Journal of Vertebrate Paleontology 25(3):602–613, September 2005 © 2005 by the Society of Vertebrate Paleontology
THE HISTOLOGY OF OSSIFIED TENDON IN DINOSAURS CHRISTOPHER L. ORGAN1,* and JASON ADAMS2,** 1 Lewis Hall, Department of Cell Biology and Neuroscience, Montana State University, Bozeman, Montana 59717, U.S.A.; 2 Department of Paleontology, Museum of the Rockies, Montana State University, Bozeman, Montana 59717, U.S.A.
1
ABSTRACT—Intratendinous ossification is widespread in dinosaurs (including birds). Although intratendinous ossification in living birds is well understood, the physiological process of tendon metaplasia and associated histological variability in Dinosauria are not. Therefore, ossified tendons were histologically sampled across extinct dinosaurian clades. Tendons of living birds and alligators were also sampled. Despite various anatomical locations and large differences in body size, ossified tendons were found to possess uniform microstructure even in specimens that do not normally experience intratendinous ossification (such as Spinostropheus and Camarasaurus). The ossified tendons of non-avian dinosaurs are largely indistinguishable from skeletal bone with respect to microanatomical features. However, ossified tendons in birds lack periosteal bone and associated fibrolamellar structures associated with ornithischian dinosaur tendons. Variation in periosteal bone occurs along the length of individual tendons. Ossified tendons from marginocephalians are unique in that they have large quantities of anisotropic fibrolamellar bone, while those from pachycephalosaurids have radial vascularity.
INTRODUCTION Tendons are an important part of the vertebrate musculoskeletal system that transfer forces from muscle to bone. Tendon ossification is considered unusual because it only occurs pathologically in mammals, with the exception of the kangaroo’s tail (Kram and Dawson, 1998). But intratendinous ossification is common in many dinosaurian clades, including birds (Vanden Berge and Storer, 1995) and in nearly all ornithischian dinosaurs, where their presence is synapomorphic of the clade (Sereno, 1999). Avian ossified tendons are found in many body parts such as the leg (Landis and Silver, 2002), back (Rydzewski, 1935), neck (Boas, 1929), and wing (Vanden Berge and Storer, 1995). Ornithischian tendons are morphologically similar to (though much larger than) ossified tendons found in birds. However, whereas in Neornithes they occur throughout the body, ossified tendons are found only along the vertebral column in ornithischian dinosaurs. Small ornithopod dinosaurs have tendon bundles that run along the dorsal aspect of the spinal column (Forster, 1990). Hypaxial tendons are also present in the tails of these animals. Iguanodontids and hadrosaurids possess a rhomboidal trellis of ossified tendons along the dorsal side of their spinal column (Brown, 1933; Lull and Wright, 1942; Norman, 1980, 1986). Ankylosaurids have a series of angled tendons in their tails that extend anteriorly from the “club” (Coombs, 1995). Enlarged bundles of tendons are often found at the base of pachycephalosaurid neural spines (Sues and Galton, 1987). There are also fossilized ossified tendons in taxa that apparently did not normally undergo intratendinous ossification. For example, Minmi is the only ankylosaurid known to possess an epaxial trellis of ossified tendons in the dorsal region (Molnar and Frey, 1987). These isolated cases are more common than might be expected in non-avian theropods (MOR 693 and MNN TIG6) and sauropods (DNM28). Thus, ossified tendons are found in a wide phylogenetic range within Dinosauria. Because they are located in many different
* Present Address Organismic and Evolutionary Biology, Harvard University, 26 Oxford Street, Cambridge, MA 02138, corgan@oeb .harvard.edu ** Present Address 49120 KE Road, Mesa, CO 81643, adams81643@ excite.com
areas in the body, they might be expected to develop by a variety of physiological mechanisms and to have disparate functions. Historically, only isolated tendon fragments have been described histologically and there are no data on whether ossified tendon histology varies according to anatomical region within an individual or across taxa. It is also unknown whether ossified tendon microstructure changes in different parts of the body in birds, the only living taxon in which tendons regularly ossify. Therefore, the primary aim of this study was to characterize ossified tendon histology across Dinosauria and to determine their histological variability within different body parts and across taxa. This was accomplished by histological sampling of tendons from animals that normally ossify tendons and from animals with abnormally ossified tendons. Tendons from living archosaurs (crocodilians and birds) were sampled to provide a phylogenetic basis for histological interpretations. MATERIALS AND METHODS Fossil tendons were sampled from groups (Table 1) as broadly as possible within Dinosauria (Fig. 1). Fossilized ossified tendon samples consisted largely of associated material from the vertebral column of adult specimens, but several (i.e., MNN TIG6 and MOR-794) were articulated with the skeleton. All fossilized tendons were sectioned across their mid-shaft, with the exception of those taken from MOR-794, which were sampled from the dorsal, sacral, and caudal regions along the axial skeleton. Sections from MOR-794 were taken from tendon termini and shafts across the three laterally stacked layers. Histological procedure followed standard paleohistology techniques (Wilson, 1994). Tendons from extant taxa were taken directly from dissected specimens (Table 1). These animals were donated postmortem to the Museum of the Rockies. Tendons of Meleagris gallopavo were sampled from M. gastrocnemius, M. fibularis longus, and M. longus colli dorsalis thoracica from 16-week-old hens and 23-week-old toms. A tendon of Bubo virginianus was sampled from the M. extensor carpi radialis. The talar retinaculum and the M. gastrocnemius tendons were sampled from a specimen of Alligator mississippiensis. A tendon from the M. longus colli dorsalis thoracica was sampled from a specimen of Podilymbus podiceps.
602
ORGAN AND ADAMS—OSSIFIED TENDON HISTOLOGY TABLE 1.
