taxon th a t regularly
experiences intratendinous ossification. Therefore, the prim ary aim of this stu d y
is to characterize ossified tendon histology across D inosauria an d determ ine
their histological diversity w ithin different b o d y parts and across taxa. This is
accom plished by histological sam pling of tendons from anim als th a t norm ally
experience intratendinous ossification and from anim als th at possess abnorm al
ossified tendons. Tendons from living archosaurs (crocodilians an d birds) are
sam pled to provide a phylogenetic basis for histological interpretations.
M aterials and M ethods
Fossil tendons w ere sam pled from groups (Table 2) as b ro ad ly as possible
w ith in D inosauria (Fig. 11). Fossilized ossified tendon sam ples consist largely
of associated m aterial from the vertebral colum n of ad u lt specim ens, b u t several
(i.e. M N N TIG6 an d MOR-794) w ere articulated w ith the skeleton. A ll fossilized
tendons w ere sectioned across their m id-shaft w ith the exception of those
45
Figure 11. Phylogenetic fram ew ork used in this study. Stem -based groups are labeled
at the branches and node-based groups are labeled at nodes. I, A rchosauria; 2,
D inosauria; 3, O rnithischia; 4, N eornithischia; 5, M arginocephalia; 6, O rnithopoda;
7, H adrosauridae; 8, Saurischia; 9, Theropoda; 10, D einonychosauria; 11,
N eornithes. M odified from G authier (1986), Cracraft (1988), and Sereno (1999).
taken from MOR-794, w hich was sam pled from the dorsal, sacral, and caudal
regions along the axial skeleton. W here possible, sections from MOR-794 w ere
taken from tendon term ini and shafts across the three laterally stacked layers.
H istological procedure followed standard paleohistology techniques (Wilson
1994).
Tendons from extant taxa were taken directly from dissected specim ens
(Table 2). These anim als w ere donated to the MOR postm ortem . Meleagris
gallopavo tendons w ere sam pled from the M. gastrocnem ius, M. fibularis longus,
and M. longus colli dorsalis thoracica from 16 w eek old hens and 23 w eek old
46
Table 2. Specimens u sed in this study. A sterisks denote living taxa. Institutional
abbreviations are as follows: D enver N atu ral H istory M useum (DNM)z M useum
of the Rockies (MOR)z Royal Tyrrell M useum of Paleontology (TMP)z U niversity of
California Berkeley M useum of Paleontology (UCMP)z N ational M useum of N iger
(M N N TIG).____________________________ ___________________________________
Specimen
Specimen ID
H isto ID
-
2003-19r
DNM28
1997-9c
Outgroup (Crocodylia)
A l li g a t o r m is s ip p ie n s is *
Saurischia (Sauropoda)
C a m a r o s a u r u s le n tu s
Saurischia (Theropoda)
A llo s a u r u s f r a g i l is
MOR 693
1995-1
Ceratosaur
MNN TIG6
2003-09c
D e in o n y c h u s a n tir r h o p u s
MOR 747
2002-06
S a u r o n ith o le s te s
MOR 660
1993-2
Saurischia (Theropoda: Aves)
P o d ily m b u s p o d ic e p s *
-
2003-21r
M e le a g r is g a llo p o v o *
-
2003-20r
B u b o v ir g in ia n u s *
-
2001-10r
B r a c h y lo p h o s a u r u s c a n a d e n s is
MOR 794 ■
1999-20
E d m o n to s a u r u s
MOR1142
2001-09
H y p a c r o s a u r u s s te b in g e r i
MOR 549
T e n o n to s a u r u s t ille tti
MOR 680
1996-10
2002-04
Omithischia (Omithopoda)
Omithischia (Marginocephalia)
S ty g im o lo c h s p in if e r
UCMP 128383
1993-1
P a c h y r h in o s a u r u s c a n a d e n s is
TMP 89.55.647
2003-14c
TMP-83.36.120
2003-15c
Omithischia (Thyreophora)
E u p lo c e p h a lu s t u t u s
tom s. Bubo virginianus tendon was sam pled from the M. extensor carpi radialus.
The talor retinaculum and the M. gastrocnem ius tendons w ere sam pled from an
Alligator mississippiensis. A tendon from the M. Iongus colli dorsalis thoracica w as
sam pled from a Podilymbus podiceps.
Extant sam ples w ere fixed in 10% neutral buffered form alin for storage.
M ineralized tendons w ere n o t decalcified before sectioning. The sam ples w ere
47
stained w ith hem atoxylin and eosin (H&E), except for the Bubo, w hich w as
stained w ith toluidine blue. Im ages w ere tak en from slides u sin g an electronic
cam era. M icroanatom y term inology follows established nom enclature for bone
(de Ricqles 1980; Fancillon-Vieillot et al. 1990; Reid 1996; C urrey 2002) and
ten d o n m icroanatom y (Kannus 2000).
C om parative H istological D escription
C rocodilian Tendon H istology
C ollagen fibers consisting of fibrils (Figure 12B, fibrils n o t apparent),
arrayed longitudinally characterizes this tissue. Fibroblasts (tenocytes) ru n
along these fibers (Figure 12B), w hich are sparsely arranged an d cylindrical in
shape in longitudinal sections. Collagen fibers dem onstrate crim p com m on to
ten d o n tissue. Fibers are b o u n d together b y endotenon sheathes to form fascicles
(Figure 12A). Surrounding the endotenons is the peritenon (Figure 12A), w hich is
dem arcated by the transverse orientation of the tenocytes in this region.
Tendons in crocodilians are n o t k now n to m ineralize in norm al
developm ental processes. The sam ples from Alligator su p p o rt the claim that
crocodilian tendons are a conserved vertebrate tissue (Summers an d Koob 2002).
O rnithischian Tendon H istology
O rnithopoda W ithin H adrosauriform es, three taxa w ere sam pled,
a lam beosaur (Hypacrosaurus) and tw o h ad ro sau rs (Brachylophosaurus and
Edmontosaurus). A side from being epaxial, the exact anatom ical locations of the
tendons are u n k n o w n for Hypacrosaurus and Edmontosaurus. Tendons from the
Brachylophosaurus w ere system atically sam pled along in d iv id u al tendons from
48
Figure 12. Histological photographs of Alligator tendons. A, cross section of Achilles
tendon detailing endotenon (ET) divisions, fibroblasts (PB), and a peritenon (PT)
sheath (IOOx). B, the sam e tendon is longitudinal section (40x).
the dorsal, sacral, and caudal regions.
H adrosauriform es tendons are peripherally com posed of fibrolam ellar
bone w ith longitudinally oriented prim ary osteons. H aversian canals are also
present, becom e dense at the center (Fig. 13A), and are identical w ith sim ilar
tissue from long bone compacta. In som e larger tendons, H aversian replacem ent
is dense w ith m any generations of osteons (Fig. 13B). O steocyte lacunae and
canaliculi are not alw ays easily identifiable. Some collagen fibers are visible
betw een osteons (Fig. 13F) though the center is nearly cancellous in some
tendons. These fibers som etim es become dense tow ard the periphery, which
often contains an external fundam ental system (EPS), denoted by tightly packed
LAGs (Fig. 13A and E).
Sections taken across the three tendon layers from the Brachylophosaurus
are histologically identical. Also, there is no observable histological difference
in sam ples taken from the dorsal, sacral, or caudal regions. But, variation occurs
along the length of individual tendons. Sections taken through ten d o n term ini
49
Figure 13. Flistological photographs of H adrosauriform es tendons. A, tendon from
Hypacrosaurus (cross section, 40x). Notice m ultiple lines of arrested grow th (LAGs),
dense H aversian bone and prim ary osteons (PO). B, tendon from Edmontosaurus
(cross section, IOOx). N ote H aversian canals (HG) and collagen fibers (CF) betw een
H aversian bone. C, close u p of an H aversian canal from Hypacrosaurus denoted
by a cem ent line (CL) (cross section, 400x). D, pathological section of tendon from
Brachylophosaurus (cross section, 40x). N ote high proportion of erosian room s (ER).
E, close u p an EFS from a Brachylophosaurus tendon (cross section, IOOx). F, close
u p of collagen fibers (CF) and a prim ary osteon from a Brachylophosaurus tendon
(cross section, cross nicols, 400x).
50
Figure 14. H istological p hotographs of Tenontosaurus tendons. A, central portion of
the tendon, dense w ith H aversian canals (HG) (cross section, IOOx). N ote collagen
fibers (CF) betw een the canals. B, close u p of collagen fibers (CF) and prim ary
osteons (PO) (cross section, 400x).
are filled w ith collagen fibers and less lam ellar tissue. In Brachylophosaurus, LAGs
vary along tendon length, becom ing m ore ab u n d an t at the center.
Several sam ples w ere taken from a pathological region in the dorsal aspect
of the tail (a com m on occurrence in h ad ro sau r tails). Grossly, som e tendons in
this region continue over the pathological neural spines u n in terru p ted w hile
others are eroded. Histologically, this tissue is sim ilar to the term ini in other
Brachylophosaurus tendons (Fig. 13D). Both have large erosion room s. However,
the n u m b er and size of erosion room s in the pathological ten d o n is m uch greater.
W hile there is no evidence of the pathology's origin (e.g. bacterial infection), the
tissue suggests that active grow th and repair w as occurring at the tim e of death.
Epaxial tendons from the Tenontosaurus caudal region contain
longitudinally oriented collagen fibers and prim ary osteons in the periphery.
Several generations of H aversian canals are present at the center, w hich contains
m any large resorption cavities that form a reticular cancellous region. Like the
H ad ro sau r tendons, som e fibers are present betw een osteons even in the central
51
region w ith dense H aversian bone (Fig. 14A and B). Some ten d o n s also contain
LAGs and an EPS.
M arginocephalia Tendons from the Pachycephalosaurus contain a core
nearly identical w ith those seen in ornithopod tendons. T hat is, tissue is
characterized by an abundance of collagen fibers oriented longitudinally.
These are in terru p ted by resorption spaces, p rim ary osteons, an d H aversian
canals. Vascular orientation parallels the fiber bundles. S urrounding the core
is fibrolam ellar bone. Vascularity is reduced in this region an d has alternating
orientation from longitudinal to radial (Fig. 15A). Im m ediately outside the
core, vascularity is longitudinally oriented. M ore distally is a b a n d of obliquely
radially canals follow ed by a b an d of radially oriented canals. R adially oriented
vascular spaces dom inate the perim eter.
Spread th ro u g h the fibrolam ellar section are isolated areas of rough
b u n d le d tissue consisting of enlarged collagen fibers (Fig. 15A). These areas look
identical to the tendon's center and are globular in shape, th o u g h one is band
shaped. Osteocyte lacunae in the fibrolam ellar regions are sm aller th a n those in
regions dom inated by ro u g h bun d led collagen fibers.
Beyond being epaxial, the anatom ical location of the ten d o n from the
Pachyrhinosaurus is unknow n. A cancellous region occupies th e tendon's center
(Fig. I SB). The bone form ing this cancellous region is lamellar. The periphery
consists of fibrolam ellar tissue, w ith longitudinal vascularity (Fig. 15C).
M ultigenerational H aversian system s are p resen t at the b o rd er betw een the tw o
bone regions, b u t are obliterated centrally by the lam ellar cancellous tissue (Fig.
15D). Lines of arrested grow th are absent.
52
'
;•
I i t -
'
A-
(I
Figure 15. Histological photographs from m arginocephalians. A, com posite im age
from Stygimoloch (cross section, 40x). N ote alternating longitudinal (LC) and radial
(RC) canals. Also, note the central core of H aversian bone (HE). B, Pachyrhinosaurus
tendon (cross section, I Ox). C, sam e tendon at 40x. N ote the outer region of
fibrolam ellar bone (FLB) and interior region of lam ellar bone (LB). D, sam e tendon
at IOOx in the lam ellare region.
53
Figure 16. Histological photographs of Euplocephalus tendons. A, the w hole tendon
in cross section (IOx). B, detail of periphery in the same tendon (IOOx). N ote how
closely m ultigenerational H aversian canals (HG) extend to the surface. Also, note
that the prim ary tissue contains m any lines of arrested grow th (LAG). Erosian
room s (ER) are also present.
Thyreophora Samples from an Euplocephalus (A nkylosauridae) tail club
consist of dense H aversian bone (Fig. 16A). Some erosion room s are dispersed
through the H aversian bone. Vascular orientation is longitudinal. H aversian
system s are so developed that no trace of collagen fibers are p resen t except close
to the surface (Fig. 16B). LAGs are present in the peripheral region. Osteocyte
lacunae are larger in this region and circum ferentially oriented.
Saurischian Tendon H istology
Sauropoda The abnorm al Camarosaurus tendon contains prim ary collagen
fibers and a few prim ary osteons in the periphery (Fig. 17A). This region is not
uniform circumferentially, and in some areas it consists of m ultiple generations
of H aversian canals. O ther portions of the periphery are filled w ith relatively
avascular rough b undled tissue. Dense H aversian tissue occurs th ro u g h o u t the
center and lacks collagen fibers betw een canals. The tendon's central p art also
contains resorption cavities. N o LAGs are present in the periphery. Osteocyte
54
Figure 17. H istological photographs of Camarosaurus tendons. A, cross section of
tendon (40x). N ote the peripheral region of com rpised of collagen fibers (CF) and
the central core of H aversian bone (HE). B, close u p of central region (cross section,
IOOx) N ote collagen fibers (CF) betw een som e H aversian canals (H C).
lacunae are not specifically shaped and canaliculi are present as thin filaments.
Non-A vian T heropods Several abnorm al ossified tendons from the sacral
neural spines of an u nnam ed ceratosaur ap p ear grossly sim ilar to other ossified
tendons sam pled in this study. These are preserved sections of the m uscle-tendon
system of the M. transversospinalis (see chapter I) and thus hom ologous to the
ossified tendons in H adrosauriform es. Histologically, m ineralized collagen fibers
oriented longitudinally com prise the tendon (Fig. 18A). The perip h ery is m arked
by LAGs (Fig. I SB). Some H aversian canals are present centrally. Fibers are large
and rough b u n d led interior to the first LAG and fine outside of it. This change in
fiber structure m ay indicate the boun d ary of the original tendon. If so, this w ould
indicate that intratendinous ossification did not develop in term s of onset and
offset like ossified tendons in O rnithischians (see chapter 2).
Allosaurus tendons possess a periphery consisting of vascularized
fibrolam ellar bone tissue (Fig. ISC). Vascular orientation is longitudinal
55
Figure 18. H istological photographs of non-avian theropod tendons. A, cross
section across n eural spine (NS) and ossified tendons (OT) in a ceratosaur (IOx). B,
close u p of a tendon w ith H aversian canals (HG) (IOOx). N ote the ro u g h bu n d led
apperance of the fibers interior to the first line of arrested grow th (1st LAG). C,
an Allosaurus tendon (cross section, 40x). N ote the fibrolam ellar structure w ith
prim ary osteons (PO) and an EPS. D, central region of the sam e tendon (IOOx).
