Bone Histology in Dysalotosaurus lettowvorbecki (Ornithischia: Iguanodontia) – Variation, Growth, and Implications
Citation: H ubner TR (
Bone Histology in Dysalotosaurus lettowvorbecki (Ornithischia: Iguanodontia) - Variation, Growth, and Implications
Tom R. Hu bner 0
Vincent Laudet, Ecole Normale Superieure de Lyon, France
0 Niedersa chsisches Landesmuseum Hannover , Hannover , Germany
Background: Dysalotosaurus lettowvorbecki is a small ornithopod dinosaur known from thousands of bones and several ontogenetic stages. It was found in a single locality within the Tendaguru Formation of southeastern Tanzania, possibly representing a single herd. Dysalotosaurus provides an excellent case study for examining variation in bone microstructure and life history and helps to unravel the still mysterious growth pattern of small ornithopods. Methodology/Principal Findings: Five different skeletal elements were sampled, revealing microstructural variation between individuals, skeletal elements, cross sectional units, and ontogenetic stages. The bone wall consists of fibrolamellar bone with strong variability in vascularization and development of growth cycles. Larger bones with a high degree of utilization have high relative growth rates and seldom annuli/LAGs, whereas small and less intensively used bones have lower growth rates and a higher number of these resting lines. Due to the scarcity of annuli/LAGs, the reconstruction of the life history of Dysalotosaurus was carried out using regularly developed and alternating slow and fast growing zones. Dysalotosaurus was a precocial dinosaur, which experienced sexual maturity at ten years, had an indeterminate growth pattern, and maximum growth rates comparable to a large kangaroo. Conclusions/Significance: The variation in the bone histology of Dysalotosaurus demonstrates the influence of size, utilization, and shape of bones on relative growth rates. Annuli/LAGs are not the only type of annual growth cycles that can be used to reconstruct the life history of fossil vertebrates, but the degree of development of these lines may be of importance for the reconstruction of paleobehavior. The regular development of annuli/LAGs in subadults and adults of large ornithopods therefore reflects higher seasonal stress due to higher food demands, migration, and altricial breeding behavior. Small ornithopods often lack regularly developed annuli/LAGs due to lower food demands, no need for migration, and precocial behavior.
Ever since scientists began to work with the remains of those
extinct animals that lack direct living descendants, they dreamed
of being able to accurately reconstruct life histories and, at least
partially, social structures and behavior. Unfortunately, it is almost
impossible to obtain such fundamental information using only
morphological and/or statistical methods, because absolute
ontogenetic dates of age or time of sexual maturity are not
determinable. Size classes within a bonebed of a single species,
surface texture of bones, or degree of suture closure are examples
of tools often used to estimate relative age and ontogenetic status of
fossil animals, but these methods are always highly imprecise (e.g.
). The study of bone histology has enabled paleontologists
partially filling this methodological gap, because its insights can
provide the required absolute data in many cases (see e.g. 
for a general introduction into bone histology and common terms).
The basal iguanodontian ornithopod dinosaur Dysalotosaurus
lettowvorbecki was the subject of this study. Dysalotosaurus was found
during the famous German Tendaguru expeditions of 1909 to
1913, which took place 60 km west of the seaport of Lindi,
southeast Tanzania [11,12]. In contrast to the abundant remains
of sauropods and the stegosaur Kentrosaurus, Dysalotosaurus is known
from only a single locality, but the two closely related
monodominant bonebeds found in channel lag deposits 
produced thousands of bones of a minimum number of 100
individuals, from several growth stages, and in all degrees of
disarticulation . Although the genesis of this mass accumulation
has long been discussed as of either catastrophic or attritional
origin , the available taphonomic record currently favors
the catastrophic mortality of a single herd . Preburial
weathering and signs of scavenging (widely distributed bones,
tooth marks, a significant number of shed carnivore teeth) are
absent, which implies fast burial after death. Abrasion is also
unknown and there is only slight sorting of bones in favor of large
and/or robust elements. The bonebeds are therefore
autochthonous or parautochthonous in origin. A preservational difference
between the two bonebeds, the upper of which almost overlies the
lower, is not recognizable. Thus, a single Dysalotosaurus herd was
probably trapped in one of the numerous tidal channels of that
ancient coastal plain , drowned in a spring tide, and their
graveyard was reworked at least once by another spring tide (a
process that can take place every two weeks) resulting in the split
into two separate bonebeds. A more detailed analysis of the
taphonomy of the Dysalotosaurus quarry will be published in a
An ontogenetic series of femora of Dysalotosaurus was previously
studied by Anusuya Chinsamy-Turan  under the name
Dryosaurus lettowvorbecki. The generic name Dysalotosaurus was made
a synonym of Dryosaurus by Galton  due to many
morphological similarities between D. lettowvorbecki and D. altus. However, an
ongoing revision of the anatomy of both taxa (see also  and 
for comments) revealed numerous significant anatomical
differences in several parts of the skeleton, which clearly support the
resurrection of the genus Dysalotosaurus.