COMPARATIVE HISTOLOGICAL DESCRIPTION
Specimens used in this study Specimen
Outgroup (Crocodylia) Alligator mississippiensis* Saurischia (Sauropoda) Camarasaurus lentus Saurischia (Theropoda) Allosaurus fragilis Spinostropheus gautieri Deinonychus antirrhopus Saurornitholestes Saurischia (Theropoda: Aves) Podilymbus podiceps* Meleagris gallopavo* Bubo virginianus* Ornithischia (Ornithopoda) Brachylophosaurus canadensis Edmontosaurus Hypacrosaurus stebingeri Tenontosaurus tilletti Ornithischia (Marginocephalia) Stygimoloch spinifer Pachyrhinosaurus canadensis Ormithischia (Thyreophora) Euoplocephalus tutus
Specimen ID
Histo ID
—
2003-19r
DNM28
1997-9c
MOR 693 MNN TIG6 MOR 747 MOR 660
1995-1 2003-09c 2002-06 1993-2
MOR MOR MOR MOR
603
— — —
2003-21r 2003-20r 2001-10r
794 1142 549 680
1999-20 2001-09 1996-10 2002-04
UCMP 128383 TMP 89.55.647
1993-1 2003-14c
TMP-83.36.120
2003-15c
Asterisks denote living taxa. Institutional Abbreviations: DMM, Denver Natural History Museum; MOR, Museum of the Rockies; TMP, Royal Tyrrell Museum of Paleontology; UCMP, University of California Berkeley Museum of Paleontology; MNN TIG, Muse´ e National du Niger.
Extant samples were fixed in 10% neutral buffered formalin for storage. Mineralized tendons were not decalcified before sectioning. All samples were stained with hematoxylin and eosin (H&E), except that for Bubo, which was stained with toluidine blue. Images were taken from slides using an electronic camera. Microanatomical terminology used here follows established nomenclature for bone (Currey, 2002; de Ricqlès, 1980; FrancillonVieillot et al., 1990; Reid, 1996) and tendon microanatomy (Kannus, 2000).
Crocodilian Tendon Histology Collagen fibers consisting of longitudinally arrayed fibrils (Fig. 2B, fibrils not apparent) characterize the crocodilian tendons. Fibroblasts (tenocytes) run along these fibers (Fig. 2B), which are sparsely arranged and cylindrical in longitudinal sections. Collagen fibers demonstrate crimp common to tendon tissue. Fibers are bound together by endotenon sheaths to form fascicles (Fig. 2A). The peritenon (Fig. 2A), which is demarcated by transverse orientation of tenocytes, surrounds the endotenons. Ornithischian Tendon Histology Ornithopoda—Within Hadrosauridae, three taxa were sampled: a lambeosaur (Hypacrosaurus) and two hadrosaurs (Brachylophosaurus and Edmontosaurus). Aside from being epaxial, the exact anatomical locations of the tendons are unknown for Hypacrosaurus and Edmontosaurus. Tendons from Brachylophosaurus were systematically sampled along individual tendons from the dorsal, sacral, and caudal regions. Tendons in Hadrosauridae are peripherally composed of fibrolamellar bone with longitudinally oriented primary osteons. Haversian canals, which are also present, become dense at the center (Fig. 3A) and are identical to similar tissue from long bone compacta. In some larger tendons, Haversian replacement is dense with many generations of osteons (Fig. 3B). Osteocyte lacunae and canaliculi are not always easily identifiable. Some collagen fibers are visible between osteons (Fig. 3F), although the center is nearly cancellous in some tendons. These fibers sometimes become dense toward the periphery, which often contains an external fundamental system (EFS; Cormack, 1987) denoted by tightly packed lines of arrested growth (LAGs) (Fig. 3A and E). The pseudolamellar bone of the EFS indicates mature bone growth. Sections taken across the three tendon layers from the Brachylophosaurus are histologically identical. Also, there is no observable histological difference in samples taken from the dorsal,
FIGURE 1. Phylogenetic framework of the species used in this study. Stem-based groups are labeled at the branches and node-based groups are labeled at nodes. Modified from Gauthier (1986), Cracraft (1988), and Sereno (1999). Nodes: 1, Archosauria; 2, Dinosauria; 3, Ornithischia; 4, Neornithischia; 5, Marginocephalia; 6, Ornithopoda; 7, Hadrosauridae; 8, Saurischia; 9, Theropoda; 10, Deinonychosauria; 11, Neornithes.
604
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 3, 2005
FIGURE 2. Histological photographs of Alligator tendons. A, cross section of Achilles tendon detailing endotenon divisions, fibroblasts, and a peritenon sheath (100×). B, the same tendon in longitudinal section; endotenon divisions not obvious (40×). Abbreviations: ET, endotenon; FB, fibroblasts; PT, peritenon.