N ote the erosian room (ER), H aversian canals (HG), prim ary osteons (PO), and
centrally located LAGs.
th ro u g h o u t the tendon. Prim ary osteons w ith in a longitudinally arranged
fiber m atrix characterize the periphery. Centrally, it is com posed of m ultiple
generations of m uch larger H aversian canals and resorption room s (Fig. 18D).
The size of the latter form s a central cancellous region. Osteocyte lacunae are
generally ovoid in the fibrous tissue and flattened in lam ellar tissue. Lines of
arrested grow th occur throughout the tendon. The interior m ost line occurs just
beyond the central H aversian region. Lines of arrested grow th occur sporadically
56
tow ards the periphery, becom ing tightly packed at the subsurface form ing an
EPS.
Tail ro d bundles from Saurornitholestes are poorly preserved. They consist
of dense H aversian bone (Fig. 19A). These reach the surface in som e places in all
tendons. Few resorption cavities are present. The small perip h ery is filled w ith
longitudinally arranged collagen fibers. M ultiple LAGs are visible in all tendons.
These lines becom e tightly packed near the surface indicating an external
fundam ental system.
The tail rods (elongated prezygapophyses and chevrons) from
Deinonychus contain tightly spaced rings consisting of ordered avascular fibrous
tissue (Fig. 19C). This is fine structured tissue w ith sparse osteocyte lacunae. A
few p rim ary osteons are present in som e rods, th ough tw o contain concentric
rings of prim ary osteons in the periphery. Several other rods (Fig. 19B) contain
central H aversian systems. They are large an d adjacent to the vertebral b o d y and
have longitudinal orientation. D istal to the vertebrae, the ro d s becom e avascular.
Centrally, all rods contain a loosely packed fiber core (Fig. 19D).
N eornithes Ossified tendons in Meleagris contain a m ineralized fiber
core in w hich collagen fiber bundles are still ap p aren t in cross section (Fig.
20A). These bundles are scallop-shaped and separated from the su rro u n d in g
m atrix. These give the m ineralized ten d o n a scaly appearance. In tendons that
are heavily ossified (due to increased age or variation w ith in an individual), the
spaces am ong the scalloped fiber bundles is filled in w ith a n ew osteoid m atrix
giving the tissue a m ore fascicular appearance. Centrally, large irreg u lar erosion
room s are com m on. A few H aversian system s are present. Meleagris tendon
histology from this stu d y agrees w ith p ublished descriptions (L ieberkuhn 1860;
57
Figure 19. H istological photographs of non-avian theropod tendons "tail rods"
or ossified tendons. A, cross section from Saurornitholestes (IOx). Taphonom ic
alteration m akes detailed interpretation difficult, b u t note the dense H aversian
bone. B, ossified tendons and vertebral body (VB) of Deinonychus (cross section,
I Ox). N ote that som e exterior tendons are vascular. These are proxim al sections
close tho their origin. C, close u p of a tendon (cross section IOOx). N ote the fine
grained rings and a core of loosely arranged fibers w ith tw o H aversian canals
(HG). D, close u p u p central region (cross section, cross polar, 400x). N ote the
uniform arrangem ent of longitudinally oriented prim ary fibers.
W eidenreich 1930; A bdalla 1979).
Longitudinal sections clearly show m ineralized areas w ith fibers that
have lost their characteristic w aviness (crimp) and become straight (Fig. 20B).
Fibroblasts in m ineralized regions are hypertrophied and contain large am ounts
of cytoplasm . These cells appear sim ilar to hypertrophyied chondrocytes.
58
Figure 20. H istological photographs of N eornithes ossified tendons. A, Meleagris
tendon containing a m ineralized core su rro u n d ed by a p araten o n (PT) (cross
section, 40x). N ote the m atrix of m ineralized collagen fibers (MF), H aversian canals
(HG), and ersoian room s (ER). B, longitudinal section of sam e ten d o n (IOOx). N ote
the exterior unm ineralized fibers (UF) and the straightened m ineralized fibers (MF)
th at contains h ypertrophied fibroblasts (HF). C, tendon from Bubo (cross section,
I Ox). N ote the thick peritenon (PT) su rro u n d in g the osseous core (OC). D, close
u p of the core in the sam e tendon (IOOx). E, tendon from Podilymbus (longitudinal
section, cross polar, IOOx). N ote the chain of hypertrophied fibroblasts (HF). F,
another Podilymbus tendon (longitudinal section, 400x). N ote obliquely angled
m uscle tissue (M) attaching directly onto the surface of the ossified tendon (OT).
59
■Long chains of h ypertrophied fibroblasts ru n do w n the long axis. As w ith the
cross-sections, the central area is m ineralized w ith straight fibers. This core is
su rro u n d e d by unm ineralized tendo n tissue w ith Wavy fibers an d elongated
tenocytes.
Some tendons (i.e. m any shank tendons) extend the zone of
transform ation into the m yotendinous p a rt to produce a fan-shaped
m yotendinous origin. The zone of transform ation also extends into th e tendon
entheses (bone insertions) in m any epaxial tendons, w hich ossify to th e neural
spine sum m its. D espite this difference, there is no difference in m icroanatom y
betw een leg and epaxial tendons.
Thin-sections from Bubo appear sim ilar to those o£ Meleagris, b u t w ith a
very thick peritenon. These thin-sections w ere taken from I cm proxim al to the
M. extensor carpi radialis insertion point. A clear dem arcation occurs betw een
the central osseous core and the unm ineralized peritendon (Fig. 20C). The central
region consists of prim ary osteons and first generation H aversian canals (Fig.
20D). Between som e H aversian canals m ineralized collagen fibers can be seen.
C ollagen fibers have the sam e scalloped shape and overall b u n d le d appearance
seen in the turkey. LAGs and prim ary osteons are absent from this tissue.
Lateral tendon sections from the M. Iongus colli dorsalis thoracica in
Podilymbus are com parable w ith those from Meleagris. For exam ple, collagen
fibers ru n straight along the longitudinal axis and chains of h y p ertro p h ied
tenocytes are com m on (Fig. 20E). Fibroblast shape is fusiform w h en they are n o t
in chains. But, ossified tendons from Podilymbus lack an unm ineralized outer
tendinous sheath (peritendon). This is evident w here the M. Iongus colli dorsalis
thoracica interfaces directly w ith the ossified ten d o n w ith o u t an unm ineralized
interm ediate (Fig. 20F). Therefore, the original m yotendinous junction is replaced
60
by a fleshy attachm ent along the length of the tendon.
D iscussion
The Alligator and unm ineralized Meleagris tendons are indistinguishable.
Furtherm ore, archosaur ten d o n m icroanatom y appears identical w ith norm al
ten d o n fo u n d in m am m als (Kannus 2000). This conform s to th e claim th a t basic
ten d o n structure is conserved w ithin vertebrates (Summers and Koob 2002). It
also establishes a basis from w hich intratendinous ossification in dinosaurs can
be evaluated (i.e. norm al unm ineralized ten d o n is plesiom orphic for D inosauria).
O ssified Tendon D iversity and Evolution
D espite the general hyperostosis found in the non-avian d in o sau r ossified
tendons and the further osseous developm ent seen in the m arginocephalians,
ossified tendons have m any characteristic features. M any w orkers describe
these features in birds and non-avian dinosaurs (Lieberkuhn 1860; Broili 1922;
M oodie 1928; K oehnlein 1930; W eidenreich 1930; Johnson 1960; A bdalla 1979).
For exam ple, L ieberkuhn (1860) identifies chondrocytes in b ird tendons, later
correctly term ed h ypertrophied tenocytes b y R etterer and Lelievre (1911). These
cells are responsible for the m ineralization process. Lieberkuhn (idem) also
describes the straightened tendon collagen fibers in longitudinal section and
the scalloped fascicular appearance of fiber b u n d les in cross section. Indeed,
longitudinally isotropic fibers and vascularity are characteristic of ossified
tendons. This orientation is due to the cord-like shape of the te n d o n precursor
tissue.
The other feature th a t is characteristic of ossified tendons is tenocyte
61
hypertrophy. This cellular m orphology occurs in ossified tendons, cartilage
caps, and ten d o n and ligam ent entheses. (Haines and M ohuiddin 1968).
H aines and M ohuiddin (Idem) identified this phenom enon in tenocytes on
the unm ineralized side of the blue line (m ineralization tide m ark) in tendon
entheses. There h y pertrophied fibroblasts occur in long chains, a structure th at
appears in ossified tendons as well (Fig. 20B an d E). A lthough these cells are still
called chondrocytes due to their cellular m orphology (e.g. Suzuki et al. 2002), this
in inappropriate (Retterer and Lelievre 1911; Landis and Silver 2002).
H aines and M ohuiddin (Haines and M ohuiddin 1968) also note that
osteocyte canaliculi in m etaplastic bone are absent, stunted, or looped. This
feature w as observed in som e ossified tendons in this study. B ut this type of
canaliculi is restricted to m ineralized ten d o n tissue. The lam ellar bone (osteons)
form ed in ossified tendons is indistinguishable from norm al bone (Lieberkuhn
1860; W eidenreich 1930). Thus, highly rem olded ossified tendons lose all traces
(hypertrophied tenocytes, collagen fibers, and absent or looped canaliculi) of
their ancestral tissue. The only indicative feature left in these cases is isotropic
longitudinal vascular orientation, w hich is a p ro d u ct of the tendon's chord-like
gross m orphology.
As h as been noted (Lieberkuhn 1860; Retterer and Lelievre 1911), the
bone laid do w n d u ring m etaplasia is indistinguishable from n o rm al bone tissue.
Furtherm ore, ossified tendon's histological features are uniform across different
anatom ical regions. There is essentially no difference in m icroanatom y betw een
leg tendons and those from the back. The w in g tendon m icrostructure of Bubo is
also sim ilar to Meleagris ten d o n m icrostructure. However, the p eritenon is m uch
thicker, form ing a solid b order around the osseous core. Given the developm ent
of H aversian bone, an indicator of late stage intratendinous ossification, a thick
62
p eriten d o n m aybe a characteristic of w ing tendons. A natom ical location of the
ten d o n and associated w ing function m ay account for the thickened peritenon or
it m ay be taxic variation.
U nlike other birds, Podilymbus ossified tendons do n o t possess an
unm ineralized peritenon. This difference m ight be caused b y age-related
differences am ong specim ens. H ow ever, it show s th a t the M. Iongus colli dorsalis
thoracica can directly interface w ith the ossified ten d o n b y a fleshy attachm ent.
The hom ologous ossified tendons from Meleagris do n o t interface in a direct bone
to m uscle attachm ent. Since ossification begins centrally an d spreads outw ard
(Johnson 1960) and all ornithischian ossified tendons have LAGs, they probably
ossified the entire tendon, then grew b y a periosteum like endochondral
bones (see chapter 2), thus interfacing w ith epaxial m uscles like those seen in
Podilymbus. Because H adrosauriform es epaxial tendons are hom ologous w ith
the epaxial ossified tendons in birds (see chapter I), this is im p o rtan t for epaxial
m uscle reconstruction. Furtherm ore, the sam e type of m uscular interface can
be u sed to infer reconstruction from other epaxial ossified tendons fo u n d in
O rnithischia, such as parallel tendons in Tenontosaurus.
The high density of H aversian canals an d presence of LAGs, prim ary
osteons, and an EFS indicate th at the bone com prising the ten d o n is as old as
the skeletal elem ents. Histological stu d y of ossified tendons sh o u ld no t rely on
indiv id u al sam ples because variance in histology along ten d o n len g th could
dram atically alter interpretations. For exam ple, a fragm ent from the center of an
ossified tendon w ou ld look histologically m u ch older th an a sam ple from near a
term inus.
Sam ples taken from b o th m arginocephalians (Pachycephalosaurus and
Pachyrhinosaurus) are better ossified th an tendons from all o ther dinosaurs.
63
Pachycephalosaurus tendons possess a loosely packed core consisting of H aversian
bone, and longitudinally oriented collagen fibers. Islands of this bone type
are found th ro u g h o u t the outer region, w hich is form ed from tightly packed
fibrolam ellar tissue. The vascular orientation in the outer region alternates
betw een longitudinal and radial. This is the only know n ossified ten d o n (living
or extinct) to have such a structure. All other ossified tendons are extrem ely
isotropic, w ith vascularity and collagen fiber orientation parallel to th e long axis.
This sam ple m ay reveal a developm ental constraint. That is, all other ossified
tendons are constrained in their collagen an d vascular orientation due to the
u n d erly in g isotropic ten d o n structure and function. For som e u n k n o w n reason,
vascular orientation alternates in Pachycephalosaurus, suggesting a freedom from
this constraint. Alternately, radial vascular orientation could indicate a reduction
in axial stress d u rin g appositional osseous grow th.
The other m arginocephalian ten d o n (Pachyrhinosaurus) w as also m ore
rem odeled th a n other tendons exam ined in this study. W hile it does no t contain
the alternating vascular structure found in the pachycephalosaur, it possesses
a cortical region filled w ith fibrolam ellar bone of anisotropic fibers and a
m edullary region com posed of lam ellar tissue. U nlike non-m arginocephalians, it
lacks uniform ly oriented collagen fibers. In fact, there are no histological features
to suggest th a t this tissue is an ossified tendon. Ossification developm ent in
this sam ple is unusual, b u t not unique. Therefore, w hile the Pachyrhinosaurus
ten d o n suggests greater intratendinous ossification m aturity in ceratopsians,
developm ental plasticity and a sam ple size of one m akes interpretations
equivocal at the present time.
64
Tail Rod Tendons
Prezygapophysis elongation is a progressive tren d in Tetanurae (G authier
1986). Thus, it w ould seem th a t the "tail ro d tendons" of M aniraptorans are
ju st prezygapophysis extensions. In this scenario, they should be form ed in a
m an n er sim ilar to vertebral bone. However, the prezygapophysis and chevron
extensions of Deinonychus have been described as developing th ro u g h tendon
sheaths taking on periosteal functions (O strom 1969; O strom 1990; Reid 1997).
C onsidering the intratendinous ossification process in birds (Johnson 1960;
A bdalla 1979; Landis and Silver 2002) and h ad ro sau rs (see chapter 2), this seems
unlikely since m etaplasia begins centrally in the tendon. The p erip h ery is the
last region to m ineralize and ossify in norm al intratendinous ossification. But
this is n o t the case in the tail rod tendons. Indeed, the opposite is true. They
lose vascularity distal to the vertebral b o d y an d have an outer region of fine
fibrous bone an d a central loosely organized fiber core. This u n u su al histology
is m ost likely the result of sm all am ounts of appositional g ro w th com pared w ith
longitudinal grow th. Because there are no extant analogues for com parison, the
exact developm ental n atu re of these tail rods rem ains unknow n. But, given that
prezygapophysis elongation is a tren d in Tetanurae (G authier 1986), they are
probably periosteal bony rods.