Age Estimations via Bone Histology
In many recent tetrapods, one growth cycle commonly
represents one year of time (e.g. [5,6,8,10,21,22,23]), and this
observation has been commonly used to estimate age for extinct
tetrapods (e.g. ). This fact is the basis of the method of
However, an accurate count of the number of annuli and/or
LAGs (Lines of Arrested Growth) is often hampered by the
ontogenetic expansion of the marrow cavity and/or secondary
remodeling. This problem was often solved by the back-calculation
of the lost/obscured number of annuli/LAGs [5,28,32,33,34]) or
by the examination of an ontogenetic series (e.g. [27,35,36]).
There is also a high variability in the number of annuli/LAGs
between different individuals within a single population (e.g. ),
between different skeletal elements of one individual (e.g. [33,36],
and sometimes even in the cross section of a single bone (e.g. ).
For example, single individuals of the dinosaurs Plateosaurus
[32,33], Maiasaura , and Hypacrosaurus  show different
numbers of preserved LAGs in different skeletal elements,
depending upon the general anatomical condition and specific
growth pattern of each of these elements (e.g. cortical thickness,
growth rate, rate of remodeling etc.).
A last important point is the assumption that all annuli/LAGs
counted in a bone are indeed true annual layers. These lines can
also be generated as a result of environmental stress, such as
scarcity of food, illness, or during seasons of pairing or
reproduction . It is also possible to find double LAGs, which
are consistently close together and represent a single year. Some
tropical mammals, for instance, can even generate two cycles per
year . All these deviations from the simple annual model of
growth cycles are rarely discernable in extinct species (e.g. )
and must be treated as sources of error in the calculation of
Another actualistic method used to estimate relative age of
extinct animals is Amprinos Rule (e.g. ). Amprino 
suggested that similar bone tissues in different animals reflect
similar growth rates. It is now widely accepted that maximum
body size seems to be one of the major factors that influences
growth rates, and therefore indirectly influences bone tissue types
. There are also differences in growth rate between
different elements within a single skeleton (e.g. [32,36,40,45]) and
during ontogeny (e.g. [5,18,34,38]). However, recent studies of
birds and reptiles recognized a clear correlation between growth
rate and the size and density of vascular canals, but no correlation
between growth rate and orientation of vascular canals
[40,41,45,46]. Such a correlation seems to exist only due to
extreme environmental conditions, which force an animal to
generate extraordinarily high growth rates . Thus, Amprinos
Rule can help to estimate the growth rate of an extinct species,
but, as for skeletochronology, the results are strongly dependent on
body size, ontogenetic stage, and skeletal element and should
always be considered in comparison with other individuals,
populations, and species.
Bone Histology in Ornithopod Dinosaurs
Ornithopods are one of the best studied dinosaur groups with
regard to bone histology, because several taxa are known from
many individuals of different growth stages [18,34,36,38,48
55]. It has even proved possible to reconstruct the breeding
strategy (altricial or precocial) and life history for some taxa.
However, whereas the growth pattern of large ornithopods is
quite well understood, the bone histology of many small
ornithopods has raised more questions than answers as to their
growth patterns [6,18,48,52,55]. In particular, the scarcity or
even absence of annuli/LAGs, the usual tool for age
estimations, has considerably complicated the reconstruction of their
life history. The recent discovery that annuli/LAGs are indeed
present in Dysalotosaurus and its close relative Dryosaurus ([9,52],
in contrast to ) helped in interpretations of their growth
patterns. However, the inconsistent development of annuli/
LAGs made it necessary to examine another type of growth
cycle for the reconstruction of the life history of Dysalotosaurus
. Additional types of possible annual markers were previously
documented mainly in sauropods (e.g. [56,57,58]). The annual
development of the type of growth cycles used here has been
assumed previously , but the application of these growth
cycles in order to reconstruct life history is successfully made
here for the first time.
Observations of bone tissue types as well as vascular and fibrillar
organization in different skeletal elements of Dysalotosaurus led to
some important insights into the reasons behind these multiple
variations. Furthermore, the highly inconsistent development of
annuli/LAGs and the newly described type of annual growth
cycles resulted in a new hypothesis to explain the differences in
growth patterns between large and small ornithopods.
The description of the microstructure of the sampled bones will
be restricted to the main features of cross sectional shape,
vascularization, and development of growth cycles. Where
appropriate, the microstructure of the femur will also be compared
to the description provided by Chinsamy . A complete version
of the description summarized here is available in the supporting
material (Text S1).
In sum, 70 individual bones were sampled comprising 30
femora, 12 tibiae, 13 humeri, seven fibulae, and eight prepubic
processes, but not all of them could be used for quantitative
analyses due to insufficient preservation.