sacral, or caudal regions. However, variation occurs along individual tendons. Sections taken through the termini are filled with more collagen fibers and less lamellar tissue than sections taken through the mid-shaft. In Brachylophosaurus, LAGs vary along tendon length, becoming more abundant at the center. Several samples were taken from a pathological region in the dorsal aspect of the tail (a common occurrence in hadrosaurid tails). Grossly, some tendons in this region continue over the pathological neural spines uninterrupted, whereas others are eroded. Histologically, this tissue is similar to that in the termini of other tendons of Brachylophosaurus (Fig. 3D). Both have large erosion rooms. However, the number and size of erosion rooms in the pathological tendon are much greater. Whereas there is no evidence of the pathological origin (e.g., bacterial infection), the tissue suggests that active growth and repair were occurring at the time of death. Within Ornithopoda, tendons of Tenontosaurus were also sampled. Epaxial tendons from the caudal region contain longitudinally oriented collagen fibers and primary osteons in the periphery. Several generations of Haversian canals are present at the center, which contains many large resorption cavities that form a reticular cancellous region. As in the hadrosaurid tendons, some fibers are present between osteons, even in the central region with dense Haversian bone (Fig. 4A and B). Some tendons also contain LAGs and an EFS. Marginocephalia—Tendons from the pachycephalosaur Stygimoloch contain a core nearly identical to those seen in ornithopod tendons. That is, tissue is characterized by an abundance of longitudinally oriented collagen fibers. These are interrupted by resorption spaces and Haversian canals. Vascular orientation parallels the fiber bundles and fibrolamellar bone surrounds the core. Vascularity is reduced in this region and has alternating orientation from longitudinal to radial (Fig. 5A). Immediately outside the core, vascularity is longitudinally oriented. More distally is a band of radially oriented canals, followed by a band of longitudinally oriented canals. Radially oriented vascular spaces dominate the perimeter. Spread through the fibrolamellar section are isolated areas of roughly bundled tissue consisting of enlarged collagen fibers (Fig. 5A). These areas look identical to the tendon’s center and are globular. Osteocyte lacunae in the fibrolamellar regions are smaller than those in regions dominated by rough-bundled collagen fibers. The tendon from the ceratopsian Pachyrhinosaurus is epaxial,
though its exact location along the spinal column is unknown. A cancellous region occupies the tendon’s center (Fig. 5B). The bone that forms this cancellous region is lamellar. The periphery consists of fibrolamellar tissue with longitudinal canals (Fig. 5C). Multigenerational Haversian systems are present at the border between the two bone regions, but are obliterated centrally by the lamellar tissue (Fig. 5D). Lines of arrested growth are absent. Thyreophora—Samples from a tail of Euoplocephalus (Ankylosauridae) consist of dense Haversian bone (Fig. 6A). Some erosion rooms are dispersed through the Haversian bone. Vascular orientation is longitudinal. Haversian systems are so developed that no trace of collagen fibers is present except close to the surface (Fig. 6B). LAGs are present in the peripheral region. Osteocyte lacunae are larger in this region and circumferentially oriented. Saurischian Tendon Histology Sauropoda—The abnormal tendon of Camarasaurus contains primary collagen fibers and a few primary osteons in the periphery (Fig. 7A). This region is not uniform circumferentially, and in some areas it consists of multiple generations of Haversian canals. Other portions of the periphery are filled with relatively avascular, roughly bundled tissue. Dense Haversian tissue occurs throughout the center and lacks collagen fibers between canals. The tendon’s center also contains resorption cavities. No LAGs are present in the periphery. Osteocyte lacunae are apparent as are filamentous canaliculi. Non-Avian Theropod—Several abnormal ossified tendons from the sacral neural spines of the ceratosaurid Spinostropheus (Sereno et al., 2004) appear grossly similar to other ossified tendons sampled in this study. Histologically, longitudinally oriented mineralized collagen fibers comprise the tendon (Fig. 8A). The periphery is marked by LAGs (Fig. 8B). Haversian canals are present centrally. Fibers are large and roughly bundled interior to the first LAG and finely bundled outside it. This change in fiber structure may indicate the boundary of the original tendon. An Allosaurus tendon possesses a periphery of vascularized fibrolamellar bone tissue (Fig. 8C). Vascular orientation is longitudinal throughout the tendon. Primary osteons within a longitudinally arranged fiber matrix characterize the periphery. Centrally, the tendon is composed of multiple generations of much larger Haversian canals and resorption rooms (Fig. 8D).
ORGAN AND ADAMS—OSSIFIED TENDON HISTOLOGY
605
FIGURE 3. Histological photographs of tendons of Hadrosauridae. A, tendon from Hypacrosaurus (cross section, 40×). Notice multiple lines of arrested growth (LAGs), dense Haversian bone, and primary osteons. B, tendon from Edmontosaurus (cross section, 100×). Note Haversian canals and collagen fibers between Haversian bone. C, a Haversian canal from Hypacrosaurus denoted by a cement line (cross section, 400×). D, pathological section of tendon from Brachylophosaurus (cross section, 40×). Note the high proportion of erosion rooms. E, an EFS from a Brachylophosaurus tendon (cross section, 100×). F, collagen fibers and a simple vascular canal from a Brachylophosaurus tendon (cross section, crossed nicols, 400×). Abbreviations: CL, cement line; CF, collagen fibers; ER, erosion room; HC, Haversian canals; PO, primary osteons; SVC, simple vascular canal.
The sizes of the latter collectively form a central cancellous region. Osteocyte lacunae are generally ovoid in the fibrous tissue and flattened in lamellar tissue. Lines of arrested growth occur throughout the tendon. The interior-most line occurs just beyond the central Haversian region. Lines of arrested growth occur sporadically toward the periphery, becoming tightly packed at the subsurface, forming an EFS. Tail rods (elongated prezygapophyses and chevrons) from the coelurosaurian theropod Saurornitholestes are poorly preserved. They consist of dense Haversian bone (Fig. 9A), which extends to the surface in some places in all tendons. Few resorption
cavities are present. The small periphery is filled with longitudinally arranged collagen fibers. Multiple LAGs are visible in all tendons. These lines become tightly packed near the surface, indicating an external fundamental system. The tail rods from the coelurosaur Deinonychus contain tightly spaced rings consisting of ordered avascular fibrous tissue (Fig. 9C). This is finely structured tissue with sparse osteocyte lacunae. A few primary osteons are present in some rods, though two rods contain concentric rings of primary osteons in the periphery. Several other rods (Fig. 9B) contain central Haversian systems. They are large and longitudinally oriented adjacent to
606
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 3, 2005
FIGURE 4. Histological photographs of Tenontosaurus tendons. A, central portion of the tendon, dense with Haversian canals (cross section, 100×). Note collagen fibers between the canals. B, collagen fibers and a simple vascular canal (cross section, 400×). Abbreviations: CF, collagen fibers; HC, Haversian canals; SVC, simple vascular canal.