C onclusion
In their review p ap er on ossified tendons in birds, V anden Berge and
Storer (1995) find no physiological or biom echanical function th a t accounts
for their sporadic appearance in Aves. The current study, w hich includes nonavian dinosaurs and birds, did not bring clarity to the phylogenetic p attern of
65
in tratendinous ossification. The Bubo w ing ten d o n contains an p eriten o n sheath
m uch larger th an other ossified tendons. H ow ever, because Bubo does no t have
ossified tendons in other areas of the b o d y for com parison, it is u n k n o w n if this
feature is phylogenetic or functional. W ithin individual Meleagris, there is no
variation in term s of m etaplastic developm ent or histological structure. These
d ata seem to contradict the com m on notion th a t tendons ossify in response
to biom echanical stress (Landis and Silver 2002), w hich w ould suggest that
different tendons th ro u g h o u t the bod y ossify at different rates as they are
subjected to different tensile stresses. Thus, caution should be u sed w h en
inferring the adaptive significance of intratendinous ossification in birds.
Inferring intratendinous ossification function from extinct dinosaurs is even
m ore problem atic.
H istological features also seem to be free of body size related constraints.
That is, the histology in avian taxa ap p ear virtually identical (save the expanded
peritenon in Bubo) despite various bo d y sizes. The sam e is tru e of non-avian
dinosaurs, from the u n n am ed ceratosaur to Tenontosaurus. Thus, size m ediated
ossification, th ro u g h progressively greater tensile stress application does no t
seem to affect intratendinous ossification.
The only discernable phylogenetic difference in ten d o n histology is the
relative degree of osseous developm ent. All extinct dinosaur ten d o n s exam ined,
including the abnorm al ossified tendons, are better ossified th a n those in
living birds. M ost non-avian dinosaur ossified tendons possess LAGs, prim ary
osteons, and m any have an EPS, indicating the presence of a periosteum . W ithin
O rnithischia, m arginocephalians appear to experience even greater osseous
developm ent in epaxial tendons. This condition is u n k n o w n in living animals.
Therefore, ossified tendons in extinct non-avian dinosaurs reveal a physiological
66
process th a t w o uld rem ain un k n o w n by studying living anim als alone.
N o differences w here detected in abnorm ally (i.e. Allosaurus
or Camarosaurus tendon) vs. norm ally ossifying tendons. A nd, ten d o n
m icrostructure w as hom ogenous across th e trellis layers in a Brachylophosaurus.
But, LAGs w ere found to vary w ithin tendons along their length. It is likely
th a t original ten d o n fibers are abundantly preserved in a ceratosaur as well as
betw een central H aversian canals in m any non-avian d in o sau r tendons.
M aniraptoran tail rods do not appear to be a unique form of
intratendin ous ossification. The fine hbered structure su rro u n d in g a loosely
organized core w as n o t encountered in any other taxa, living or extinct. A
histological survey of Tetanurae zygapophyses could help illum inate the tissue's
developm ental nature. Tail tendon ro d developm ent m ay also have functional
im plications because of its im pact on caudal m uscle structure. F urther research
is needed to reconstruct the caudal m uscles associated w ith these rods in
m aniraptorans!
Classically, ossified tendons have been categorized as "grobgebundelte
grund su b stan z - faserknochen" or "ro u g h b u n d led bone tissue" b y W eidenreich
(1930). This categorization seems appropriate due to the fact th a t all bone is
m etaplastic. N o data from living or extinct dinosaurs w ere fo u n d to argue w ith
W eidenriech's designation.
67
CHAPTER 4
OSSIFIED TENDON BIOMECHANICS IN ORNITHOPOD DINOSAURS
Introduction
H adrosauriform es possess a bony rhom boidal trellis along the dorsal
aspect of the spinal colum n (Fig. 21). It is form ed by ossified ten d o n s from
the M. transversospinalis (see chapter I). A lthough the trellis is p resen t in all
archosaurs, an ossified tendon trellis is only k n o w n in H adrosauriform es (Sereno
1986; Sereno 1999) and som e birds, such as grebes (see chapter I). The trellis in
H adrosauriform es has p ro m p ted m any functional interpretations, w hich m ay
be divided into aquatic and terrestrial speculations. Various authors (Brown
1916; Colbert 1951) use the trellis to argue th at hadrosaurs w ere semi-aquatic.
The ossified ten d o n trellis is claim ed to have functioned by p ro v id in g additional
su p p o rt and pow er to lateral m ovem ents of the tail w hile sw im m ing.
H adrosauriform es are now th o u g h t to be terrestrial w ith horizontal
posture (O strom 1964; G alton 1970; N o rm an 1980; N orm an 1986). In the first
terrestrial interpretation, the trellis (in an Iguanodon bernissartensis) is suggested
to stiffen the body about the sacrum , counteracting dorsoventral loading from
the M. caudofem oralis longus (Dollo 1886). Dollo {idem) interprets these anim als
standing in an erect posture w ith the tail on the ground. A lthough iguanodons
and h ad ro sau rs are currently reconstructed w ith a horizontal po stu re (Galton
1970), the caudofem oral m uscles w ere still probably im portant for locom otion as
the prim ary fem oral retractor (Gatesy 1995).
A nother terrestrial interpretation suggests th at the ossified tendons trellis
68
I meter
Figure 21. R econstruction of Brachylophosaurus (on top) and Tenontosaurus (bottom).
The relative sizes depicted are based on MOR 794 and MOR 682, the specim ens
used in this study.
in H adrosauriform es resisted sagging of the body about the hips (Ostrom 1964).
The trellis arrangem ent is used as an argum ent for terrestrial habits because of its
sim ilarity to a balanced cantilever bridge, w hich bears w eight th ro u g h loading
cross m em bers. Intratendinous ossification is used as prim ary su p p o rtin g data
for this argum ent, because bone has greater tensile strength th an tendon. This
argum ent also uses H adrosauriform es taphonom y (their vertebral colum n is
usually preserved "ram -rod" straight) as su pporting evidence (Lull and W right
1942; Bultynck 1992). A nother argum ent against the sem i-aquatic interpretation
suggests that the trellis produced m ediolateral tail stiffness (N orm an 1986).
O ther authors claim that ossified tendons did not stiffen the vertebral
colum n at all by noting that long tendons can ossify w hen transm itting large
69
Figure 22. Trellis of ossified tendons from the posterior dorsal region of a
Brachylophosaurus (MOR 794). A nterior is to the right.
loads (Coombs 1995). This is thought to im prove m uscle-tendon function by
increasing tendon strength and reducing strain. Coombs (idem) notes that m any
birds have ossified tendons and proxim ally located muscles in their legs. But
som e birds also have ossified tendons th ro u g h o u t the tru n k and neck (Vanden
Berge and Storer 1995) w here tendon length is not always greater than muscle
length. A nd, there is no functional pattern in birds. For exam ple, ow ls and cranes
possess ossified tendons b u t their sister groups lack them (Vanden Berge and
Storer 1995). In neither case do these birds appear to exhibit different behaviors
that w ould explain intratendinous ossification (Vanden Berge and Storer 1995).
The ossified tendon trellis in H adrosauriform es has been explained
70
Figure 23. Phylogenetic fram ew ork used in this stu d y based on Sereno (1999). N odebased taxa are as follows: I, O rnithopoda; 2, Ankylopollexia; 3, H adrosauriform es.
Bold taxa are m odeled in this study.
using functional m orphology as restricting and prom oting lateral m otion of the
tail. A rgum ents like these, w ithout biom echanical testing, have been used to
justify contradictory claims. Specifically, ossified tendons have been used as an
argum ent for and against sem i-aquatic hypotheses. Biomechanical analysis of
ossified tendons in dinosaurs m ay provide a m ore robust m ethod for evaluating
their function.
This study has three prim ary aims. The first is to estim ate the
biom echanical influence of ossified tendons in basal ornithopods w ith parallel
ossified tendons. The second is to estim ate the biom echanical influence of the
ossified tendon trellis in H adrosauriform es. Two hypotheses are evaluated for
b oth taxic m odels. The first is Dollo's (1886) assertion that the trellis acted like
71
a brace to resist caudofem oral loading d u rin g locom otion (fem oral loading
hypothesis). The second is O strom 's (1964) hypothesis th a t the ossified tendon
trellis passively reduced sagging (hogging) about the hips (hogging hypothesis).
Both hypotheses are evaluated using static finite elem ent com puter m odeling
to determ ine the biom echanics of the plesiom orphic parallel n etw o rk a n d the
apom orphic ossified tendon trellis.
The th ird aim is to determ ine functional differences betw een the m odels
in a phylogenetic fram ew ork. This com parison is u sed to elucidate possible
functional ossified tendon evolution in O rnithopoda. Recent paleo-biom echanical
studies evaluate their m odels using phylogenetically extant relatives for w hich
em pirical d ata can be collected (Gatesy 1995; C arrano and Biewener 1999;
R ayheld et al. 2001; H utchinson and Garcia 2002). Such a phylogenetic approach
is m ore robust for testing functional hypotheses about extinct organism s
(W eisham pel 1995; W itm er 1995). This stu d y includes experim ental data from an
Alligator mississippiensis for m odel developm ent.
M aterials and M ethods
Two articulated specim ens from the collections at the M useum s of the
Rockies (MOR) w ere u sed to create the d in o sau r m odels in this study: a prim itive
(Tenontosaurus tilletti, M OR 682) and derived (Brachylophosaurus canadensis, M OR
794) ornithopod. M OR 682 possesses a com plete tail and M OR 794 a roughly 90%
com plete tail. O nly preserved elem ents w ere used. A th ird m odel w as created
from a m o u n ted juvenile Alligator (MOR OST 130) skeleton.
Skeletal elem ents w ere m easured directly from specim ens w ith m easuring
tape, calipers, an d a protractor. Individual centra w ere m easu red for length
72
(anteroposterior), height (dorsoventral), and w id th (m ediolateral). The neural
spines and chevrons w ere m easured for length (anteroposterior), height
(dorsoventral), w id th (mediolateral), and inclination (from the vertical axis).
In som e cases (i.e. the Brachylophosaurus sacrum ), w id th h a d to be estim ated as
an average of the surro u n d in g vertebrae because sections w ere im bedded in
m atrix. These m orphologic d ata w ere u sed to create three dim ensional finite
elem ent m odels in Visual A nalysis 3.0 (IBS 2000), a structural engineering
program . D orsoventral corresponds to the Y-axis, m ediolateral to the Z-axis, and
anteroposterior to the X-axis. C entra w ere m odeled as short cylinders w hose
circular cross-sectional diam eter was determ ined b y the average of w id th and
h eight m easurem ents.
P in joints and fixed joints w ere u sed in Visual Analysis. Joints are fully
constrained by default, allow ing full force and m om ent transfer betw een
m em bers (IBS 2000). The neural spines w ere attached to the centra b y fixed
connections th at do not allow translation or rotation. This w as done to m odel
the fused neu ral spine and centrum . Individual vertebrae, chevrons, tendons,
an d ligam ents w ere joined to one another b y p in connections th a t allow
rotation betw een m em bers. Such non-fused joints w ere partially constrained,
m eaning th a t less th an 100% transfer of m om ents and forces occur (IBS 2000).
V isual A nalysis approxim ates a partially constrained joint th ro u g h endzones.
Endzones are short regions at m em ber ends th a t have altered stiffness (E) values.
D eterm ining endzone properties for the dinosaur m odels w as estim ated from
experim ental data of the Alligator (see below).
N eu ral spine apical ends w ere connected w ith m em bers th a t h a d the
cross-sectional area of tendons. The m odels w ere su pported at every sacral
centrum by fixed supports th a t do no t allow translation or rotation. All elem ents
73
A
L (Neural Spine)
B
W (Neural Spine)
— H ( Neural Spine)
— H (Centrum)
L (Centrum)
W (Centrum)
Figure 24. D iagram of m odeled vertebra w ith indicated dim ensions in lateral view
(A) and frontal view (B). Closed circle indicates a fixed connection.
w ere given bone m aterial properties (see below), except the ligam entum apicum
dorsalis, w hich was given the m aterial properties of tendon.
Two load cases w ere then applied to each model. Each load case consisted
of tw o point loads: one applied to the dorsal region and one to the tail region
on the centra. All loads had a -y (ventral) direction and 2.0 kN m agnitudes in
the first case and 10.0 kN in the second case. Only tw o load m agnitudes were
used because the m odels w ere linear (uniform geometries, m aterial property
distributions, and load applications). Skeletal w eight was autom atically
calculated by Visual A nalysis and included in all load cases. The results from
these tw o load cases w ere used to generate data at I kN intervals by linear
extrapolation.
In each case, loads w ere placed at the center of m ass (COM) for each
region. A cylinder was assum ed for the dorsal region and a cone for the tail
74
region. In the brachylophosaur m odel, the dorsal loads w ere placed on the ten th
dorsal v ertebra (1.0 m posterior from the first dorsal) and tail loads on the ten th
caudal vertebra (3.4 m posterior from the first dorsal). In the ten o n to sau r m odel,
the dorsal loads w ere placed on the ten th dorsal vertebra (0.4 m posterior from
the first dorsal) and tail loads on the fifteenth caudal vertebra (1.8 m posterior
from the first dorsal).
The preceding steps w ere u sed to produce three m odel variants for each
dinosaur. O nly the vertebral colum n and ligam entum apicum dorsalis com prised
the first m odel variant. Tendons w ere ad d ed to the vertebral colum n in the
second m odel variant. Based on the conservative structure of epaxial m uscle
tendons in non-avian dinosaurs (see chapter I), parallel tendons from the M.
Iongissim us dorsi and a trellis from the slips of the M. transversospinalis w ere
included in b o th dinosaur m odels. The tenontosaur m odel also included parallel
tendons in the ventral aspect of the tail. In the th ird m odel variant, the tendons
w ere ossified by giving them bone m aterial properties. In the ten o n to sau r m odel
only the parallel tendons w ere ossified in contrast to the brachylophosaur m odel
w here only the trellis w as ossified. C om parison am ong the three m odel variants
for each anim al allow ed phylogenetic, sensitivity and functional analyses.
U ltim ate stress values w ere then u sed to determ ine w h en individual
m em bers w ould fail (see below). M em ber failure was used to indicate the
m axim um load th a t the m odel could w ith stan d before structural failure. Because
m uscles w ere n o t included the actual bearable loads w ere u n d o u b ted ly greater
th an in these m odels. Therefore, the results are interpreted in term s of relative
spinal colum n deflection, increased ability to resist failure, an d sensitivity to
m aterial properties, not absolute values.
The alligator m odel consisted of only th e tail because experim ental
75
deflection d ata could n o t be obtained from the dorsal region d u e to struggling.
The load (tail w eight) w as placed on the fifteenth caudal vertebra (0.1 m posterior
from the first caudal) in the alligator m odel. O nly one load case w as used, w hich
corresponded to the live alligator's tail w eight. This was d eterm ined by using
the circum ference at the base of the tail (13.5 cm) and the tail len g th (37.0 cm) to
calculate volum e (assum ing a cone). A ssum ing the density of w ater (I g/cm3),
an d gravitational acceleration (9.8 m/s2), tail m ass w as calculated to be 179.0 g
an d its w eight 1.8 N. This value is appropriate given the anim al's w eight of 730.0
g the day the deflection data w ere taken. Tail deflection data w ere taken from
the alligator by allow ing its tail to hang over the table's edge. The skeleton's tail
length u sed to construct the m odel was 35.0 cm and the live alligator's tail length
w as 37.0 cm.