Bone Histology of the Femur of Dysalotosaurus
Description. The femoral cross section is generally triangular
in shape and becomes more slender close to the base of the fourth
trochanter (see Fig. 1AD for the general orientation). The
respective cross sections of figure 1 in Chinsamy  are
inconsistently oriented, so that the larger section (from a left
femur) is oriented with its anteromedial wall facing ventrally and
the smaller section (from a right femur) is oriented with its
posteromedial corner in that way.
The edge of the marrow cavity is well defined and mainly
consistent, but undulations and cavities are often present internal
to the anterior corner. No spongiosa were observed within the
marrow cavity. A layer of endosteally deposited lamellar bone may
be developed in variable thicknesses around the marrow cavity,
although it never forms a completely surrounding band. One
reason is the resorptive posterior edge of the cavity (e.g. Figs. 1B,
The compact bone wall consists mainly of two types of bone
tissue. Most of it is composed of periosteal fibrolamellar bone
tissue with woven fibered matrix and numerous primary osteons
(Fig. 2AD). Only the anterior corner shows sometimes strongly
birefringent parallel-fibered matrix (Fig. 2EF) and is most likely
the same region mentioned for a large femur by Chinsamy .
The second tissue type, compacted coarse cancellous bone
(CCCB), is of endosteal origin and mostly restricted to the anterior
corner and adjacent areas (marked in red in the sketches of Fig. 1B,
D; 2GH). In more distal sections, the amount of CCCB relative
to fibrolamellar bone, and the average size of the innermost canals
of CCCB, increases.
The vascularization (sensu lato, following ) is very variable in
terms of the size of the canals and overall density. Most of the
vascular canals are well-developed primary osteons. Generally, the
size and density are greatest in the thickest parts of the primary
bone wall (posteromedial corner, Fig. 3AB) and lowest, with
relatively more matrix between the primary osteons, in the
thinnest parts (Fig. 3CD). The latter also include the anterior
corner of the femur, because the CCCB wedge takes up the inner
part of the bone wall in this area and the outer primary bone looks
compressed (Fig. 3EF). The opposite relationship exists for the
degree of organization of vascular canals, where it is highest in the
thinner parts of the primary bone wall (longitudinal to laminar
orientations) and very low in the thickest parts (plexiform to
sometimes reticular orientations; compare Figs. 3CD with 3AB).
An additional tendency is the general increase of vascular
organization from inner parts of the bone wall towards the
periosteal surface. However, the laminar type of vascularization is
the most abundant. The smallest, longitudinal, and fairly
wellorganized primary osteons are observable in the innermost areas
of the primary bone wall around the anterior corner. There are
relatively thick bands of matrix, which isolate these osteons from
each other and which resemble a knitted pattern (Figs. 3EH).
The posterolateral corner represents a special area of the bone
wall (Fig. 4AB). Here, primary osteons are less well developed,
larger on average, and more randomly shaped and oriented than
in all other cross sectional units (Fig. 4CD). This area, which will
be called the Posterolateral Plug in the following text, represents a
very abrupt change within the organization of bone tissue. The
general course of growth cycles, bone laminae, and the orientation
of vascular canals stops at the border of the Posterolateral Plug
(Fig. 4B) and only distinct annuli/LAGs can be followed through
it. This area is most prominent in sections slightly distal to the
midshaft and becomes less prominent proximally, towards the
fourth trochanter. A similar structure is sometimes visible in the
outer cortex of the anterior corner of more proximal sections, and
in larger sections (Fig. 4EH). This cluster, however, does not
significantly disturb the general organization of the tissue and is
also far less widespread.
The zonation pattern is also highly variable. Annuli/LAGs are
present (in contrast to ; Fig. 5AB), but only in 10 out of 30
sampled femora. There is additionally no correlation between the
size of the bone and the number of annuli/LAGs (compare Fig. 1B,
D). None of the cross sections record more than one or two
annuli/LAGs. Nevertheless, these are the only growth cycles that
can be followed around the cross section.
Another type of growth cycle is much more abundant, but less
distinctive than annuli/LAGs because it is often only clearly visible
under polarized light (Fig. 5CD). This type is most developed
within the lateral side of the posterior wall close to the
Posterolateral Plug (Figs. 1AD; 5E). It consists of weakly
birefringent fast growing zones (viewed under polarized light)
with mainly longitudinally oriented collagen fibrils, as well as
numerous and dense primary osteons that show a relatively lesser
degree of organization (Fig. 5FG). The fast growing zones
alternate with more strongly birefringent slow growing zones,
which consist of mainly transversely oriented collagen fibrils and
less dense and more circumferentially oriented primary osteons
that show a relatively higher degree of organization (Fig. 5F, H).
The transition from the fast to the slow growing zone is diffuse.
Only the external rim of the slow growing zones is definable and
possible annuli/LAGs occur mainly in this area. Thus, one growth
cycle consists of an internal fast growing zone and an external slow
growing zone. The slow growing zones often merge together in the
thin parts of the primary bone wall (especially anteriorly) or split
up towards thicker parts, where they even vanish in some areas.