the vertebral body. Distal to the vertebrae, the rods become avascular. Centrally, all rods contain a loosely packed fiber core (Fig. 9D). Neornithes—Ossified tendons in the turkey Meleagris contain a mineralized fiber core in which collagen fiber bundles are still apparent in cross section (Fig. 10A). These bundles are scallopshaped and separated from the surrounding matrix, giving the mineralized tendon a scaly appearance. In tendons that are heavily ossified, due to advanced age or variation within an individual, the spaces among the scalloped fiber bundles are filled with a new osteoid matrix that gives the tissue a more fascicular appearance. Centrally, large irregular erosion rooms are common. A few Haversian systems are present. Our observations of the tendon histology of Meleagris agree with published descriptions (Abdalla, 1979; Lieberkuhn, 1860; Weidenreich, 1930). Longitudinal sections clearly show mineralized areas with fibers that have lost their characteristic waviness (crimp) and become straight (Fig. 10B). Fibroblasts in mineralized regions are hypertrophied and contain large amounts of cytoplasm. These cells look much like hypertrophied chondrocytes. Long chains of hypertrophied fibroblasts run down the long axis. As with the cross sections, the central area is mineralized with straight fibers. This core is surrounded by unmineralized tendon tissue that has wavy fibers and elongated tenocytes. Some tendons (e.g., many shank tendons) extend the zone of transformation into the myotendinous part to produce a fanshaped myotendinous origin. This morphology is similar to the flared ends of ornithischian ossified tendons, which were also probably myotendinous. The zone of transformation also extends into the tendon entheses (bone insertions) in many epaxial tendons, which ossify to the neural spine summits. Despite this difference, there is no difference in microanatomy between leg and epaxial tendons. Thin sections from tendons in the owl Bubo look similar to those of Meleagris, but have a very thick peritenon. These thin sections were taken 1 cm proximal to the insertion point of the M. extensor carpi radialis. A clear demarcation occurs between the central osseous core and the unmineralized peritendon (Fig. 10C). The central region consists of primary osteons and firstgeneration Haversian canals (Fig. 10D). Mineralized collagen fibers can be seen between some Haversian canals. Collagen fibers have the same scalloped shape and overall bundled appearance seen in turkey tendons. LAGs and primary osteons are absent from this tissue.
Lateral tendon sections from the tendons of M. longus colli dorsalis thoracica in the grebe Podilymbus are comparable to those from Meleagris. For example, collagen fibers run straight along the longitudinal axis, and chains of hypertrophied tenocytes are common (Fig. 10E). Fibroblast shape is fusiform when they are not in chains. Ossified tendons from Podilymbus lack an unmineralized outer tendinous sheath (peritendon). This is evident where the M. longus colli dorsalis thoracica interfaces directly with the ossified tendon without an unmineralized intermediate (Fig. 10F). Therefore, the original myotendinous junction is replaced by a fleshy attachment along the length of the tendon. DISCUSSION The Alligator and unmineralized avian tendons are histologically indistinguishable (compare Figs. 2 and 10). Furthermore, archosaur and mammalian tendon microanatomy appear identical (Kannus, 2000). These data support the claim that tendon microanatomy is apomorphic for Vertebrata (Summers and Koob, 2002). Within vertebrates, the only derived microanatomy in tendons appears to occur in the ossified tendons of dinosaurs. The plesiomorphic microanatomy of archosaur tendons establishes a basis from which intratendinous ossification in dinosaurs can be compared. Various authors have described ossified tendon microanatomy in birds and non-avian dinosaurs (Abdalla, 1979; Broili, 1922; Johnson, 1960; Koehnlein, 1930; Lieberkuhn, 1860; Moodie, 1928; Weidenreich, 1930). Ossified tendons have been categorized with “grobgebündelte Grundsubstanz–Faserknochen” or “rough bundled bone tissue” by Weidenreich (1930). Lieberkuhn also described the straight collagen fibers in longitudinal section and the scalloped fascicular appearance of fiber bundles in cross section. Indeed, longitudinally isotropic fibers and vascularity are characteristic of ossified tendons. This orientation reflects the cord-like shape of the tendon precursor tissue. Lieberkuhn (1860) identified chondrocytes in bird tendons, later correctly termed hypertrophied tenocytes by Retterer and Lelievre (1911). These cells are responsible for the mineralization process (Landis and Silver, 2002). Biomineralization occurs in ossified tendons, cartilage caps, and tendon and ligament entheses, collectively termed metaplastic bone (Haines and Mohuiddin, 1968; Koehnlein, 1930). Specifically, metaplastic bone is the transformation of fully differentiated tissue into bone. Haines and Mohuiddin (ibid.) identified this phenomenon in
ORGAN AND ADAMS—OSSIFIED TENDON HISTOLOGY
607
FIGURE 5. Histological photographs from marginocephalian tendons. A, composite image from Stygimoloch (cross section, 40×). Note the alternating longitudinal (LC) and radial (RC) canals. Also, note the central core of Haversian bone. B, tendon of Pachyrhinosaurus (cross section, 10×). C, same tendon at 40×. Note the outer region of fibrolamellar bone and interior region of lamellar bone. D, same tendon at 100× in the lamellar region. Abbreviations: LB, lamellar bone, LC, longitudinal canals; RC, radial canals; FLB, fibrolamellar bone.
tenocytes on the unmineralized side of the blue line (mineralization tide mark) in tendon entheses. In this region, hypertrophied fibroblasts occur in long chains, a structure that appears in ossified tendons as well (Figs. 10B and E). Although these cells are still often called chondrocytes due to their cellular morphology (e.g., Suzuki et al., 2002), this is inappropriate given the cell’s original location and function (Landis and Silver, 2002; Retterer and Lelievre, 1911). Haines and Mohuiddin (1968) also noted that osteocyte canaliculi in metaplastic bone are absent, stunted, or looped. This feature was observed in some ossified tendons in this study. But this type of canaliculi is restricted to mineralized tendon tissue. The lamellar bone (osteons) formed in ossified tendons is indis-
tinguishable from that of normal bone, and this observation agrees with previous studies (Lieberkuhn, 1860; Weidenreich, 1930). Thus, highly remodeled ossified tendons lose all traces (hypertrophied tenocytes, collagen fibers, and absent or looped canaliculi) of original tendon tissue and metaplastic bone. The only indicative feature left in these cases is isotropic longitudinal vascular orientation, which is a product of the tendon’s cord-like gross morphology. Metaplasia is thought to be a distinct process of ossification because fibroblasts take on the role of osteoblasts (Haines and Mohuiddin, 1968). But histological characteristics of metaplastic bone also occur during endochondral and intramembranous bone formation, which continues through a process of remodel-
608
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 3, 2005
FIGURE 6. Histological photographs of tendons of Euoplocephalus. A, the whole tendon in cross section (10×). B, detail of periphery in the same tendon (100×). Note how close to the surface multigenerational Haversian canals extend. Also, note that the primary tissue contains many lines of arrested growth. Erosion rooms are also present. Abbreviations: ER, erosion room; LAG, lines of arrested growth; HC, Haversian canals.