M aterial Properties
Three factors w ere u sed in choosing m aterial properties from the
literature. The first w as to choose bones from sim ilar anatom ical regions. The
second w as to choose bones w ith potentially sim ilar functions. The th ird was to
select bone properties from phylogenetically related taxa over those of distantly
related taxa. M aterial properties for bone an d ten d o n are com m on in the
literature, th o u g h few studies investigate non-m am m als (A lexander 1988; C urrey
1999a; C urrey 1999b; Cubo and Casinos 2000). Tendon m od u lu s of elasticity (E)
values for various m am m als range from I GPa (Rigby et al. 1959) to 1.7 GPa,
w ith a m ean of 1.5 GPa (Bennett et al. 1986). A n average E value of 1.6 GPa
has also been suggested (A lexander 1988). Uncalcified avian tendons possess
m aterial properties sim ilar to those in m am m als (Bennett and Stafford 1988). The
relatively stiff value of 1.6 GPa for E w as u sed because it is less likely to inflate
76
differences betw een bone and tendon m aterial properties.
N o m aterial properties on vertebrae from living archosaurs have been
rep o rted in the literature, so data from h u m an vertebrae w ere u sed. The .
m o dulus of elasticity for h u m an vertebral bone is highly variable (Roy et al.
1999). The low er end of E for hu m an vertebrae ranges from 2.5 GPa, w hile the
u p p e r end has been reported as h igh as 25.8 GPa (Guo 2001). A n E value of
6.6 GPa for h u m a n vertebrae (Ladd et al. 1998) w as chosen for its interm ediate
value. This range seems reasonable, given th a t the elastic m o d u lu s in bending
w as determ ined to be 12.0 GPa for Alligator fem urs and 5.6 GPa for Crocodylus
frontals (C urrey 1988). Various bones from 29 avian species range from 7.524.0 GPa (Cubo and Casinos 2000). Like the elastic m odulus u sed for tendons,
this relatively flexible value for bone (6.6 GPa) is less likely to overestim ate the
differences betw een bone and tendon.
The elastic m odulus for ossified tendons has been d eterm ined to be 17.7
GPa in Grus antigone (Currey 1999a). H ow ever, the value u sed for bone in this
stu d y w ill also be u sed for ossified tendons for consistency of properties across
elem ents.
O ther m aterial properties w ere required in Visual A nalysis, such as
density and Poisson's ratio. A v alu e of 0.25 (Ladd et al. 1998) w as u sed for
Poisson's ratio and 7.8 kN /m 3 (Nigg and H erzog 1999) for density. H ow ever,
w h en changed to double and half of these values, neither w as fo u n d to affect the
m odel's final results.
U ltim ate Stresses
Bone's ultim ate bending stress (138.6 MPa) w as calculated b y m ultiplying
the m o d u lu s of elasticity for bone u sed in this stu d y (6.6 GPa) b y th e ultim ate
77
strain (0.02) for Alligator fem ur in tension (C urrey 1990). A nother, nonlinear,
m ethod to calculate the ultim ate bending stress derived from em pirical data
(C urrey 1999a) produced a value of 75.1 M Pa using the equation BS = 15.1E0.85.
This value is low com pared w ith bending values found in avian long bones th at
range from 148.8 - 201.10 M Pa (Cubo and Casinos 2000), h u m a n fem ora (208.6
MPa), an d bovine (223.8) tibia (M artin et al. 1998). A n ultim ate b e n d in g stress
of 138.6 M Pa w as u sed because the initial calculation was m ore consistent w ith
m ost published d ata on bending strengths (M artin et al. 1998; N igg an d H erzog
1999).
Bone is w eakest in shearing and ultim ate shear stress does n o t vary as
greatly as b ending or ultim ate tensile stress. Values range from 50.4-73.0 MPa,
for h u m a n com pact bone (Reilly and B urstein 1975; Saha 1977; M artin et al. 1998;
Turner et al. 2001) w ith a m ean of 62.0 MPa, the value used in this s tu d y Like
ultim ate shear stress, ultim ate com pressive stress does no t seem to vary greatly.
U ltim ate com pressive stress ranges aro u n d 200.0 MPa (M artin et al. 1998; N igg
an d H erzog 1999).
The ultim ate tensile stress for tendons u sed in this stu d y w as 100.0 MPa,
an average determ ined from m am m alian tail and digit tendons (Bennett et al.
1986). Tendinous ultim ate tensile stress has been reported to range from 20.6 147.0 M Pa (Viidik 1966). These data com pare w ell w ith uncalcified tendons from
a heron, w hich ranges from 77.0-131.0 M Pa (Bennett and Stafford 1988). A value
of 271.0 M Pa w as reported for an ossified ten d o n from Grus antigone, (Currey
1999a). The ultim ate tensile stress for bone from terrestrial anim als ranges from
108.0 M Pa for Alligator fem ur to 254.0 M Pa for Grus tibiotarsus (C urrey 1990)
w ith an average of 166.0 MPa, w hich w as u sed in this study. This average is a
close fit to the "typical" value of 150.0 M Pa (M artin et al. 1998; N igg and H erzog
78
Table 3. U ltim ate stress values used for this study. The ultim ate stresses for ossified
tendons are identical to bone except for tensile stress. See text to see how these
values are obtained.
Ultimate Stress (MPa)
Tendon
Ossified Tendon
Bone
Tension cru
100
271
166
Compression cju
-
200
Bending Cut
-
138
200
138
Shearing ctm
-
62
62
1999; C urrey 2001).
The m aterial properties above w ere taken from a variety of sources,
each em ploying slightly different m ethods, and specim ens of different species
an d ages. Therefore, the values for these properties have all b een chosen
conservatively. The ultim ate tensile strains w ere calculated (E = a/s) for each
m aterial (tendon, ossified tendon, and bone) an d listed in table 4. These values
sit com fortably w ithin the ultim ate tensile strain range for bone of 0.0045-0.13
(C urrey 1990). The ultim ate strain for ten d o n w as slightly underestim ated, w hich
ranges from 0.08-0.1 (Rigby et al. 1959)> and ossified tendon in th e Grus, w hich
w as em pirically determ ined to be 0.062 (C urrey 1990). Low er strain values w ere
n o t surprising considering the conservative values chosen for E. H ow ever, the
ultim ate tensile strain of tendons, ossified tendons, and bone calculated in table
4 agree w ell w ith values em pirically obtained from turkeys (Bennett and Stafford
1988).
Alligator Results and Partially R estrained Joint Properties
The alligator m odel produced a dorsoventral tail deflection of -6.4 cm
w ith fully constrained joints. As expected, this is far sm aller th a n the -32.0 cm
deflection m easured from the live alligator. But, it allow ed for partially restrained
79
Table 4. C alculated ultim ate tensile strain values. U ltim ate tensile strains are
w ith in the range for the respective m aterials, helping ensure the validity of the
m o dulus of elacticity and ultim ate tensile stress used in this study.
Tendon
Ossified Tendon
Bone
E (GPa)
1.6
6 .6
6 .6
A,,.=™,= (MPa)
100
271
0.063
0.041
166
0.025
^ i i . te n s i l e
joint (endzone) properties to be selected. Endzones w ere placed only in regions
w here tw o m em bers joined by p in connections (i.e. tendon-neural spine
connections and non-sacral centrum -centrum connections). The alligator m odel
w as found to be very sensitive to endzone properties, th o u g h partially restrained
fixed joints h a d little im pact on the results (increased deflection b y 0.1 cm).
E ndzone properties w ith 0.2 cm w idths and 10% m em ber stiffness p ro duced best
fit a deflection of -30.91 cm.
U sing endzones (0.2 cm w idth, 10% stiffness) the alligator m odel
p ro d u ced a m axim um bending stress (about the y-axis) at the first caudal
n eural spine of -13.0 MPa, well w ithin bone's ultim ate ben d in g stress (108 MPa
for alligator fem ur, C urrey 1990). The m odel experienced a m axim um shear
stress (Vy) of 2.2 MPa, far below (50.4-73.0 MPa) ultim ate shear stress com m on
values. The highest axial stress (33.8 MPa) w as generated at the apical ligam ent
connecting the last sacral neural spine to the first caudal neu ral spine. Again, this
value w as w ell below the ultim ate tensile stress of tendon (around 100.0 MPa,
Bennett, Ker et al. 1986). Thus, dorsal epaxial m em bers (just u n d e r the skin) w ere
u n d e r the greatest tensile stress, conform ing to expected results (A lexander and
Jayes 1981).
The alligator m odel results appear valid, insofar th at stresses did not
80
Figure 25. D iagram of the Alligator m odel show ing placem ent of endzones. The
shaded areas betw een the centra and at the apical tendons are the locations of
endzones. These short regions (0.2 cm) have altered stiffness values.
exceed m aterial strength. Furtherm ore, these stresses are w ithin the value range
incurred by the limb bones of Alligator d u rin g locomotion (Blob and Biewener
1999). Therefore, cautious application of m odeling the two dinosaurs is justified.
Results
Tenontosaur M odel Results
Tenontosaur m odel results are sum m arized in table 5. All three
tenontosaur m odel variants experience m ore tail deflection th an dorsal deflection
(Fig. 26). Tail deflection is also reduced m ore by the incorporation of tendons
and their subsequent ossification than in the dorsal region. Therefore, the tail
is m ore heavily influenced in term s of deflection than the tru n k by inclusion of
tendons and their ossification. In both anatom ical regions, the vertebral colum n
is stiffened m ore by tendon inclusion than by ossification. In the tail, however,
an increase in stiffness results from intratendinous ossification w here the dorsal
81
Table 5. Sum m ary of tenontosaur m odel results. These results are u sed to
extrapolate deflection and stress values for loads from 0 to 10 kN at I kN intervals.
The negative values for the deflection (dy) indicate the ventral direction in the
dorsal (D) and caudal (C) regions. The negative values for the b en d in g loads (Fbz)
indicate a counter-clockw ise direction in left lateral view about th e z-axis (about
the central transverse axis) in the n eu ral spine of the first sacral vertebra (SI)
and last sacral vertebra (S5). Fa denotes axial tensile stress in the apical ligam ent
betw een the last dorsal and first sacral (T16-S1) and betw een the last sacral and the
first caudal (S5-C1) n eural spines.
Deflection (cm)
Stress (MPa)
D y(D )
Dy(C)
Load
(kN)
Model Tl: No Tendons
Fbz (SI)
Fbz (S5)
Fa (D16-S1)
Fa (S5-C1)
-669.40
-238.61
-1187.35
543.48
2699.01
607.4
3022.45
1425.47
7082.80
-94.48
-470.32
258.84
1130.94
151.51
753.71
373.54
1852.34
-105.73
-526.27
234.33
995.57
138.33
688.14
326.33
1618.37
-110.68
-2.00
-549.82
-10.00
Model T2: Tendons
:3288.48
-275.46
-17.32
-2.00
-1333.84
-85.97
-10.00
Model T3: Ossified Tendons
-2.00
-10.00
-14.90
-74.04
-235.62
-1140.94
region is h ard ly affected.
The greatest bending stress occurs about the first and last sacral neural
spine arches (Fig. 27). This corresponds to peak bending m om ents about the
neu ral arch, a phenom enon reported to occur in m echanical testing of h u m an
vertebrae (Cyron et al. 1976; C yron and H u tto n 1978). As show n in table 5,
ben d in g stresses about the last sacral n eu ral spine are greater th an bending
stresses in the first sacral neural spine at a given load. Furtherm ore, the
b en d in g stress in the neural spines is reduced m ore b y ten d o n inclusion th an by
ossification.
Axial stress is greatest in the apical ligam ent betw een the posterior (S5-C1)
n eu ral spines (Fig. 28). This situation is identical w ith th at seen in the alligator
82
Tenontosaur Model: Deflection vs. Load
— ■— D. No Tendon
Tendon
—' ■
D. Gss Tendon
— D— C. No Tendon
- - O - - C . Tendon
— O " C. Oss Tendon
1000 '
Load (kN)
Figure 26. Tenontosaur deflection by load. The caudal region experiences m ore
deflection than the dorsal region. Also, note that the spinal colum n is stiffened
m ore by the inclusion of tendons than by ossification.
T e n o n to sau r Model: B ending S tre ss vs. Load
3000.00
2500.00
2000.00
— ■— Sl-No Tendon
• • • * • • S I -Tendon
• -S I-O ssT en d o n
1500.00
— o— S 5-No Tendon
& - 85-Tendon
- O - S S - O s s Tendon
1000.00
500 00
Load (kN)
Figure 27. Tenontosaur bending stress by load. The last neural spine of the sacrum
(S5) experiences m ore b ending than the first sacral neural spine.
83
Tenontosaur Model: Axial Stress vs. Load
8000.00
7000.00
6000.00
—
D16/ S1 No Tendon
• - * - • [ > 16/ 81- Tendon
—
D16/S1-Oss Tendon
— O— S5/C1-No Tendon
-•< y --S 5 /C 1 -T en d o n
- O - S 5/C 1-0ss Tendon
3000.00
2000.00
1000.00
L o a d (kN)
Figure 28. Tenontosaur axial stress by load in the apical ligam ent connecting the
last dorsal (D16) and first sacral (SI), and the last sacral (S5) and first caudal (Cl).
The S5-C1 ligam ent experiences m ore tension than the D16-S1 ligam ent. A ddition
of tendons affects axial stress m ore than ossification.
m odel and is consistent w ith lum bar loading in m am m als (A lexander and Jayes
1981). Inclusion of tendons significantly reduces axial stress in the D16-S1 and S5C l ligam ents. In contrast, intratendinous ossification has little im pact axial stress
reduction. W hile intratendinous ossification low ers the axial stress in the D16-S1
and S5-C1 ligam ents, it also increases the axial stress in the ossified tendons over
the posterior sacrum and puts the ossified tendons in the ventral aspect of the tail
into com pression, though at very small m agnitudes.
Shearing is small in the alligator m odel and in both d in o sau r m odels as
well. H ow ever, despite a small m agnitude, intratendinous ossification increases
the shearing stress experienced in the tenontosaur model. W hen the ossified
tendon architecture is considered (parallel to the vertebral axis), such an increase
84
Table 6. U ltim ate loads determ ined from m axim um stresses in m odel variants
w ith no tendons (NT), tendons (T), and ossified tendons (OT). These loads w ere
calculated using the ultim ate stresses (y-value) to calculate the load (x-value) from
the linear plots above. In the dorsal region, m axim um tension is in the apical
ligam ent betw een the last dorsal and first sacral and in the caudal region betw een
the last sacral and first caudal. In the dorsal region m axim um b en d in g occurs in the
first sacral neu ral spine and in the caudal region in the last sacral n eu ral spine.