One has therefore to check carefully their number and extension
by repeatedly rotating the cross sections under polarized light. The
Posterolateral Plug interrupts the course of these growth cycles
completely (Figs. 4A; 5CE).
Five out of six of the largest sectioned femora show a transition
(Mark of Initial Sexual Maturity MISM, see below) from the
generally distinct sequence of growth cycles internally to a much
more uniform area externally. The latter resembles a very thick
slow growing zone and only a very weak internal zonation is
recognizable (Figs. 1AB; 5E; 6AE).
Secondary remodeling is very rare, which differs from the
remarks of Chinsamy . There are only local occurrences of
scattered secondary osteons, concentrated mainly in the
transitional area between the primary bone tissue and the CCCB
(Fig. 6FH). Isolated osteons are also present within the latter
(Fig. 2GH). Other isolated occurrences are located within the
Posterolateral Plug (Fig. 4CD) and sometimes in the external part
of the anterior corner (Fig. 2EF). However, sections from more
distal parts of the femur have greater numbers of secondary
osteons throughout the cortex.
The comparison of longitudinal sections of a large femur and of
the smallest sampled femur (Fig. 7) reveals that the amount and
area occupied by pads of calcified cartilage decreases with size, but
there is still a substantial amount present in the large specimen.
The large specimen is much better ossified than the small
specimen, consisting of a dense meshwork of trabecular bone.
However, there is a concentration of bony straps in the epiphyseal
centre of the small specimen, which reaches almost to the distal
Ontogenetic Stages in Femora
Due to the highly variable features within the shaft, between
different femoral cross sections, and even within a single section,
ontogenetic stages are difficult to distinguish. The use of most of
the features, such as the degree of development of primary osteons,
vascularization pattern, or secondary remodeling, was therefore
limited, and there is often a smooth transition between successive
ontogenetic stages. However, useful indicators of ontogenetic stage
are, in addition to absolute size, the number of growth cycles and
the degree of development of distinct areas, such as the
Stage 1 or Embryonic/Perinatal Stage. This stage, already
described in some other ornithopods [36,51,52], is not represented
in the sampled femora of Dysalotosaurus, and the overall size of
other known specimens indicates that none of the preserved
femora would fit into this stage.
Stage 2 or Early Juvenile Stage (Fig. 8AD; Tab. 1). The
marrow cavity is very large compared to the bone wall thickness
(see also ). The internal anterior wedge, if present, consists of
CCCB that is not yet compacted. The posterolateral corner and
the respective Plug are weakly pronounced. The periosteal
compact bone tissue has a high number of longitudinal vascular
canals. The primary osteons are often isolated from each other by
thick bands of well-organized and relatively uniformly birefringent
woven-fibered matrix (knitted texture; Fig. 8BD). Particularly in
the internal part anteriorly, only simple vascular canals are
present. There is at most one slow growing zone developed at the
external edge of the cortex (Fig. 8B).
Stage 3 or Late Juvenile Stage (Tab. 1). The external
circumferential profile is more pronounced and the Posterolateral
Plug is well visible. The drift of the marrow cavity from
approximately anterior to posterior is in progress, which is
indicated by the well-compacted CCCB of a larger anterior
wedge as well as a deeper incision into the posterior bone wall (this
is also dependent on the sectioned level). The primary osteons are
more numerous and there is a decrease in the proportion of
knitted texture. There are the first occurrences of isolated
secondary osteons. Growth cycles are well distinguishable and
reach two to three in number (Fig. 8EH).
Stage 4 or Sexually Immature Stage (Tab. 1). The
development of the external cross-sectional profile as well as of
distinct areas (e.g. the Posterolateral Plug) is now complete
(Fig. 1CD). The anterior wedge of the CCCB is more
pronounced, although this also depends on the relative position
of the cross section within the shaft. The marrow cavity is deeply
incised into the posterior wall (Fig. 1D). The density of
welldeveloped primary osteons is very high in the thick and fast
growing parts of the sections. Secondary osteons are more
abundant and can also occur in the Posterolateral Plug and the
anterior corner (Fig. 2EF). The number of growth cycles is three
Stage 5 or Sexually Mature Stage (Tab. 1). The units of
the cross sectional bone wall are strongly diversified (Fig. 1AB).
The anteroposterior migration of the marrow cavity interrupts up
to four growth cycles posteriorly (Figs. 2AB; 5E). Secondary
osteons are numerous forming clusters anteriorly and
posterolaterally at different distances from the external surface
(Fig. 4CD). The number of growth cycles reaches up to nine and
the transition from well-distinguishable fast and slow growing
zones internally to the diffuse and more uniform wide zone
externally is visible in five of the largest cross sections (Figs. 1AB;
4A; 5CE; 6AE).