ing and periosteal growth that obliterates the precursor tissue. All bone is initially formed from precursor tissue, whether cartilaginous, mesenchymal, or tendinous. Only the timing and degree of ossification distinguishes metaplasia from the process of normal bone formation. Therefore, “metaplastic bone” should only refer to the late timing and limited degree of ossification. This delay can be part of the normal developmental process, as in avian ossified tendons (Johnson, 1960), part of the transition of tissue types at entheses (Haines and Mohuiddin, 1968), or part of a wide range of pathological conditions such as mechanical stress, injury, disease, and aging (Rothschild, 1987). Mature ossified tendons in birds are composed of metaplastic bone because ossification begins during the juvenile growth stage (Landis and Silver, 2002) and halts before periosteal growth and remodeling occur. Ossified tendons in mature non-avian dinosaurs are largely devoid of metaplastic bone because they begin to ossify in the embryo (Rothschild and Tanke, 1992) and are composed of remodeled and periosteal bone tissue. The difference between these tissue types is almost certainly due to changes in growth strategies that evolved on the line to birds (de Ricqlès et al., 2001; Horner et al., 1999). The high density of Haversian canals and the presence of LAGs, primary osteons, and an EFS indicate that the bone comprising non-avian ossified tendons is as old as the skeletal ele-
ments (Horner et al., 1999). However, thin sections taken along the length of individual ossified tendons from the hadrosaurid Brachylophosaurus reveal that LAGs, primary osteons, and the EFS are variable. Consequently, whenever possible, histological study of ossified tendons should not rely on isolated samples because variance in histology along tendon length could dramatically alter interpretations. For example, a fragment from the center of an ossified tendon would look histologically much older than a sample from near a terminus. The histological features of ossified tendons are uniform across different anatomical regions. There is essentially no difference in microanatomy between leg tendons and those from the back. The wing tendon microstructure of Bubo is also similar to the tendon microstructure of Meleagris. However, the peritenon is much thicker, forming a solid border around the osseous core. Given the development of Haversian bone, an indicator of late-stage intratendinous ossification, a thick peritenon may be a characteristic of ossified tendons in wings. The anatomical location of the tendon and associated wing function may account for the thickened peritenon, or it may be taxic variation. Unlike those of other birds, ossified tendons of Podilymbus do not possess an unmineralized peritenon. This difference might be caused by age-related differences among specimens. However, it shows that the M. longus colli dorsalis thoracica can directly
FIGURE 7. Histological photographs of tendons of Camarasaurus. A, cross section of tendon (40×). Note the peripheral region composed of collagen fibers and the central core of Haversian bone. B, the central region of the tendon (cross section, 100×). Note the collagen fibers between some Haversian canals. Abbreviations: CF, collagen fibers; HB, Haversian bone; HC, Haversian canals.
ORGAN AND ADAMS—OSSIFIED TENDON HISTOLOGY
609
FIGURE 8. Histological photographs of non-avian theropod tendons. A, cross section across neural spine and ossified tendons in Spinostropheus (10×). B, same field, tendon with Haversian canals (100×). Note the roughly bundled appearance of the fibers interior to the first line of arrested growth (1st LAG). C, a tendon of Allosaurus (cross section, 40×). Note the fibrolamellar structure with primary osteons and an EFS. D, central region of the same tendon (100×). Note the erosion room, Haversian canals, primary osteons, and centrally located LAGs. Abbreviations: ER, erosion room; HC, Haversian canals; LAG, line of arrested growth; NS, neural spine; OT, ossified tendon; PO, primary osteons.
interface with the ossified tendon by a fleshy attachment. The homologous (in the taxic sense) epaxial ossified tendons in Meleagris do not interface in a direct bone-to-muscle attachment. The epaxial tendons of Hadrosauriformes are homologous (in the taxic sense) with those of birds (Organ, 2003). Because most ossified tendons from non-avian dinosaurs are highly ossified, they probably interfaced with epaxial muscles like those in Podilymbus (by a direct fleshy attachment). Samples taken from both marginocephalians (Stygimoloch and Pachyrhinosaurus) have more mature bone than tendons from all other dinosaurs. Stygimoloch tendons possess a loosely packed core consisting of Haversian bone and longitudinally oriented collagen fibers. Islands of this bone type are found throughout the outer region, which is formed from tightly packed fibrolamellar tissue. This suggests that several tendons formed into a single large ossified tendon. The vascular orientation in the outer region alternates between longitudinal and radial. This “plywood-like” structure has been reported in pterosaur bones (de Ricqlès et al., 2000). Pterosaurs had very thin bone walls and this structure was presumably related to increased material strength similar to histological adaptations to torsional loading in birds (de Margerie, 2002). The microanatomy of the Stygimoloch tendon is the only known ossified tendon to have such a structure. All other ossified tendons are extremely isotropic, because vascular canals and collagen fibers are oriented parallel to the long axis. This sample may reveal relaxation of a developmental constraint. That is, all other ossi-
fied tendons are constrained in their collagen and vascular orientation due to the underlying isotropic tendon structure and function. The plywood-like vascular orientation suggests that non-tensile mechanical stress impacted pachycephalosaurid tendons. The other marginocephalian tendon from Pachyrhinosaurus was also more remodeled than other tendons examined in this study. While it does not contain the alternating vascular structure found in Stygimoloch, it possesses a cortical region filled with fibrolamellar bone of anisotropic fibers and a medullary region composed of lamellar tissue. Unlike non-marginocephalians, it lacks uniformly oriented collagen fibers. In fact, there are no histological features to suggest that this tissue is an ossified tendon. Although the Pachyrhinosaurus tendon suggests greater intratendinous ossification maturity in ceratopsians, age or developmental plasticity could account for its unique histology. Small samples sizes for the marginocephalian tendons necessitate further corroboration of the anisotropic vascularity and osseous maturity encountered in this study. Tail Rod Tendons Prezygapophyses elongate progressively in Tetanurae (Gauthier, 1986). Thus, it would seem that the “tail rod tendons” of coelurosaurians are just prezygapophyseal extensions. In this scenario, they should be formed much like vertebral bone. However, the prezygapophysis and chevron extensions of Deinony-
610
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 3, 2005
FIGURE 9. Histological photographs of non-avian theropod ‘tail rods’. A, cross section of multiple ‘tail rods’ from Saurornitholestes (10×). Taphonomic alteration makes detailed interpretation difficult, but note the dense Haversian bone. B, multiple ‘tail rods’ and vertebral body of Deinonychus (cross section, 10×). Note that some exterior rods are vascular. These are proximal sections close to their origin. C, a single ‘tail rod’ (cross section 100×). Note the fine-grained rings and a core of loosely arranged fibers with two Haversian canals. D, central region of a ‘tail rod’ (cross section, crossed polars, 400×). Note the uniform arrangement of longitudinally oriented primary fibers. Abbreviations: HC, Haversian canals; VB, vertebral body.