Tenontosaur Model
Brachylophosaur Model
Dorsal loads (kN)
Caudal loads
Dorsal loads (kN)
Caudal loads
_____________________________________ (kN)_________________________________ (kN)
NT
T
OT
NT
T
OT
NT
T
OT
NT
T
OT
Tension
0.32
1.3
1.4
0.12
0.52
0.60 '
1.3
6.1
20.7
1.7
7.1
24.7
Bending
1.2
2.9
2.6
0.50
0.90
1.0
4.4
8.5
16.5
4.1
14.9
44.3
is expected as loads are distributed transversely to the n eu ral spines. D espite this
effect, shearing w as so far below other stresses th at it is no t considered further.
The tenontosaur m odel fails at the S5-C1 apical ligam ent in tension at 0.12
RN. A ddition of the tendons increases failure load to 0.52 kN (a 320% increase).
Intratendinous ossification allows the structure to w ith stan d an additional 0.08
kN . Tendon inclusion decreases tail deflection (at failure load) from 55.2 cm
w ith no tendons, to 27.3 cm w ith tendons (50.6% increase in stiffness), to 23.3 cm
w ith ossified tendons (57.8% increase in stiffness). D orsal deflection at failure
load is reduced from 7.7 cm w ith no tendons to 1.22 cm w ith ten d o n s (84.1%
increase in stiffness) to 1.0 cm w ith ossified tendons (86.6% increase in stiffness).
A lthough the dorsal region is stiffened m ore th an the tail on a percent basis, the
tail is influenced m ore in absolute term s. In the tenontosaur m o d el w ith ossified
tendons, the m axim um tension shifts from th e S5-C1 apical ligam ent to the apical
ossified tendon over the posterior sacrum . H ow ever, w h en this apical ligam ent
fails the tensile stress in the ossified ten d o n is 210.0 MPa, below th e m axim um
85
tensile stress of ossified tendon (271.0 MPa). Therefore, the w eakest m em ber
rem ains the S5-C1 apical ligam ent, despite inclusion of tendons.
B rachylophosaur M odel Results
The results for the brachylophosaur m odels are sum m arized in table 7.
All three brachylophosaur m odel variants experience m ore tail deflection than
dorsal deflection (Fig. 29). Tail deflection is also reduced m ore b y incorporating
tendons and subsequent ossification th an in the dorsal region. Therefore, like the
tenontosaur m odel, the tail is m ore influenced in term s of deflection b y tendon
inclusion th a n ossification. The brachylophosaur m odel is also stiffened m ore
by the adding tendons th a n through intratendinous ossification, regardless of
anatom ical region.
The greatest bending stress occurs at the first and last sacral n eu ral spine
arches. This result agrees w ith the tenontosaur and alligator m odels. As show n
in figure 30, bending stresses about the last sacral neural spine are roughly the
sam e (though slightly greater) th an the b en d in g stresses at the first n eu ral spine
for any given load. Tendon inclusion significantly low ers the b en d in g m om ents
at the n eural spines arches. Intratendinous ossification low ers these bending
stresses even m ore, th o u g h not as m uch as the addition of tendons. This effect is
caused by the dorsoventral neural spine height.
A m axim um bending m om ent of -5.9 kN*m is p ro duced at the last
sacral neural spine at failure load. This is m u ch low er th an the -25.0 kN*m
b e n d in g m om ent across the hips calculated in Iguanodon usin g beam theory
(A lexander 1985). H ow ever, A lexander notes th a t this value is probably inflated
appreciably because his m odel assum ed a p o in t support, w here as these m odels
are su p p o rted at every sacral centrum . Therefore, this study's results do no t
86
Table 7. Sum m ary of brachylophosaur m odel results. These results are u sed to
extrapolate deflection and stress values for loads from 0 to 10 kN at I kN intervals.
The negative values for the deflection (dy) indicate the ventral direction in the
dorsal (D) and caudal (C) regions. N egative values for the b en d in g loads (Fbz)
indicate a counter-clockw ise direction in left lateral view about th e z-axis (about
the central transverse axis) in the neural spine of the first sacral vertebra (SI) and
last sacral vertebra (S9). Fa denotes axial tensile stress in the interspinous ligam ent
betw een the last dorsal and first sacral (D18-S1) and betw een the last sacral and the
first caudal (S9-C1) neural spines.
Stress (MPa)
Deflection (cm)
Load
Dy (D)
Dy (C)
Fbz ( SI)
Fbz (S9)
Fa (DlS-Sl)I
Fa (S9-C1)
-52.53
-220.74
-65.93
70.49
326.10
143.49
658.06
117.04
541.43
30.38
139.66
8.11
405.0
(kN)
Model BI: No Tendons
-35:04
-2.00
-151.50
-10.00
Model B2: Tendons
-11:92
-8.33
-2.00
-50.56
-30.15
-10.00
Model B3: Ossified Tendons
-2.00
-10.00
-4.20
-11.96
-3.60
-15.15
- 302.36
-35.62
-162.94
20.40
83.80
35.26
. 161.23
-18.57
-84.93
6.93
31.82
10.73
49.02
contradict previous attem pts to u n d erstan d b en d in g stress across the hips in
H adrosauriform es.
Axial stress is greatest in the apical ligam ent betw een the anterior (dl8-Sl)
neural spines (Fig. 31). This situation differs from the alligator and tenontosaur
m odels, w hich experience the greatest degree of tension in the apical ligam ents
betw een the last sacral and first caudal n eu ral spines. Inclusion of tendons
substantially reduces axial stress in both anterior (dl8-Sl) and posterior (S9-C1)
apical ligam ents.
The brachylophosaur m odel fails first in tension (at 1.3 kN ) in the D18S l apical ligam ent (Table 6). Tendon inclusion increases failure load to 6.1 kN
(a 361.6% increase). Intratendinous ossification increases failure load to 20.7(a
87
Brachylophosaur Model: Deflection vs. Load
— ■— D. No Tendon
Tendon
— *■ " D. Oss Tendon
— O— C. No Tendon
- - O - - C . Tendon
- O - - C. Qss Tendon
L oad (kN)
Figure 29. B rachylophosaur deflection by load. The degree of deflection is largely
d ep en d an t u p o n inclusion of tendons not by ossification.The caudal region is
deflected m ore than the dorsal region, though overlap exists.
B ra c hy lo ph o s au r Model: Be nd in g S tre ss vs. Load
— »— SI-No Tendon
si-Tendon
-
-S I-O ssT en d o n
— o— S 9-No Tendon
•••*••• S 9-Tendon
- O -S G -O ss Tendon
... i---
L o a d (RN)
Figure 30. B rachylophosaur bending stress by load. N ote that the bending stress,
like deflection, is largely dependant on the presence of tendons no t ossification.
The first and last neural spines of the sacrum (SI and S9) experience roughly the
sam e degree of bending.
88
B rachylophosair Model: A»al Stress vs. Load
7 0 000
50000
— ■— D18/S1 No Tendon
• • * • • 0 1 8 / 5 1 - Tendon
- ■ * - • D18/S1-Oss Tendon
— D— S O /d-N o Tendon
S9/C1-Tendon
- O - SO/C 1-O ss Tendon
S 40000
I
30000
2 0 000
0
1
2
3
4
5
6
7
8
0
1
0
Load(KN)
Figure 31. B rachylophosaur axial stress by load in the interspinous ligam ent
connecting the D18-S1 and S9-C1 ligam ents. The apical ligam ent betw een D18-S1
experiences m ore tension than the one betw een S9-C1.
1459.9% increase). The tendons decrease tail deflection (at failure load) from 38.1
cm w ith no tendons, to 8.6 cm w ith tendons a (77.4 increase in stiffness), to 2.6 cm
w ith ossified tendons (93.1% increase in stiffness). Dorsal deflection at this load is
reduced from 25.1 cm w ith no tendons to 6.5 cm w ith tendons (74.2% increase in
stiffness), to 3.5 cm w ith ossified tendons (85.9% increase in stiffness).
Deflection in the m odels is m ainly affected by length of the loaded region
and neural spine height. The brachylophosaur m odel is constructed from a
slightly truncated tail, w hich if complete, w ould be roughly 50 cm longer. To
evaluate this influence, the three brachylophosaur m odel variants w ere m odified
to include the last p art of the tail. This increased tail deflection by 29.5% at 2.0 kN
and 18.8% at 10.0 kN, inflating the difference betw een dorsal and tail deflection.
At failure (1.3 kN) in the m odel variant w ith no tendons, deflection increased
89
from 38.1 cm to 49.94 cm (a 31.1% increase). In the ossified ten d o n m odel variant
w ith a com plete tail, deflection increased from 2.6 cm to 6.6 cm at failure load.
The m odel w ith the com plete tail p ro d u ced a difference of 2.8 cm betw een
dorsal and caudal deflection, w hile the difference in the ten o n to sau r m odel is
47.5 cm at failure load. Therefore, the dorsal and tail regions are roughly equally
affected by ten d o n biom echanics. Furtherm ore, the m odel variants w ith the fully
reconstructed tail experience sm all increases in stress (only a 2.0% increase in
ben d in g an d less th an 0.01% increase in axial stress). C onsidering these data,
the original m odel w ith a truncated tail w ill suffice for stru ctu ral analysis and
hypothesis testing.
D iscussion
It has been noted that, despite the prom ise of using biom echanics to infer
function in extinct organism s, m any obstacles rem ain (Lauder 1981; Bryant and
Seym our 1990; L auder 1991a; Lauder 1991b; L auder et al. 1993; L auder 1995;
L auder an d Reilly 1996). For exam ple factors such as posture, m ass, m uscular
force production, neurological factors (i.e. tim ing in m uscle activation), and so
on m u st be inferred. Each of these factors is an inferential step in w hich errors
can accrue. There is evidence th at organism s can evolve novel function directly
from m orphology w hile neurological activation patterns rem ain conserved
across the group of interest (Lauder 1991a). C om puter m odels th a t infer m uscle
m ass and force production have produced interesting results (H utchinson and
Garcia 2002), th o u g h the m ethods used here m ore closely resem ble attem pts to
u n d e rstan d the skeleton m echanically (Rayfield et al. 2001).
Stress values produced in the m odels of this stu d y w o u ld m ost likely
90
be low er in living anim als as various connective tissues and m uscular loads
w ould help absorb them . Therefore, the results probably overestim ate stresses
experienced b y the m em bers for any given load. D espite this lim itation, the
com puter m odels can be u sed to answ er specific structural questions. For
exam ple, incorporation of ossified tendons into the m odels w ere expected to
place m em bers (skeletal elem ents) into com pression th at w ere initially in tension,
because bone is generally stronger in com pression rath er th a n tension (Nigg and
H erzog 1999). C om pression im pacts the brachylophosaur m odel b y affecting
ben d in g of the neural spines. The angle of inclination in the M. transversospinalis
trellis reduces bending on the neural spine by com pressive (-y) load application.
This effect is inherent in the trellis architecture, b u t m ore p ro n o u n ced in the
brachylophosaur m odel because of the taller n eu ral spines.
The tenontosaur m odel contains hypaxial, in addition to epaxial tendons,
th a t experience com pression (a m axim um of -59.38 MPa in the 2.0 kN load case)
far below tensile m agnitudes. The structure's stresses are m ostly distributed
in the epaxial ossified tendons. Indeed, w h en hypaxial tendons are rem oved,
deflection, tensile and bending stress increase less th an one percent. Therefore,
hypaxial ossified tendons in prim itive ornithopods likely h a d little structural
function in term s of dorsoventral loading.
Some have suggested th at the trellis of ossified tendons stiffened the tail
laterally (N orm an 1986). This conjecture w as tested by lateral lo ad cases w here
loads w ere applied at the tail's center of m ass in the m ediolateral direction.
There w as no difference betw een m odel variants in either anim al. H ow ever, the
m odeled tendons are applied im m ediately to the neural spines in these models.
Lateral location from the neutral axis of the vertebral colum n w o u ld increase the
tendon's ability to stiffen the vertebral colum n. But proxim al application seems
91
appropriate given archosaur epaxial anatom y (see chapter I).
It is also possible th at epaxial anatom y in hadrosaurs predisposes the
tail to structural failure. Indeed pathologies on the neural spines in the tail of
h ad ro sau rs is n o t uncom m on (Rothschild and Tanke 1992). These pathologies
consist of broken, fused, and eroded n eural spines and ossified tendons. N othing
in the brachylophosaur m odels suggest th a t these pathologies w ere caused by
tail biom echanics due to dorsoventral loading. That is, the region in the tail in
w hich these pathologies occur is not a m echanical w eak point. Therefore, it is
likely th a t som e stress other than loading from b o d y m ass caused failure, such as
m ating, agonistic behavior, or predation.
E valuating the Fem oral and H ogging H ypotheses
The dinosaur m odel results m ay also be u sed to evaluate the
fem oral loading and hogging hypotheses. Both hypotheses w ere m ad e for
H adrosauriform es, b u t can also be used against the Tenontosaurus data. The
fem oral loading hypothesis is supported if the m odels reduce tail deflection and
stress m ore th an dorsal deflection and stress. This is true for the tenontosaur
m odel, w here the tail experiences a greater percentage increase in stiffness (7.2%
com pared to 2.5% in the dorsal region), and deflects m ore in absolute term s
(23.3 cm com pared to 1.0 cm in the ossified ten d o n variant at failure load) due
to its longer length. Therefore, the fem oral loading hypothesis is su p p o rted
b y the tenontosaur m odel. The brachylophosaur m odel experiences reduced
deflection nearly equally in the dorsal and tail regions on a percentage basis w ith
the inclusion of the tendons and their ossification (Table 8). In absolute term s,
the tail deflects 2.6 cm com pared w ith 3.5 cm in the ossified ten d o n variant at
failure load. A lthough a com plete tail w o u ld increase caudal deflection past
92
Table 8. Percentage changes in the m odels w h en tendons are ossified. The deflection
ro w percentages indicate to h ow m uch less the m odel deflected com pared to the
m odel w ith tendons. Bending stress percentages indicate h o w m u ch less the first
sacral n eural spine (dorsal column) and last sacral n eural spine (caudal colum n)
w ere loaded in bending com pared to the m odel w ith a netw ork of tendons. Axial
stress percentages indicate h ow m uch less th e interspinous ligam ents betw een the
last dorsal and first sacral neural spines (dorsal column) and the last sacral and
first caudal neural spines (caudal column) w ere loaded in tension, com pared to the
m odel w ith a netw ork of tendons.
% Changes
Deflection
Bending a
Axial a
Tenontosaur Model
Brachylophosaur Model
Dorsal
-2.5
4.7
-2.1
Dorsal
-11.7
-25.6
-17.1
Caudal
-7.2
-5.0
-3.4
Caudal
-15.7
-18.8
-18.3 .
dorsal deflection, they w ould still be subequal. A lthough the d ata do no t favor
the fem oral loading hypothesis, it can n o t be rejected because the tail is still being
stiffened b y intratendinous ossification.