Bone Histology of the Tibia of Dysalotosaurus
Description. The cross-sectional shape of the tibia is almost
egg-like in distal sections and almost circular in proximal sections,
but there is always a straight anterior wall, which opposes the
fibula when in articulation (Fig. 1EH). The shape of the marrow
cavity is more symmetrical than the external outline and the rim is
mostly well defined and straight. A slight shift of the marrow cavity
medially is observed in later ontogenetic stages.
An endosteal layer is developed almost exclusively in medium to
large sections (Fig. 9AB; see Tab. 2 for comparable sizes of
samples) with its maximum thickness in the anteromedial or
anterolateral corner. With one exception, the endosteal layer
never completely surrounds the marrow cavity.
As in the femora, the tibial cross sections consist generally of
fibrolamellar bone tissue with a high density of well-developed
primary osteons, which are predominantly organized in a laminar
pattern (Fig. 9AB). The variability in size, density, and
organization of vascular canals/primary osteons is also
comparable to that seen in femora (Fig. 9). CCCB may occur as a wedge in
the anterolateral corner internally, which extends far into the
cortex only in the two largest cross sections (Figs. 1GH; 9EG). In
most of the smaller sections (see below), as well as in the proximal
sections, CCCB is absent. A structure similar to the femoral
Posterolateral Plug is visible in the middle cortex of this corner
(Fig. 9E, H), although its extent within the tibial shaft is much
smaller than in the femur.
The zonation pattern is also similar to that of the femora, with
very few annuli/LAGs and with growth cycles mainly consisting of
fast and slow growing zones (Figs. 1F, H; 9A, E; 10; Tab. 2). The
growth cycles are best preserved in the anterior and/or medial
walls. A transition from distinct growth cycles internally to a
uniform slow growing area externally, as occurs in five large
femora, is not visible in the two large tibiae.
Secondary remodeling is even rarer than in femora. The only
area with preserved secondary osteons is the anterolateral corner
of large (SMNS T3; GZG.V 6791, see Tab. 2) and more distal
sections. Scattered examples are found mainly in the outer area of
the CCCB wedge and within the Anterolateral Plug (Fig. 9H).
In one of the large tibial cross sections (SMNS T3), at the
anterior edge of the marrow cavity, an unusual bone tissue is
preserved (Figs. 1GH; 11). It is strongly cancellous with
irregularly shaped caverns of various sizes. It is weakly birefringent
under polarized light. It is also clearly separated from the compact
bone wall by an endosteal layer (Fig. 11CD, GH). Some of this
tissue was also found inside two large caverns within the
CCCBwedge (Fig. 11AD). All these features, and the absence of any
external pathologies (including a thickening of the bone wall or
bilaterally symmetrical occurrence of unusual tissue as a sign for
osteopetrosis [60,61]), indicate that this tissue belongs to the
endosteal type of tissue called medullary bone, which has already
been documented in three other dinosaur taxa [62,63] (but see
). This tissue is known among living vertebrates only in birds
and it functions as storage for the calcium needed for the
development of eggs in breeding females. Thus, medullary bone
tissue is also a marker for sexually mature females around the
breeding period .
Ontogenetic Stages in the Tibiae
The recognition of distinct ontogenetic stages in the tibiae is
more difficult than in the femora, because there are fewer tibial
sections available, and because most of the available specimens,
belonging to a medium size range (see stage 3 below, Tab. 2), are
probably of the same immature stage. However, the differences
between these and the younger and older stages are substantial,
ilaed ilaed ilaed ilaed ilad edm edm ,inp freo irssa
axm ilaed rtseoom ilaed ilaed rtseoom ilaedm rtseoom ilaed rtseoom irrteon rtseeom ilaed rtseoop rtseoop ilaed ilaedm ilaedm reThoug itaeDm iyeen
T m p m m p + p m p a o m .5 .4 m .1 .8 ag
W .8 .2 .9 .5 .1 .1 .3 .1 .9 .8 p .9 2 1 1 3 0
B 1 5 3 3 _ 7 4 4 4 4 _ 6 6 5 1 1 1 1 1 up ityv Teh
n a ;G t s
a l- ty ce so en
isno edMicav .25 .96 .801 .29 .49 .29 .98 .87 .711 .99 .211 01 01 .151 .941 .602 81 .271 .191 freen -tnpA iltckh
t a um
rsaeud l-eadM .87 .811 _ .151 _ .271 .641 .331 .181 .261 _ .702 12 .452 43 .863 73 .253 .273 ;ithdC lieodm iinMm
owing mainly to the preserved number of growth cycles and the
number and distribution of secondary osteons.
Stage 1 or Embryonic Stage. As in the femora, this stage is
unknown in the tibiae.