chus have been described as developing through tendon sheaths that have taken on periosteal functions (Ostrom, 1969, 1990; Reid, 1997). Considering the intratendinous ossification process in birds (Abdalla, 1979; Johnson, 1960; Landis and Silver, 2002), this seems unlikely because metaplasia begins inside the peritenon. The periphery is the last region to mineralize and ossify in normal intratendinous ossification. But this is not the case in the ‘tail rod’ tendons. Indeed, the opposite seems true. They lose vascularity distal to the vertebral body and have an outer region of fine fibrous bone and a central core of loosely organized fibers. This unusual histology is most likely the result of small amounts of appositional growth compared with large amounts of longitudinal growth. Because there are no extant analogues for comparison, the development of these tail rods remains unknown. However, given their histology and the trend of prezygapophyseal elongation in Tetanurae (Gauthier, 1986), they are decidedly not ossified tendons, but rather vertebral extensions. A histological survey of tetanuran zygapophyses could help illuminate the tissue’s developmental nature. Zygapophyseal development may also have functional implications because of their impact on caudal muscle structure. Further research is needed to reconstruct the caudal muscles associated with these rods in coelurosaurians. A similar question to that of coelurosaurian tail rods was addressed by Unwin et al. (1996) on the pteroid bone in pterosaurs. The pteroid is a wrist bone apomorphic to pterosaurs that con-
trolled the propatagium. Using histological evidence, Unwin et al. concluded that the pteroid is a neomorphic bone and not a metaplastic ossified tendon. Given the remodeling and periosteal growth in non-avian dinosaur ossified tendons, the pteroid could have had a tendinous origin (assuming pterosaurs grew like nonavian dinosaurs). However, the remnants of rough bundled tendon fibers are absent and the gross morphology of the pteroid is different from all known ossified tendons. Ossified tendons also lack articular surfaces, a feature the pteroid clearly possesses. Therefore, the results of this study support the hypothesis that the pteroid is a neomorphic bone. CONCLUSIONS In their review paper on ossified tendons in birds, Vanden Berge and Storer (1995) found no physiological or biomechanical function that accounts for their sporadic occurrence in Aves. The wing tendon of Bubo contains a peritenon sheath much larger than that of other ossified tendons. However, because Bubo does not have ossified tendons in other areas of the body for comparison, it is unknown if this feature is phylogenetic or functional. Within individual specimens of Meleagris, we found no variation in metaplastic development or histological structure among tendons from different anatomical locations. Because different anatomical locations are under different stresses, these data seem to contradict the common notion that tendons ossify
ORGAN AND ADAMS—OSSIFIED TENDON HISTOLOGY
611
FIGURE 10. Histological photographs of ossified tendons of Neornithes. A, tendon of Meleagris containing a mineralized core surrounded by a paratenon (cross section, 40×). Compare the matrix of mineralized collagen fibers, Haversian canals, and erosion rooms with Figure 2. B, longitudinal section of same tendon (100×). Note the exterior unmineralized fibers and the straightened mineralized fibers that contain hypertrophied fibroblasts. C, single tendon from Bubo at bifurcation of osseous core (cross section, 10×). Note the thick peritenon surrounding the osseous cores. D, the core of the same tendon (100×). E, tendon from Podilymbus (longitudinal section, crossed polars, 100×). Note the chain of hypertrophied fibroblasts. F, another Podilymbus tendon (longitudinal section, 400×). Note obliquely angled muscle tissue attaching directly onto the surface of the ossified tendon. Abbreviations: ER, erosion room; HC, Haversian canals; HF, hypertrophied fibroblasts; MF, mineralized collagen fibers; M, muscle tissue; OC, osseous core; OT, ossified tendon; PT, paratenon; PO, primary osteons, UF, unmineralized fibers.