Equal stiffness in the dorsal and tail regions supports the hogging
hypothesis. This is not the case for the tenontosaur m odel, w hich experiences a
difference betw een dorsal and tail deflection of 47,5 cm (in the ossified tendon
m odel at failure load). Thus, the hogging hypothesis can be rejected for the
prim itive ornithopod. Differences betw een dorsal and tail deflection in the
brachylophosaur m odel (ossified ten d o n variant) are only 9.2 cm an d only 3.1
cm in the m odel w ith a com plete tail. O strom (1964) noted th a t trellis ossification
w o u ld provide additional structural su p p o rt to resist dorsal tension and ventral
com pression of the vertebral colum n due to hogging about the hips. A ccording
to the m odels, this is the case as the trellis is loaded in tension, n o t com pression.
Furtherm ore, the hogging hypothesis is a structural claim th a t posits a reduction
in b en ding stress about the hips. The ossified trellis functions this w ay by
93
reducing b en ding stress approxim ately 25.6% in the dorsal region and 18.8%
in the caudal region. It also reduces axial stress b y over 17% in b o th anatom ical
regions. In contrast, there is a 2.1% (dorsally) and 3.4% (caudally) difference in
the tenontosaur m odel's ability to reduce axial stress. Since th e m odels fail in
tension, the equal degree of axial stress reduction in the brachylophosaur m odel
supports the hogging hypothesis. The ossified trellis' ability to low er internal
stresses is largely due to the architecture of the structure enhanced b y tall neural
spines. M oreover, the brachylophosaur m odel fails first in tension in the anterior
apical ligam ent (D18-S1), w hich suggests th a t dorsal loading is ah im portant
factor in this animal. The degree to w hich tensile stress is red u ced is the degree to
w hich epaxial m uscles are relieved of load bearing functions, w hich is substantial
in the b rachylophosaur m odel. Thus, results favor the hogging hypotheses for
the brachylophosaur m odel.
A com parison can be m ade betw een th e tw o taxa m odeled in term s of
ossifying different epaxial muscles. The tenontosaur m odel can be tak en as
the prim itive condition for O rnithopoda. Based on the m odel's results, the
plesiom orphic ossification pattern m ay have been an ad aptation resulting in tail
stiffness. As Dollo n o ted in 1886, this w ould have locom otor im plications because
of the caudofem oral m uscles couple the h in d limb and tail into a functional u n it
(Gatesy 1995). If tail stiffness increased the m om ent arm for fem oral retraction,
ten d o n ossification could be selectively advantageous. Ossified ten d o n s w ould
reduce required loads from the epaxial m usculature to counteract fem oral
m uscles and passively stiffen the tail and increase safety factors. Therefore,
in tratendinous ossification m ay have selected for because of locom otor function.
O ssification of the M. transversospinalis tendons (the trellis) in
H adrosauriform es increases tail stiffness b y 15.7% (where the tenontosaur
94
m odel reduced deflection by only 7.2%). Like Tenontosaurus, this likely im pacted
caudofem oral loading of the tail d u rin g fem oral retraction. H ow ever, the dorsal
region is also im pacted by the ossified trellis to roughly the sam e degree. This
effect of the ossified trellis m ay have been im portant in H adrosauriform es,
w hich w ere large anim als, in term s of passively su p p o rtin g th e b o d y from
hogging about the hips. M oreover, the ossified trellis greatly enhances the
vertebral colum n's ability to bear large loads (237.9% greater th a n unossified
tendons com pared to 15.1% in the tenontosaur model). Or, stated another way,
the trellis greatly reduces bending and axial stresses suffered b y th e vertebral
colum n. It is interesting to note th at b o th the tenontosaur an d brachylophosaur
m odels fail from tensile loading and th at ossified tendons reduce tensile stress
on the vertebral colum n m ore than ben d in g stress. In addition, they also have
higher ultim ate tensile stress values th an ten d o n or bone and are th u s capable
of w ith standing larger tensile stresses. This increases the vertebral column's
ability to w ith stan d loading in tw o ways: b y low ering internal stresses and by
increasing the m axim um tolerable stress. Furtherm ore, the brachylophosaur
m odel w ith ossified tendons reduces tensile stress by over tw ice th a t of the
tenontosaur m odel w ith ossified tendons. Again, this function m ay be related to
the evolution of large size in these animals.
Conclusions
Ossified ten d o n biom echanics in ornithopods w ere exam ined using
com puter finite elem ent m odeling. W hile m any claims about ossified tendon
function have been m ade, none have been rigorously tested. This stu d y is a first
attem pt to do so and has produced reasonable results, considering the outcom e
95
of a m odeled alligator tail and the location of m axim um stresses in the dinosaur
m odels. A m odel produced from the anatom ical data of a Tenontosaurus tilletti
w as u sed to represent the prim itive ornithopod condition th a t possessed parallel
ossified tendons from the M. Iongissim us dorsi. A nother m odel constructed from
a Brachylophosaurus canadensis w as u sed to represent the derived condition found
in H adrosauriform es th at possessed a trellis of ossified tendons from the M.
transversospinalis.
The tenontosaur m odel w ith ossified tendons reduced the stresses
incurred on vertebral elem ents, though to a lesser degree th an the
brachylophosaur m odel, and actually placed the neural spines u n d e r greater
shear (though the m agnitude w as insignificantly small). The apom orphic ossified
ten d o n trellis found in H adrosauriform es stiffens the vertebral colum n m ostly
th ro u g h ten d o n architecture and neural spine height. M oreover, the ossified
trellis reduced the stresses incurred on the vertebral colum n m ore th an the
ossified tendons in the tenontosaur m odel, th u s allow ing far greater loads to be
safely born. Since bo th m odels w ere stiffened by the addition of ossified tendons,
they probably functioned by passively su p p o rtin g the tail an d trunk, thus
reducing the need for epaxial m uscular exertion.
Two hypotheses w ere evaluated concerning the function of ossified
tendons. The hogging hypothesis suggests th a t the vertebral colum n is stiffened
in b o th the dorsal and tail regions to resist buckling about the h ip s due to body
w eight. The fem oral loading hypothesis suggests th a t ossified ten d o n s stiffened
the tail to increase the lever arm for fem oral retraction. The h o g g in g hypothesis
w as rejected and the fem oral loading hypothesis su p p o rted for the tenontosaur
m odel. N either hypothesis could be rejected for the brachylophosaur model.
Also, the effectiveness of ossified tendons to reduce stress an d deflection is 2
96
to 3 tim es th a t of the plesiom orphic ten d o n architecture O ssification of the M.
transversospinalis trellis in H adrosauriform es m ay have therefore functioned in
tw o ways: in locom otion and in supp o rtin g large body size.
97
APPENDIX
M ORPHOLOGIC DATA FOR CHAPTER 4 MODELS
98
Table 9. M orphologic data from M OR 794 (Brachylophosaurus canadensis). Length
is in the anterior-posterior direction, height is in the dorsoventral direction, and
w id th is in the m ediolateral direction. H eight for the n eu ral spine w as m easured
along its longitudinal axis (not vertical). These m easurem ents for the neural spine
w ere taken at their dorsal sum m its. 0? w as m easured at the base of the neural arch
w ith a protractor and is the angle betw een the vertical axis an d th e n eural spine.
Dorsal Section
Centra
Vert.
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Neural spines
L (cm)
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
H (cm)
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
16.0
W
(cm)
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
L (cm)
1.0
1.5
1.5
2.0
3.0
7.0
7.0
7.0
8.0
8.0
8.5
9.0
9.5
9.5
9.5
9.5
9.5
9.5
H (cm)
2.0
3.0
7.0
13.0
13.0
17.0
20.0
22.0
23.0
26.0
28.0
28.0
28.0
28.0
31.0
33.0
33.0
33.0
W
(cm)
0.3
0.3
0.5
0.7
0.8
0.8
1.0
1.0
1.0
1.0
1.0
1.1
' 1.1
1.1
1.1
1.1
1.1
1.1
0
60.0
60.0
50.0
40.0
30.0
20.0
15.0
10.0
5.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
W
(cm)
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
10.0
Sacral Section
Centra
Vert.
I
2
3
4
5
6
7
8
Neural spines
L (cm)
9.0
9.0
9.0
9.0
9.0
9.0
9.0
9.0
W
H (cm) (cm)
15.0
11.0
15.0
11.0
15.0
11.0
15.0
11.0
15.0
11.0
15.0
11.0
15.0
11.0
15.0
11.0
L (cm)
9.5
9.5
9.5
9.0
9.0
9.0
8.0
8.0
H (cm)
35.0
35.0
35.0
37.0
37.0
37.0
37.0
37.0
0
99
9
9.0
15.0
11.0
8.0
40.0
1.1
20.0
W
(cm)
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.9
0.9
0.9
0.9
0.9
0.9
0:8
0.8
0.8
0.8
0.8
0.7
0.7
0.7
0.7
0.7
0.7
0.7
0.6
0.6
0.6
0.6
0.5
0.5
0
25.0
27.0
30.0
31.0
33.0
33.0
35.0
38.0
40.0
43.0
50.0
50.0
50.0
50.0
53.0
53.0
57.0
62.0
62.0
62.0
62.0
63.0
63.0
63.0
65.0
65.0
65.0
66.0
67.0
69.0
69.0
Caudal Section
Centra
Vert.
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Neural spines
W
L (cm) H (cm) (cm)
7.9
13.5
10.7 .
8.0
13.0
10.5
7.9
13.0
10.3
7.9
13.8
10.2
13.7
7.8
10.0
7.9
13.5
9.8
8.7
13.4
9.6
8.5
13.0
9.5
8.1
12.5
9.3
8.5
12.5
9.1
8.2
12.0
8.9
7.9
11.5
8.8
7.7
11.0
8.6
10.5
8.4
7.8
7.7
10.1
8.2
10.0
8.1
8.0
8.0 '
9.7
7.9
9.5
7.7
7.8
8.5
.9.4
7.5
8.2
9.4
7.4
7.2
8.0
8.6
8.0
8.2
7.0
7.0
7.8
8.2
8.0
7.8
6.9
7.9
6.8
7.6
7.7
6.6
7.6
7.5
6.5
7.6
6.4
7.4
7.3
7.2
6.2
7.4
7.2
7.3
6.1
6.0
7.3
7.0
L (cm) H (cm)
7.1 ■ 43.5
7.1
43.5
6.8
41.0
6.7
40.5
6.5
38.0
6.5
37.5
6.2
35.0
6
33.0
5.7
32.0
5
29.5
5.7
29.0
26.0
3.8
3.2
21.0
3.1
26.0
5.1
26.0
5
28.0
27.5
3 .8
3.7
26.0
3.4
26.0
3.4
25.0
3.1
25.0
23.0
3.8
23.0
3.8
3.7
23.0
3.7
23.0
3.5
22.0
3.4
21.0
3.3
21.0
3.3
20.0
3.2
20.0
3.1
19.0
100
Table 10. M orphologic data from M OR 682 (Tenontosaurus tilletti). Length is in the
anterior-posterior direction, height is in the dorsoventral direction, and w id th is
in the m ediolateral direction. H eight for the neural spine w as m easured along
its longitudinal axis (not vertical). These m easurem ents for the neural spine
w ere taken at their dorsal sum m its. N eural spine 0 ?was m easu red at the base of
the n eu ral arch w ith a protractor and is the angle betw een th e vertical axis and
the n eural spine. C hevron 0 ? w as m easured at the base of the chevron w ith a
protractor an d is the angle betw een the vertical axis and the chevron spine.
Dorsal Section
Centra
Neural spines
Vert.
L
H
W
L
H
W
0
I
3.0
3.0
2.7
2.0
3.7
0.5
40.0
2
3.0
3.2
2.9
2.1
3.9
0.5
30.0
3
3.0
3.4
3.0
2.3
5.0
0.6
20.0
4
3 .2
3.4
3.4
3.0
5.1
0.6
17.0
5
3 .4
3.4
3.8
3.6
5.2
0.6
15.0
6
3.7
3.4
3.9
3.9
5.2
0.6
6.0
7
5.0
3.4
3.9
3.3
5.2
0.6
0.0
8
5.0
3.5
3.1
3.4
6.0
0.7
5.0
9
5.0
3.7
3.3
3.5
6.7
0.8
10.0
10
5.0
3.9
3.4
3.5
6.7
0.9
10.0
11
5.0
5.0
3.5
3.5
6.7
1.1
10.0
12
5.0
5.0
5.0
3.5
6.7
1.3
10.0
13
5.0
5.0
5.5
3.2
7.5
1.2
10.0
14
5.0
5.0
5.2
3.0
8.8
1.2
10.0
15
5.0
5.0
5.2
3.0
8.8
1.2
10.0
16
5.0
5.0
5.2
3.0
8.8
1.2
10.0
Sacral Section
Neural spines
Centra
Vert.
L
H
W
L
H
W
0
I
3.7
5.0
5.0
3.5
8.0
0.7
5.0
2
5.0
5.0
5.0
3.5
9.0
0.7
5.0
3
5.0
5.0
5.0
3.5
8.0
0.7
5.0
4
5.0
5.0
5.0
3.5
8.0
0.7
5.0
5
3.5
5.0
5.0
3.5
8.0
0.7
5.0
Caudal Section
101
37
38
39
40
L
H
3.7
5.2
3.6 ■ 5.3
3.5
5.4
3.5
5.5
3.5
5.5
3.6
3.7
3.7
3.5
5.0 . 3.4
5.4
3.3
3.2
5.7
6.0
3.1
3.2
6.0
3.2
6.0
6.2
3.1
6.5
3.0
6.5
3.0
6.5
3.0
6.5
3.0
6.5
3.0
6.2
3.5
6.0
3.0
5.6
3.0
5.3
3.0 '
2.7
5.1
2.5
5.0
2.5
3.9
3.7
2.5
2.4
3.7
3.7
2.3
2.2
3.5
2.1
3.4
1.9
3.3
1.7
3.1
3.0
1.6
1.5
3.0
1.4
3.0
1.4
3.0
3.5
1.4
1.4
3.0
1.3
3.0
Neural spines
L
H
W
3.0
12.0
3.8
3.0
12.0
3.7
3.0
12.0
3.5
3.0
1.5
3.2
3.0
13.0
2.7
3.2
12.5
3.2
2.5
12.0
3.1
2.2
10.0
3.0
2.0
9.0
3.0
2.0
8.0
7.0
3.0
2.0
3.0
2.0
7.0
3.0
2.0
7.0
3.0
1.8
6.5
3.0
1.5
6.5
3.0
1.5
6.0
3.0
1.5
6.0
1.2
5.5
3.0
3.0
1.0
5.0
3.5
1.0
3.6
3.2
1.0
3.5
3.2
0.8
2.8
3.2
2.5
0.8
0.7
2.5
3.0
2.5
0.7
3.5
2.5 ■ 0.6
3.4
0.6
3.1
2.5
2.5
0.5
3.0
2.5
0.5
2.5
2.5
0.4
2.0
2.5
0.4
1.0
0.3
2.3
1.0
2.2
0.3
1.0
0.2
0.7
2.1
2.0
1.9
1.8
1.7
1.5
1.5
W
5.0
0.8
1.0
1.2
1.0
0.9
0.8
0.7
0.7
0.7
0.7
0.7
0.7
0.7
0.6
0.6
0.6
0.6
0.6
0.5
0.5
0.5
0.5
0.5
0.5
0.4
0.4'
0.4
0.4
0.3
0.3
0.3
0.3
0.2
0.2
-
0
0.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
10.0
15.0
17.0
20.0
22.0
25.0
27.0
30.0
33.0
35.0
35.0
37.0
37.0
40.0
40.0
40.0
40.0
40.01
40.0
40.0
40.0
40.0
40.0
-
Chevrons
L
H
1.5
18.0
1.6
17.5
17.0
2.1
16.0
2.5
15.5
2.6
16.0
2.8
17.5
2.6
17.0
2.5
16:0
2.2
15.5
2.0
15.0
1.7
13.5
1.5
13.0
1.3
13.5
1.1
13.0
1.0
12.0
0.9
11.0
0.8
10.0
0.7
9:0
0.6
8.0
0.6
7.0
0.5
6.0
0.5
5.0
0.5
3.0
0.5
3.0
0.5
3.0
0.4
2.0
0.4
2.0
0.4
2.0
0.4
1.7
0.3
1.0
0.3
1.0
0.3
1.0
0.2
0.7
•-
-
-
-
CO
Centra
Vert.