Stage 2 or Early Juvenile Stage. Probably only a single tibia
belongs to this stage (GPIT/RE/3795; Fig. 12; Tab. 2). The
different units of the cross section differ only slightly from each
other. An Anterolateral Plug, secondary osteons, CCCB, an
endosteal layer, and resorption activity of the marrow cavity are
all absent. Primary osteons are present, but they are still under
development. Simple, laminarly organized, longitudinal canals are
common, but radial orientations are also visible in the anterolateral
corner (Fig. 12AB, EF). If at all present, only the beginning of the
first slow growing zone is visible at the outer edge of the bone wall.
Stage 3 or Late Juvenile to Sexually Immature
Stage. These cross sections possess much better differentiated
units including the Anterolateral Plug, which occur in distal
sections within the shaft. CCCB, secondary remodeling, and
resorption by the marrow cavity are observed in some sections.
The knitted pattern is now only preserved in the inner cortex,
whereas primary osteons are now well developed and widely
distributed. At least two to three growth cycles are present
(Fig. 1EF; 10CF).
Stage 4 or Sexually Mature Stage. The two largest samples
(SMNS T3; GZG.V 6791) belong to this stage. The cross-sectional
units are strongly differentiated and the bone wall thickness is
highly variable (Fig. 1GH). The CCCB tissue forms a large
wedge, which reaches far into the cortex anterolaterally. There is a
distinct swirl-like Anterolateral Plug within the anterolateral
corner (Fig. 9E). Simple juvenile vascularization is preserved
only as a relict in some of the innermost parts (Fig. 9AB).
Secondary osteons are more abundant within the Anterolateral
Plug (Fig. 9E, G). Primary osteons are dense and numerous. The
number of growth cycles exceeds three. Finally, medullary bone
may be found in one of the cross sections of this stage (Fig. 11).
Bone Histology of the Humerus of Dysalotosaurus
Description. The shape of the cross sections varies from a
lateromedially wide and flat oval outline distally to an almost
circular oval shape more proximally (Fig. 1IL). CCCB is very
rare and only visible in various units in the most distal sections and
in the anterolateral part in the most proximal sections. More
common is the development of an endosteal layer, although it
never surrounds the marrow cavity completely. Proximal sections
often possess a thick but short wedge of endosteal bone in the
anterolateral corner of the cavity (Fig. 13AB).
The bone matrix of the primary compact bone wall consists
mainly of fibrolamellar bone tissue, although the anterolateral
corner can be built by parallel-fibered tissue in some of the more
proximal sections (Fig. 13C). However, this Anterolateral Plug is
only visible in mid diaphyseal and proximal sections and is much
less distinct than in femora and tibiae.
Primary osteons are numerous and dense, but there are high
numbers of relatively smaller and longitudinal osteons with a
strongly birefringent single ring of lamellar infilling (Fig. 13D).
Such small primary osteons are absent in femora and tibiae, but
the relative amount of well-developed larger primary osteons as
well as their density is the same. The dominant type is again the
laminar organization (Fig. 13F, H). In some proximal sections,
convoluting radial canals can be found, which often extend
throughout the whole thickness of the cortex (Fig. 13G).
Annuli/LAGs are more abundant than in femora and tibiae,
but their distribution is still very inconsistent (Fig. 13EF, H).
Secondary osteons are very rare. They are mainly located at the
edge of the CCCB in the most distal or proximal sections, but they
mainly occur close to the internal margin of the anterolateral
corner along the edge of the short endosteal layer (Fig. 13D) or
within the Anterolateral Plug.
Ontogenetic Stages in Humeri
The differentiation of humeral cross sections into ontogenetic
stages is much more ambiguous than in the femora and tibiae. The
only clear features are the size and the number of growth cycles.
axm llrrttaaeeon irtaaeenodm ilrtaeeondm ilaed llrrttaaeeon ilrtaeeondm llrrttaaeeo ll-rrtt/aaeeon llrrttaaaeeon irscaoonopm irtssceeeonw
TBW .a31 ./444 .a44 .84m .a83 .a83 a4n .a84 .017 _ teh sso
edM icav 4 ./527 .89 .87 .28 .36 .87 ./_19 21 .361 itsem fseo
Stage 1 or Embryonic Stage. As in the other sectioned
elements, this stage is not preserved.
Stage 2 or Juvenile Stage. The smallest sections with not
more than a single growth cycle belong to this stage (Tab. 3;
Fig. 13AB, D, G). The slow growing part (zone, annulus, or LAG)
exists close to or at the outer rim of the bone wall. The degree of
organization of the vascular canals is low, so that plexiform to
sometimes reticular tissue type predominates.
Stage 3 or Post-Juvenile Stage. All remaining cross sections
belong to this stage and a further subdivision is not possible. The
number of growth cycles exceeds one and the laminar vascular
pattern predominates (Tab. 3; Fig. 13C, EF, H).