in response to biomechanical stress (Landis and Silver, 2002), which implies that different tendons throughout the body ossify at different rates as they are subjected to different tensile stresses. Ossified epaxial tendons occur over the notarium, which is entirely rigid, suggesting that stress had little to do with the cause of their ossification. Thus, caution should be used when inferring the adaptive significance of intratendinous ossification in birds. Inferring the function of intratendinous ossification in extinct dinosaurs is even more problematic. Histological features also seem to be free of body-size-related constraints as well. That is, the histology of tendons in birds
appears virtually identical (save the expanded peritenon in Bubo) irrespective of body size. The same is true of non-avian dinosaurs, from Spinostropheus to Tenontosaurus. Furthermore, ossified tendons are normally absent in large non-dinosaurian vertebrates. Thus, size mediated ossification, through progressively greater tensile stress application, does not seem to affect intratendinous ossification. Why do tendons commonly ossify in dinosaurs while they are absent in other vertebrate groups? As discussed above, body size, anatomical location, and mechanical stress can be eliminated as possible causes. The absence of apparent adaptive func-
612
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 3, 2005
tion may imply that ossified tendons were not the direct product of natural selection. Consequently, there are two possible explanations. The first is that a physiological propensity to ossify tendons evolved early in dinosaurs and then drifted to fixation in different subgroups. According to this scenario, selection for ossified tendons was neutral and population sizes were likely restricted in multiple groups. The second, more probable, explanation is that intratendinous ossification is like a spandrel (Gould and Lewontin, 1979) whose evolution was not the direct product of selection, but correlated to physiological mineral use or selection for skeletal co-ossification. For example, there is a progressive trend in most dinosaurian clades for vertebral fusion with the sacrum (Sereno, 1999). In Aves, co-ossification of the skeleton has resulted in the synsacrum, pygostyle, and notarium. Ossified tendons might be a consequence of selection for ossification in these structures. This interpretation does not imply that ossified tendons lack specific functions in every instance; only that adaptive function cannot be assumed a priori. Biomechanical modeling and developmental studies are needed to clarify the biological significance of ossified tendons in dinosaurs. ACKNOWLEDGMENTS We thank Paul Sereno, William Simpson, Janet Hinshaw, and Jim Gardner for access to and discussion of specimens. We also thank Robert Storer for insightful discussions of intratendinous ossification in birds. The Seder Ridge Turkey Farm, Ruth Elsey (Rockefeller Wildlife Refuge, Louisiana Department of Wildlife and Fisheries), and the Bozeman Raptor Center kindly donated extant specimens. Erich Staudacher helped to translate papers in Old German. Martha Middlebrooks, Tobin Hieronymus, Walter Coombs, Jack Horner, Cynthia Marshall, and Armand de Ricqlès are greatly thanked for invaluable discussions of histology, as is Ellen Lamm for instruction and technical assistance with thin sectioning. Kevin O’Neill, Matt Lavin, Gwen Jacobs, and Deborah King are thanked for their guidance and support. Comments from Nicole Hobbs, Tim Van Tassel, and Katie Organ improved the manuscript. We are particularly grateful to Kevin Padian and an anonymous reviewer for extremely helpful comments. This research was funded in part by a doctoral grant from the International Society of Biomechanics, the Museum of the Rockies (Department of Paleontology), and the Department of Cell Biology and Neuroscience at Montana State University. LITERATURE CITED Abdalla, O. 1979. Ossification and mineralization in the tendons of the chicken (Gallus domesticus). Journal of Anatomy 129:351–359. Boas, J. E. V. 1929. Biologisch-anatomische studien uber den hals der vogel. Det Kongelige Danske Videnskabernes Selskabs skrifter Naturvidenskabelig og Mathematick Afdelning Series 9:105–222. Broili, F. 1922. Uber den feineren Bau der ⬘verknochnerten Sehnen⬘ (verknochteren Muskeln) von Trachodon. Anatomischer Anzeiger 55:465–475. Brown, B. 1933. A gigantic ceratopsian dinosaur, Triceratops maximus, new species. American Museum Novitates 649:1–9. Coombs, W. P. J. 1995. Ankylosaurian tail clubs of middle Campanian to early Maastrichtian age from western North America, with description of a tiny tail club from Alberta and discussion of tail orientation and tail club function. Canadian Journal of Earth Sciences 32: 902–912. Cormack, D. 1987. Ham’s Histology. Lippincott, New York, 732 pp. Currey, J. D. 2002. Bones: Structure and Mechanics. Princeton University Press, New Jersey, 480 pp. de Margerie, E. 2002. Laminar bone as an adaptation to torsional loads in flapping flight. Journal of Anatomy 201:521. de Ricqlès, A. J. 1980. Tissue structure of dinosaur bone: functional significance and possible relation to dinosaur physiology; pp. 103–139 in R. D. K. Thomas and C. Olson (eds.), A Cold Look at the