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28 ,
29
30
31
32
33
34
35
36
-
W
0.7
0.8 0.8
0.8
0.7
0.7
0.7
0.6
0.6
0.5
0.5
0.5
0.5
0.5
0.5
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.3
0.3
0.3
0.3
0.3
0.3
0.3
0.2
0.2
0.2
0.2
-
0
30.0
22.0
15.0
15.0
15.0
22.0
30.0
25.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
40.0
-
-
-
0 .8
102
41
42
43
44
45
46
47
48
49
50
51
52
53
54
3.0
2.9
2.8
2.6
2.5
2.5
2.5
2.5
2.3
2.4
2.4
2.4
2.4
2.3
1.2
1.1
1.0
1.0
1.0
0.9
0.9
0.7
0.5
0.5
0.5
0.4
0.4
0.4
1.5
1.5
1.5
1.5
1.5
1.4
1.4
1.3
1.3
1.2
1.2
1.0
0.8
0.9
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
‘
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
103
REFERENCES CITED
A bdalla, O. (1979). "Ossification and m ineralization in the tendons of the chicken
(Gallus domesticus)." Tournal of A natom y 129: 351-359.
A lexander, R. M. (1985). "M echanics of posture and gait of som e large
dinosaurs." Zoological Tournal of the Linnean Society 83:1-25.
A lexander, R. M. (1988). Elastic M echanism s in A nim al M ovem ent. N ew York,
C am bridge U niversity Press.
A lexander, R. M. and A. S. Jayes (1981). "Estim ates of the ben d in g m om ents
exerted by the lum bar and abdom inal m uscles of som e m am m als."
Tournal of Zoology L ondon 194: 291-303.
Baumel, J. J., A. S. King, J. E. Breazile, H. E. Evans and J. C. V anden Berge (1993).
H andbook of Avian Anatom y: N om ina A natom ica A vium . Cam bridge,
M assachusetts, Publications of the N u tall O rnithological Club.
Baum el, J. J., J. A. W ilson and D. R. Bergren (1990). "The ventilatory m ovem ents
of the avian pelvis and tail: function of the m uscles of the tail region of the
pigeon (Columbia Uvia)." Tournal of Experim ental Biology 151: 263-277.
Benjamin, M., T. K um ai, S. Milz, B. M. Boszczyk, A. A. Boszczyk and J. R.
Ralphs (2002). "The skeletal attachm ent of tendons - te n d o n 'entheses'."
C om parative Biochemistry and Physiology, Part A 133: 931-945.
Benjamin, M. an d J. R. Ralphs (1998). "Fibrocartilage in tendons an d ligam ents an adaptation to com pressive load." Tournal of A natom y 193: 481-494.
Bennett, M. B., R. F. Ker, N. J. D im ery an d R. M. A lexander (1986). "M echanical
properties of various m am m alian tendons." Toumal of Zoology. London
209 (A): 537-548.
Bennett, M. B. and J. A. Stafford (1988). "Tensile properties of calcified and
uncalcified avian tendons." Tournal of Zoology. L ondon 214: 343-351.
Blob, R. W. and A. A. Biewener (1999). "In vivo locom otor strain in the hindlim b
104
bones of Alligator mississippiensis and Iguana iguana: im plications for the
evolution of lim b bone safety factor and non-spraw ling lim b posture."
Tournal of Experim ental Biology 202:1023-1046.
Boas, J. E. V. (1929). "Biologisch-anatomische Studien uber den H als d er Vogel."
D. KgL D anske Vidensk. Selk. Skr. N aturvidensk. M athem . Afd. Series
9(1): 105-222.
Broili, F. (1922). "U ber d en feineren Bau d er ,,verknochnerten Sehnen"
(verknochteren M uskeln) von Trachodon." A natom ischer A nzeiger 55:
465-475.
Brown, B. (1916). " Corythosaurus casuarius: skeleton, m usculature an d epiderm is."
Bulletin of the A m erican M useum of N atu ral H istory 35: 709-716.
Brown, B. (1933). "A gigantic ceratopsian dinosaur, Triceratops maximus, new
species." A m erican M useum N ovitates 649:1-9.
Bryant, H. N. and A. P. Russell (1992). "The role of phylogenetic analysis in
the inference of unpreserved attributes of extinct taxa." Philosophical
Transactions of the Roval Society of London, Series B 337: 405-418.
Bryant, H. N. and K. L. Seym our (1990). "O bservations and com m ents on the
reliability of muscle reconstruction in fossil vertebrates." Tournal of
M orphology 206:109-117.
Bultynck, P. (1992). "An assessm ent of posture and gait in Iguanodon
bernissartensis." Bulletin De U Institut Royal Des Sciences N aturelles De
Belgique 63: 5-11.
C arrano, M. T. and A. A. Biewener (1999). "Experim ental alteration of limb
posture in the chicken (Callus gallus) and its bearing on the use of birds as
analogs for dinosaur locom otion." Tournal of M orphology 240(3): 237-249.
C arrano, M. T. and J. R. H utchinson (2002). "Pelvic and hindlim b m usculature of
Tyrannosaurus rex (Dinosauria: Theropoda)." Tournal of M orphology 253:
207-228.
C astanet, ]., H. Fancillon-Veillot, F. J. M eunier and A. J. de Ricqles (1993). Bone
and Individual Aging. Bone. B. K. Hall. Boca Raton, Florida, CRC. 7: 329-
105
339.
Colbert, E. H. (1951). "E nvironm ent and adaptations of certain dinosaurs."
Biology Review 26: 265-284.
Coom bs, W. R J. (1995). "A nkylosaurian tail clubs of m iddle C am panian to early
M aastrichtian age from w estern N o rth America, w ith description of a
tiny tail club from A lberta and discussion of tail orientation an d tail club
function." C anadian Tournal of Earth Sciences 32: 902-912.
Cracraft, J. (1988). The m ajor clades of birds. The Phylogeny and Classification
of the Tetrapods. M. Benton. Oxford, C larendon Press. Systematics
Association, Special Vol. No. 35A: 333-355.
Cubo, J. and A. Casinos (2000). "M echanical properties and chem ical com position
of avian long bones." E uropean Tournal of M orphology 38(2): 112-121.
C u rre y J. D. (1988). "The effect of porosity an d m ineral content o n th e Young's
m odulus of elasticity of com pact bone." Tournal of Biomechanics 21(2):
131-139.
C u rrey J. D. (1990). "Physical characteristics affecting the tensile failure
properties of com pact bone." Tournal of Biomechanics 23(8): 837-844.
C u rre y J. D. (1999a). "The design of m ineralized h a rd tissues for their m echanical
functions." The Tournal of Experim ental Biology 202: 3285-3294.
C u rre y J. D. (1999b). "W hat determ ines the bending strength of com pact bone?"
The Tournal of Experim ental Biology 202: 2495-2503.
C u rrey J. D. (2001). O ntogenetic changes in com pact bone m aterial properties.
Bone M echanics H andbook. S. C. Cowin. Boca Raton, CRC Press: 19,1-16.
C u rrey J. D. (2002). Bones: Structure and M echanics. N ew Jersey, Princeton
U niversity Press.
Cyron, B. M. and W. C. H utton (1978). "The fatigue strength of the lum bar n eu ral
arch in spondylolysis." Toumal of Bone and Toint Surgery 60B: 234-238.
106
Cyronz B. M v W. C. H u tto n and J. D. G. Troup (1976). "Spondylolytic fractures."
Journal of Bone and Toint Surgery 58B: 462-466.
de Pinna, M. C. C. (1991). "Concepts and tests of hom ology in th e cladistic
paradigm ." Cladistics 7: 367-394.
de Ricqlesz A. J. (1980). Tissue structure of dinosaur bone: functional significance
an d possible relation to dinosaur physiology. A Cold Look at the W armBlooded D inosaurs. R. D. K. Thom as and C. Olson. W estviewz Boulder,
CO, A m erican A ssociation for the A dvancem ent of Science, Selected
Sym posium 28:103-139.
de Ricqles, A. J., K. Radian and J. R. H orner (2001). The bone histology of basal
birds in phylogenetic and ontogenetic perspectives. N ew Perspectives on
the O rigin a nd Early Evolution of Birds: Proceedings of the International
Sym posium in H onor of John H. O strom . J. G authier and L. F. Gall. N ew
H aven, CT, Peabody M useum of N atu ral H istory: 411-426.
de Ricqles, A. J., K. Radian, J. R. H orner and H. Fancillon-Veillot (2000).
"Paleohistology of the bones of pterosaurs (Reptilia: A rchosauria):
an ato m y ontogeny, and biom echanical im plications." Zoological Journal
of the L innean Society 129: 349-385.
Dilkes, D. W. (2000). "A ppendicular m yology of the h ad ro sau rian dinosaur
M aiasaura peeblesorum from the Late Cretaceous (C am panian) of
M ontana." Transactions of the Royal Society of E dinburgh. E arth Science
90: 87-125.
Dollo, L. (1886). "N ote sur Ies ligam ents ossifies des dinosauriens de Befnissart."
Archives de Biologie 7: 249-264.
Erickson, G. M., K. Curry-Rogers and S. A. Yerby (2001). "D inosaurian grow th
patterns and rap id avian grow th rates." N ature 412: 429-433.
Fancillon-Vieillot, H., V. de Buffrenil, J. Castanet, J. G eraudie, F. J. M eunier, J.
Y. Sire, L. Zylberberg and A. J. de Ricqles (1990). M icrostructure and
m ineralization of verebrate skeletal tissues. Skeletal Biom ineralization:
Patterns, Processes and Evolutionary Trends. J. G. Carter. N ew York, Van
N ostrand Reinhold. I: 471-530.
107
Felisbino, S. L. and H. R Carvalho (1999). "Identification an d distribution of type
VI collagen in tendon fibrocartilages." Toumal of Subm icroscopic Cytology
an d Pathology 31:187-195.
Forster, C. A. (1990). "The posteranial skeleton of the ornithopod dinosaur
Tenontosaurus tilletti." Tournal of Vertebrate Paleontology 10(3): 273-294.
Frey, E. (1982). D er Bau des B ew egungsapparates de Krokodile u n d seine
F unktion bei der auatischen Fortbew egung. Fakultat Biologie. Tubingen.
Frey, E. (1985). Biomechanics of terrestrial locom otion in crocodiles. Principles
of C onstruction in Fossil and Recent R eptiles. Sonderforschungsbereich.
Stuttgart and Tubingen, U niversity of Stuttgart/U niversity of Tubingen. 4:
145-167.
Frey, E., J. Riess and S. Tarsitano (1989). "The axial tail m usculature of recent
crocodiles and its phyletic im plications." A m erican Zoologist 29: 857-862.
G alton, R M. (1969). The pelvic m usculature of the dinosaur h y p so lophodon
(reptilia: ornithischia). N ew haven, CT, Peabody M useum of N atural
History, No. 131: 64.
G alton, R M. (1970). "The p osture of h ad ro sau rian dinosaurs." Tournal of
Paleontology 44(3): 464-473.
G alton, R M. (1981). "D ryosaurus, a hypsilophodontid dinosaur from the
U pper Jurassic of N o rth A m erica and Africa; poster anial skeleton."
Palaeontologische Zeitschrift 55(3-4): 271-312.
Case, J. R (1981). Axial M usculature. Biology of the Reptilia. C. Gans and T. S.
Parsons. London, A cadem ic Press: 355-435.
G atesy S. M. (1994). "N eurom uscular diversity in archosaur deep dorsal thigh
m uscles." Brain Behavior and Evolution 43:1-14.
G atesy S. M. (1995). Functional evolution of the hindlim b and tail from basal
theropods to birds. Functional M orphology in V ertebrate Paleontology. J.
J. Thom ason. N ew York, C am bridge U niversity Press: 219-234.
G authier, J. (1986). "Saurischian m onophyly and the origin of birds." California
108
A cadem y of Sciences. M em oir 8:1-55.
G uoz X. E. (2001). M echanical properties of cortical and cancellous bone tissue.
Bone M echanics H andbook. S. C. Cowin. Boca Ratonz CRC Press: IOz 1-19.
H aines, R. W. and A. M ohuiddin (1968). "M etaplastic bone." Tournal of A natom y
103(3): 527-538.
H arveyz E. B.z H. E. Kaiser and L. E. R osenberg (1968). A n atlas of the dom estic
turk ey (Meleagris gallopavo). M yology and Osteology. US A tom ic Energy
C om m ission. W ashington DC.
H aw kins, J. A., C. E. H ughes and R. W. Scotland (1997). "Prim ary hom ology
assessm ent, characters and character states." Cladistics 13: 275-283.
H orner, J. R., A. J. de Ricqles and K. Padian (1999). "Variation in dinosaur
skeletochronology: im plications for age assessm ent an d physiology."
Paleobiology 25(3): 295-304.
H orner, J. R., A. J. de Ricqles and K. Radian (2000). "Long bone histology of the
h a d ro sau rid dinosaur M aiasaura peeblesorum : grow th dynam ics and
physiology based on an ontogenetic series of skeletal elem ents." Journal of
V ertebrate Paleontology 20(1): 115-129.
H orner, J. R., K. Radian an d A. J. de Ricqles (2001). "C om parative osteology of
som e em bryonic and perinatal archosaurs: developm ental an d behavioral
im plications for dinosaurs." Paleobiology 27(1): 39-58.
H utchinson, J. R. (2001a). "The evolution of fem oral osteology an d soft tissues on
the line to extant birds (N eornithines)." Zoological Journal of the Linnean
Society 131(2): 169-197.