Bone Histology of the Fibula of Dysalotosaurus
Due to the scarcity of preservation of fibulae, cross sections
could only be produced from levels very close to or within their
proximal metaphysis. Therefore, periosteal compact bone is, if at
all, often present as a thin layer surrounding parts of the bone wall
externally and it was impossible to get a truthful count of growth
The overall shape of the cross sections is oval to kidney-like with
very thick and strongly curved bone walls anteriorly and
posteriorly. Most of the outer rim of the marrow cavity is poorly
defined because of wide cavernous spaces surrounded by a loose
network of trabeculae. An endosteal layer can only be observed
along the thinner lateral and medial walls. This band of lamellar
bone is very thick posteromedially (Fig. 14 AD).
The thin layer of periosteal primary compact bone consists of
fibrolamellar bone tissue, although the primary osteons are often
relatively small and scattered. Endochondral bone tissue is often
developed between this peripheral fibrolamellar bone and the
The medial wall differs strongly from the other units, because it
is heavily altered by dense Sharpeys fibers, so that the area is
strongly birefringent under polarized light (Fig. 14EF). The bone
matrix seems to be completely metaplastic in origin and the
vascular canals are simple, elongated, and oriented parallel to the
Secondary osteons are very common in these metaphyseal cross
sections. The CCCB is not involved, but its external border and
most of the endochondral tissue is strongly remodeled. Internal
and mid cortical areas of the posterior corner may even consist of
dense haversian tissue of at least two generations of secondary
osteons (Fig. 14 GH). The medial wall is affected by very coarse
remodeling (Fig. 14EF), because the scattered secondary osteons
are much larger.
In the cross section of the large fibula GPIT/RE/5109, possible
medullary bone is preserved internal to a part of the endosteal
layer that fans out (Fig. 14AD). The medullary bone tissue also
differs from the thick layer of CCCB external to the endosteal
layer by the lack of birefringent lamellar bone typical for the latter,
by the complete lack of osteonal development, and by a much
higher density of osteocyte lacunae within its reticular network.
Bone Histology of the Prepubic Process of the Pubis of
The sections cut directly at the maximum lateromedial width of
the prepubic process have a wide oval shape (Fig. 15A). Sections
taken more distally/anteriorly to the maximum width of the
prepubic process have a triangular to lamp shade-like external
outline (Fig. 15B).
The periosteal compact bone wall is very thin compared to the
overall diameter of the cross sections. There is no consistent
internal margin, because a single large marrow cavity is absent.
However, some of the internal cavities are quite large. These
cavities are always of resorptive origin, because remnants of
periosteal compact bone are often still preserved in some of the
thicker trabeculae (Fig. 15CD).
This tissue consists of vascular fibrolamellar bone.
Welldeveloped primary osteons are mainly visible in the dorsal and
medial parts of the bone wall, but they are not very dense and
mostly longitudinal in arrangement (Fig. 15E). Mainly ventrally,
primary osteons are rare, relatively small, and weakly developed.
Here, the matrix is often almost opaque and the often simple
vascular canals are also longitudinally organized (Fig. 15F).
Growth cycles are very rare, but there are at least one to two
annuli/LAGs preserved in some sections.
Numerous small secondary osteons occur in the trabeculae
between the large pseudo-cavities as well as in the internal areas of
the periosteal compact bone wall, where they sometimes form
haversian tissue. Their abundance decreases towards the medial
The combination and correlation of the fractional values of the
growth cycles for each group of cross sections resulted in a quite
consistent number of years represented by these cycles. Thus, the
combined growth cycles in femur group one (sections from the top
of the proximodistal shelf close to the middle of the shaft) represent
11 years, those of femur group two (sections from the base of the
fourth trochanter) represent 12 years (Fig. 16), and those of tibia
group one (sections well within the distal third of the shaft)
represent 11 years (Fig. 17). The only group for humeri represents
ten years recorded by all combined growth cycles (Fig. 18),
although several cycles were probably not recognized (compare
with Tab. 3). The remaining groups three and four in femora as
well as group two in tibiae contain only three to four cross sections
without enough preserved growth cycles for a secure correlation.
The Mark of Initial Sexual Maturity (MISM) in femora always
correlates with an age of approximately 9.5 years in femur group
one and 10.5 years in femur group two (Figs. 16; 19).
To calculate the respective body masses for the correlated
growth cycles with the Developmental Mass Extrapolation method
, and to calculate the sigmoidal growth curves, it was necessary
to calculate the maximum body mass. The largest femoral
specimen (MB.R.2144) represents a body mass of 115.3 kg using
the method of Anderson et al.  for bipeds. In the same way,
the respective body mass at the MISM was calculated as 32.44 kg
on average for femur group one and 31.96 kg for femur group
By using the first nine (femur group one) to ten (femur group
two) secured growth cycle values, the respective values of the
MISM, and the maximum body mass, four sigmoidal growth
curves were created. The remaining growth cycle values,
representing unsecured growth cycles external to the MISM, were
plotted into the curves subsequently (Fig. 19). The manual shift of
these values by one year on average resulted in the ideal fit to their
respective growth curves. At the end, a total of 13 years of life of
Dysalotosaurus are represented by the observed and correlated
growth cycles in the femoral cross sections of groups one and two
The now known values of the four parameters of each of the
four growth curves were used to calculate the respective values for
all known femora of Dysalotosaurus. The largest sampled femur
(SMNS F2, group two) would therefore represent an age of 16.5
years (body mass after ) or 16.3 years (body mass after ).