Warm-Blooded Dinosaurs. American Association for the Advancement of Science, Selected Symposium 28, Westview, Boulder, CO. de Ricqlès, A. J., K. Padian, and J. R. Horner. 2001. The bone histology of basal birds in phylogenetic and ontogenetic perspectives; pp. 411–426 in J. Gauthier and L. F. Gall (eds.), New Perspectives on the Origin and Early Evolution of Birds: Proceedings of the International Symposium in Honor of John H. Ostrom. Peabody Museum of Natural History, New Haven, Connecticut. de Ricqlès, A. J., K. Padian, J. R. Horner, and H. Francillon-Veillot. 2000. Paleohistology of the bones of pterosaurs (Reptilia: Archosauria): anatomy, ontogeny, and biomechanical implications. Zoological Journal of the Linnean Society 129:349–385. Forster, C. A. 1990. The postcranial skeleton of the ornithopod dinosaur Tenontosaurus tilletti. Journal of Vertebrate Paleontology 10: 273–294. Francillon-Vieillot, H., V. de Buffrénil, J. Castanet, J. Géraudie, F. J. Meunier, J. Y. Sire, L. Zylberberg, and A. J. de Ricqlès. 1990. Microstructure and mineralization of vertebrate skeletal tissues; pp. 471–530 in J. G. Carter (ed.), Skeletal Biomineralization: Patterns, Processes and Evolutionary Trends. Van Nostrand Reinhold, New York. Gauthier, J. 1986. Saurischian monophyly and the origin of birds. California Academy of Sciences, Memoir 8:1–55. Gould, S. J., and R. C. Lewontin. 1979. The spandrels of San Marco and the Panglossion paradigm: a critique of the adaptationist programme. Proceedings of the Royal Society of London, Series B 205:581–598. Haines, R. W., and A. Mohuiddin. 1968. Metaplastic bone. Journal of Anatomy 103:527–538. Horner, J. R., A. J. de Ricqlès, and K. Padian. 1999. Variation in dinosaur skeletochronology: implications for age assessment and physiology. Paleobiology 25:295–304. Johnson, L. C. 1960. Mineralization of Turkey Leg Tendon. I. Histology and Histochemistry of Mineralization; pp. 117–128 in R. F. Sognnaes (ed.), Calcification in Biological Systems. American Association for the Advancement of Science, Washington. Kannus, P. 2000. Structure of the tendon connective tissue. Scandinavian Journal of Medicine and Science in Sports 10:312–320. Koehnlein, H. 1930. Knochenbildung in Sehnen. Archiv LX Klinik Chirurgischen 163:147–172. Kram, R., and T. J. Dawson. 1998. Energetics and biomechanics of locomotion by red kangaroos (Macropus rufus). Comparative Biochemistry and Physiology, Part B 120:41–49. Landis, W. J., and F. H. Silver. 2002. The structure and function of normally mineralizing avian tendons. Comparative Biochemistry and Physiology, Part A 133:1135–1157. Lieberkuhn, N. 1860. Uber die Ossification. Die Ossification des Sehnengewebes. Archiv für Anatomie, Physiologie und Wissenschaftliche Medicin 1860:824–846. Lull, R. S., and N. E. Wright. 1942. Hadrosaurian dinosaurs of North America. Geological Society of America, Special Papers 40:242. Molnar, R. E., and E. Frey. 1987. The paravertebral elements of the Australian ankylosaur Minmi (Reptilia: Ornithischia, Cretaceous). Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen 75: 19–37. Moodie, R. L. 1928. The histological nature of ossified tendons found in dinosaurs. American Museum Novitates 311:1–15. Norman, D. B. 1980. On the ornithischian dinosaur Iguanodon bernissartensis from the lower Cretaceous of Bernissart (Belgium). L’Institut Royal des Sciences Naturelles de Belgique, Memoire 178: 1–103. Norman, D. B. 1986. On the anatomy of Iguanodon atherfieldensis (Ornithischia: Ornithopoda). Bulletin de l’Institut Royal des Sciences Naturelles de Belgique, Sciences de la Terre 56:281–372. Organ, C. L. 2003. Epaxial muscles and ossified tendons in dinosaurs: Anatomy, development, histology, and biomechanics. Ph.D. dissertation, Department of Cell Biology and Neuroscience, Montana State University, Bozeman, Montana, 115 pp. Ostrom, J. H. 1969. Osteology of Deinonychus antirrhopus, an unusual theropod from the lower Cretaceous of Montana. Peabody Museum of Natural History, Yale University, Bulletin 30:1–165. Ostrom, J. H. 1990. Dromaeosauridae; pp. 269–279 in D. B. Weishampel, P. Dodson, and H. Osmólska (eds.), The Dinosauria. University of California Press, Berkeley. Reid, R. E. H. 1996. Bone histology of the Cleveland-Lloyd dinosaurs
ORGAN AND ADAMS—OSSIFIED TENDON HISTOLOGY and of dinosaurs in general, part I: introduction to bone tissues. Brigham Young University Geology Studies 41:25–72. Reid, R. E. H. 1997. Histology of bones and teeth; pp. 329–339 in P. J. Currie and K. Padian (eds.), Encyclopedia of Dinosaurs. Academic Press, San Diego, California. Retterer, E., and A. Lelièvre. 1911. Phénomènes cytologiques des tendons des oiseaux en voie d’ossification. Compte Rendu des Seances de la Société de Biologie, Paris 71:596–599. Rothschild, B. M. 1987. Diffuse idiopathic skeletal hyperostosis as reflected in the paleontologic record: dinosaurs and early mammals. Seminars in Arthritis and Rheumatism 17:119–125. Rothschild, B. M., and D. Tanke. 1992. Paleopathology of vertebrates: insights to lifestyle and health in the geological record. Geoscience Canada 19:73–82. Rydzewski, W. 1935. Études sur le notarium chez les Oiseaux. Acta Ornithologica Musei Zoologici Polonici 1:403–427. Sereno, P. C. 1999. The evolution of dinosaurs. Science 284:2137–2147. Sereno, P. C., J. A. Wilson, and J. L. Conrad. 2004. New dinosaurs link southern landmasses in the mid-Cretaceous. Proceedings of the Royal Society of London, Series B 271:1325–1330. Sues, H.-D., and P. M. Galton. 1987. Anatomy and classification of the
613
North American Pachycephalosauria (Dinosauria: Ornithischia). Palaeontographica 198:1–40. Summers, A. P., and T. J. Koob. 2002. The evolution of tendon - morphology and material properties. Comparative Biochemistry and Physiology, Part A 133:1159–1170. Suzuki, D., G. Murakami, and N. Minoura. 2002. Histology of the bonetendon interfaces of limb muscles in lizards. Annals of Anatomy 184:363–377. Unwin, D. M., E. Frey, D. M. Martill, J. B. Clarke, and J. Riess. 1996. On the nature of the pteroid in pterosaurs. Proceedings of the Royal Society of London, Series B 263:45–52. Vanden Berge, J. C., and R. W. Storer. 1995. Intratendinous ossification in birds: a review. Journal of Morphology 226:47–77. Weidenreich, F. 1930. Das Knochengewebe (Grobgebündelte Grundsubstanz - Faserknochen); pp. 409–416 in Handbuch der Mikroskopischen Anatomie de Menschen. Springer, Berlin. Wilson, J. W. 1994. Histological techniques; pp. 329–339 in P. Leiggi and P. May (eds.), Vertebrate Paleontological Techniques. Cambridge University Press, NY. Submitted 7 October 2004; accepted 4 January 2005.