H utchinson, J. R. (2001b). "The evolution of pelvic osteology an d soft tissues on
the line to extant birds (N eornithines)." Zoological Journal of the Linnean
Society 131(2): 123-168.
H utchinson, J. R. and M. Garcia (2002). "Tyrannosaurus w as n o t a fast runner."
N atu re 415:1018-1021.
IES (2000). Visual A nalysis 4.0. Bozeman.
109
Johnson, L. C. (1960). M ineralization of Turkey Leg Tendon. I. H istology and
H istochem istry of M ineralization. Calcification in Biological System s. R.
R Sognnaes. W ashington, A m erican Association for the A dvancem ent of
Science: 117-128.
K annus, P. (2000). "Structure of the tendon connective tissue." Scandinavian
Tournal of M edicine and Science in Sports 10: 312-320.
K irkendall, D. T. and W. E. G arrett (1997). "Function and biom echanics of
tendons." Scandinavian Tournal of M edicine and Science in Sports 7: 62-66.
K oehnlein, H. (1930). "K nochenbildung in Sehnen." Arch. LX Klin. Chir. 163:147172.
K usuzaki, K., N. Kageyama, H. Shinjo, H. Takeshita, H. M urata, S. H ashiguchi, T.
A shihara and Y. H irasaw a (2000). "D evelopm ent of bone canaliculi d u rin g
bone repair." Bone 27(5): 655-659.
Ladd, A. J., J. H. Kinney, D. L. H a u p t and S. A. G oldstein (1998). "Finite-elem ent
m odeling of trabecular bone: com parison w ith m echanical testing and
determ ination of tissue m odulus." Tournal of O rthopedic Research 16(5):
622- 628 .
Landis, W. J., J. J. Librizzi, M. G. D un n and R H. Silver (1995). "A stu d y of the
relationship betw een m ineral content an d m echanical properties of tu rk ey
gastrocnem ius tendon." Tournal of Bone M ineral Research 10(6): 859-867.
Landis, W. J. and R H. Silver (2002). "The structure and function of norm ally
m ineralizing avian tendons." C om parative Biochemistry and Physiology
P a rtA 133:1135-1157.
Lauder, G. V. (1981). "Form and function: structural analysis in evolutionary
m orphology." Paleobiology 7(4): 430-442.
Lauder, G. V. (1991a). Biomechanics and evolution: integrating physical and
historical biology in the study of com plex systems. Biomechanics in
E volution. J. M. V. R ayner and R. J. W ootton. C am bridge, N ew York,
C am bridge U niversity Press: 1-19.
Lauder, G. V. (1991b). A n evolutionary perspective on the concept of efficiency:
no
h ow does it evolve? Efficiency and econom y in anim al physiology. R. W.
Blake. N ew York, C am bridge U niversity Press: 169-184.
Lauder, G. V. (1995). O n the inference of function from structure. Functional
M orphology in Vertebrate Paleontology. J. J. Thom ason. N ew York,
C am bridge U niversity Press: 1-18.
Lauder, G. V., A. M. Leroi and M. R. Rose (1993). "A daptations and history."
Trends in Ecology and Evolution 8(8): 294-297.
Lauder, G. V. and S. M. Reilly (1996). The m echanistic bases of behavioral
evolution: a m ultivariate analysis of m usculoskeletal function.
Phylogenies and the com parative m ethod in anim al behavior. E. R
M artins. N ew York, O xford U niversity Press: 104-137.
Lieberkuhn, N. (1860). "U ber die ossification. Die Ossification des
Sehnengew ebes." Arch. Anat. Physiol. W issens. M ed. 1860: 824r846.
Likins, R. C., K. A. Piez and M. L. K unde (1960). M ineralization of Turkey
Leg Tendon. III. Chem ical N ature of the Protein and M ineral Phases.
Calcification in Biological System s. R. F. Sognnaes. W ashington, A m erican
A ssociation for the A dvancem ent of Science: 117-128.
Lull, R. S. an d N. E. W right (1942). "H ad ro sau rian dinosaurs of N o rth America."
Geological Society of America, Special Papers 40: 242.
M artin, R. B., D. B. Burr an d N. A. Sharkey (1998). Skeletal tissue m echanics. N ew
York, Springer.
M olnar, R. E. and E. Frey (1987). "The paravertebral elem ents of the A ustralian
ankylosaur M inm i (Reptilia: O rnithischia, Cretaceous)." N eues Tahrbuch
fu r Geologie u n d Palaontologie, A bhandlungen 75(1): 19-37.
M oodie, R. L. (1927). ,,Studies in Paleopathology, XX. Vertebral lesions in
the sabre-tooth. Pleistocene of California, resem bling the so-called
m yositis ossificans progressiva com pared w th certain ossifications in the
dinosaurs." A nnals of M edical FIistory 9(1): 91-102.
M oodie, R. L. (1928). "The histological n ature of ossified tendons found in
dinosaurs." A m erican M useum N ovitates 311:1-15.
Ill
N igg, B. M. and W. H erzog (1999). Biomechanics of the M usculo-Skeletal System .
Chichester, N ew York, Wiley.
N ishi, S. (1938). M uskeln des R um ples. H an d b u ch der vergleichebnden
A natom ie der W irbeltiere. Bd. 5. L. Bolk, E. G dppert, E. Kallius and W.
Lubosch. Berlin, U rban u n d Scwarzenberg: 351-446.
N orm an, D. B. (1980). "O n the ornithischian dinosaur Iguanodon bernissartensis
from the low er Cretaceous of B ernissart (Belgium)." U In stitu t Royal des
Sciences N aturelles de Belgique M em oire 178:1-103.
N orm an, D. B. (1986). "O n the anatom y of Iguanodon atherfieldensis (Ornithischia:
O rnithopoda)." Bulletin de TInstitut Royal des Sciences N aturalles de
Belgique. Sciences de la Terre 56: 281-372.
N ylen, M. U., D. B. Scott and V. M. M osley (1960). M ineralization of Turkey
Leg Tendon. II. Collagen-M ineral Relations Revealed b y electron and
X-Ray Microscopy. Calcification in Biological System s. R. F. Sognnaes.
W ashington, A m erican Association for the A dvancem ent of Science: 117128.
O strom , J. H. (1964). "A reconsideration of the paleoecology of h ad ro sau rian
dinosaurs." A m erican Tournal of Science 262: 975-997.
O strom , J. H. (1969). "O steology of Deinonychus antirrhopus, an u n u su al theropod
from the low er Cretaceous of M ontana." Peabody M useum of N atural
History. Yale U niversity Bulletin 30:165.
O strom , J. H. (1990). D rom aeosauridae. The D inosauria. D. B. W eisham pel, R
D odson and H. Osmolska. Berkeley, U niversity of C alifornia Press: 269279.
Parks, W A. (1924). "D yoplosaurus acutosqam eus, a new genus an d species of
arm oured dinosaur, and notes on a skeleton of Prosaurolophus maximus."
Geology Series. U niversity of Toronto StudiesilSk 50.
Patterson, C. (1982). M orphological characters and homology. Systematics
A ssociation Special Volume 21: Problem s of Phylogenetic R ecosntruction.
K. A. Joysey and A. E. Friday. N ew York, Academic Press: 21-74.
112
Prieto-M arquez, A. (2001). O steology and Variation of B rachylophosaum s
canadensis (D inosauria, H adrosauridae) from the U p p er Cretaceous
Ju dith River F orm ation of M ontana. D epartm ent of Earth Sciences.
Bozeman, M ontana State University: 390.
R ayheld, E. J., D. B. N orm an, C. C. H orner, J. R. H orner, P. M. Smith, J. J.
T hom ason and R U pchurch (2001). "C ranial design an d function in a large
theropod dinosaur." N ature 409:1033-1037.
Reid, R. E. H. (1996). "Bone histology of the Cleveland-Llyod dinosaurs and of
dinosaurs in general, p a rt I: introduction: introduction to bone tissues."
Brigham Young U niversity Geology Studies 41: 25-72.
Reid, R. E. H. (1997). H istology of bones and teeth. Encyclopedia of D inosaurs. R
J. C urrie and K. Padian. San Diego, CA, A cadem ic Press: 329-339.
Reilly, D. T. and A. H. B urstein (1975). "The elastic and ultim ate properties of
com pact bone tissue." Journal of Biomechanics 8: 393-405.
Retterer, E. and A. Lehevre (1911). "Phenonm enes cytologiques des tendons des
oiseaux en voie d'ossification." Soc. Biol. C.R. Seanc. M em. 71: 596-599.
Rigby, B. J., N. H irai, J. D. Spikes and H. Eyring (1959). "The m echanical
properties of ra t tail tendon." Tournal of G enefal Physiology 43: 265-283.
Romer, A. S. (1923a). "C rocodilian pelvic m uscles and their avian and reptilian
hom ologues." Bulletin of the A m erican M useum of N atu ral H istory 48:
533-552.
Romer, A. S. (1923b). "The pelvic m usculature of saurischian dinosaurs." Bulletin
of the A m erican M useum of N atu ral H istory 48: 605-617.
Romer, A. S. (1927). "The pelvic m usculature of ornithischian dinosaurs." Acta
Zoologica 8(3): 225-275.
Rooney, R (1994). Intratendinous Ossification. Bone Volume 8: M echanism s of
Bone D evelopm ent and G row th. B. K. Hall. Boca Raton, CRC Press: 47-84.
Roth, V. L. (1994). W ithin and betw een organism s: replicators, lineages, and
hom ologues. H om ology: the hierarchical basis of com parative biology. B.
113
K. Hall. London, Academ ic Press.
R othschild, B. M. (1987). "Diffuse Idiopathic skeletal hyperostosis as reflected
in the paleontologic record: dinosaurs and early m am m als." Sem inars in
A rthritis and R heum atism 17(2): 119-125.
Rothschild, B. M. and L. D. M artin (1993). Paleopathology D isease in the Fossil
R ecord. Boca Raton, CRC Press.
Rothschild, B. M. and D. Tanke (1992). "Paleopathology of Vertebrates: Insights
to Lifestyle and H ealth in the Geological Record." Geoscience C anada
19(2): 73-82.
Roy, M. E., J.-Y. Rho, T. Y. Tsui, N. D. Evans and G. M. Pharr (1999). "M echanical
and m orphological variation of the h u m an lum bar vertebral cortical and
trabecular bone." Tournal of Biomedical M aterial Research 44:191-197.
R ydzew ski, W. (1935). «Etudes sur Ie n o tariu m chez Ies Oiseaux.» Acta
O rnithologica M usei Zoologici Polonici I: 403-427.
Saha, S. (1977). "L ongitudinal shear properties of h u m an bone an d its
constituents, and the associated failure m echanism s." Tournal of M aterial
Science 12:1798-1806.
Santa Luca, A. P. (1980). "The postcranial skeleton of Heierodontosaurus tucki
(Reptilia, Ornithischia) from the Strom berg of South Africa." A nnals of the
South A frican M useum 79(7): 159-211.
Sereno, R C. (1986). "Phylogeny of the bird -h ip p ed dinosaurs (order
O rnithischia)." N ational G eographic Research 2(2): 234-256.
Sereno, R C. (1999). "The evolution of dinosaurs." Science 284: 2137-2147.
Sibley, C. G. (1994). "O n the phylogeny and classification of living birds." Tournal
of Avian Biology 25: 87-92.
Stark, J. M. and A. C hinsam y (2002). "Bone m icrostructure and developm ental
plasticity in birds and other dinosaurs." Tournal of M orphology 254: 232246.
114
Storerz R. W. (1982). "Fused thoracic vertebrae in birds: Their occurence and
possible significance." Tournal of the Yamshina Institute of O rnithology 14:
86-95.
Sues, H.-D. and P. M. G alton (1987). "A natom y and classification of the
N o rth A m erican Pachycephalosauria (Dinosauria: O rnithischia)."
Palaeontographica 198:1-40.
Sum m ers, A. R and T. J. Koob (2002). "The evolution of ten d o n - m orphology and
m aterial properties." C om parative Biochemistry and Physiology. Part A
133:1159-1170.
Suzuki, D., G. M urakam i an d N. M inoura (2002). "H istology of th e bone-tendon
interfaces of limb m uscles in lizards." A nnals of A natom y 184: 363-377.
Traub, W., T. A rad and S. W einer (1989). "Three-dim ensional o rd ered distribution
of crystals in turkey tendon collagen fibers." Proceedings of the N ational
A cadem y of Sciences 86: 9822-9826.
Turner, C. H., T. W ang and D. B. Burr (2001). "Shear Strength a n d Fatigue
Properties of H u m an Cortical Bone D eterm ined from P u re Shear Test."
Calcified Tissue International 69: 373-378.
Vallois, H. V. (1922). "Les transform ations de la m usculature de Tepisome chez
Ies vertebres." Arch. M orph, gen, exper. 13:1-538.
V anden Berge, J. C. and R. W Storer (1995). "Intratendinous ossification in birds:
a review." Tournal of M orphology 226(1): 47-77.
Viidik, A. (1966). Biomechanics and functional adaptation of ten d o n s and joint
ligam ents. Studies on the A natom y an d Function of Bone an d Toints. F. G.
Evans. N ew York, Springer: 17-38.
Walker, A. D. (1977). Evolution of the pelvis in birds and dinosaurs. Problem s
of V ertebrate Evolution. S. M. A ndrew s, R. S. Miles and A. D. Walker.
London, A cadem ic Press: 319-357.
W eidenreich, F. (1930). Das K nochengew ebe (G robgebundelte G rundsubstanz
- Faserknochen). H andbuch-der M ikroskopischen A natom ie de M enschen.
Berlin, Springer: 409-416.
115
W eisham pelz D. B. (1995). Fossils, function, and phytogeny. Functional
M orphology in Vertebrate Paleontology. J. J. Thom ason. N ew York,
C am bridge U niversity Press: 34-54.
W eisham pel, D. B. and J. R. H orner (1990). H adrosauridae. The D inosauria.
D. B. W eishampel, P. D odson and H. Osm olska. Berkeley U niversity of
California Press: 534-561.
W ilson, J. W. (1994). H istological techniques. Vertebrate paleontological
techniques. R Leiggi and R May. NY, C am bridge U niversity Press. 205234: 329-339.
W inkler, D. A., R A. M urry and L. L. Jacobs (1997). "A new species of
Tenontosaurus (Dinosauria; O rnithopoda) from the Early Cretaceous of
Texas." Tournal of Vertebrate Paleontology 17(2): 330-348.
W itmer, L. M. (1995). The extant phylogenetic bracket and the im portance of
reconstructing soft tissues in fossils. Functional M orphology in Vertebrate
Paleontology. J. J. Thom ason. N ew York, C am bridge U niversity Press: 1933.
Zusi, R. L. an d G. D. Bentz (1984). "M yology of the P urple-throated Carib
(Eulam pis jugularis) and other hum m ingbirds (Aves: Trochilidae)."
Sm ithsonian C ontributions to Zoology 385:1-70.