The age of the third largest femur found in the collections
(R12277) would then represent an age of 19.7 years (after ) or
19.3 years (after ) (Fig. 20).
The MISM is located well between the lower and middle third
of the growth curves, if body mass is plotted versus age (Fig. 21).
Thus, the growth rate of body mass is still accelerating after this
mark and reaches its maximum in the 14th year with a daily
increase of 24 to 26 grams (for femur group two). However, by
plotting the respective values of the distal mediolateral width of
femora or their midshaft circumference (representing body size)
versus age, the MISM is then located very close to the inflection
point of the curve (between accelerating and decelerating growth
rate: Fig. 21). Finally, the relative body size of Dysalotosaurus at the
MISM reaches 62.1% for the femoral distal mediolateral width
and 63.4% for the midshaft circumference when compared to the
known maximum body size.
Variation within Bone Tissues in Dysalotosaurus
Variation within bone tissues in Dysalotosaurus is exhibited
between different individuals, within the ontogenetic series, within
a skeleton, within a bone, and even within a cross section. This
variation also clearly demonstrates that comparative bone
histology is only significant when the sampling is standardized
among several skeletal elements and the relative ontogenetic stage
is considered (e.g. [6,36,38]).
Variation between Different Skeletal Elements. The
bone wall of the main weight bearing long bones (femora, tibiae)
of Dysalotosaurus are naturally thicker than in the sampled humeri,
fibulae, and prepubic processes. Interestingly, the relative growth
rate is also higher in these long bones compared to the other
sampled elements, which is inferred from the overall development,
density, and organizational degree of vascular canals (see e.g.
[8,36,39,41,46]). Femora and tibiae possess a comparatively
higher amount of well developed primary osteons and larger
areas with plexiform or even reticular vascularization than humeri
and prepubic processes. Thus, as in Maiasaura  and Plateosaurus
[32,33], different skeletal elements grow at different rates during
A possible explanation for growth rate changes may be the
absolute size of the respective element within the skeleton
combined with the degree of utilization, which includes two
components: (1) the degree the element participates in weight
bearing and (2) the functional demand on the bone. In the case of
the biped Dysalotosaurus, the femur and tibia are the largest and
Figure 16. Fractional growth cycle values of femur group two are correlated to age. MISM = Mark of Initial Sexual Maturity. F is the
abbreviation for femur. Each of the following numbers corresponds to the respective specimens in Tab. 1. Some specimens were sampled at least
twice so that additional letters (a, b) advert to the respective section used for this correlation.
Figure 17. Fractional growth cycle values of tibia group one are correlated to age. T is the abbreviation for tibia. Each of the following
numbers corresponds to the respective specimens in Tab. 2. Some specimens were sampled at least twice so that additional letters (b) advert to the
respective section used for this correlation.
Figure 18. Fractional growth cycle values of the single group of humeri are correlated to age. H is the abbreviation for humerus. Each
of the following numbers corresponds to the respective specimens in Tab. 3.
primary weight-bearing bones intensively used for locomotion.
The humerus is comparatively much smaller (in the only preserved
individual dy I, exhibited in Berlin, app. 57% the length of its
femur) and was likely not used in weight bearing or locomotion. It
is therefore not surprising to find it less densely vascularized within
a relatively thinner bone wall. The sampled prepubic process is
even more different than the femur and tibia in these characters,
because it serves only as muscle attachment site and is not involved
in active movements or in bearing weight.
Similar tendencies are visible in other tetrapods, but it strongly
depends on their respective skeletal bauplan. The humerus of the
therapsid Diictodon reached higher relative growth rates than its
femur , because it was probably used for digging in addition to
weight bearing. This is also observed more extensively in the
common mole (Talpa europea) by Enlow & Brown , where the
large humerus is well vascularized and the much thinner cortex of
the smaller tibia is almost avascular indicating the tibia had a
much slower relative growth. It is not as simple in birds and
pterosaurs, because the demand on active forelimbs, mainly for
flying, against weight bearing hindlimbs is highly speculative.
However, there are at least indications that the absolute size of
bones in pterosaurs , in penguins , and in some dinosaurs
(see e.g. [32,36]) is correlated with relative growth rate in these
groups as well. Although there are no subsumable differences in
the vascularization pattern between elements in recent ratite
skeletons, the flightless habit almost predicts much lower growth
rates for the forelimb elements compared to the elements of the
hindlimb . This is also comparable to biped dinosaurs, such as
Allosaurus (see e.g. ) and Dysalotosaurus, or facultative quadruped
dinosaurs with a strong size difference between fore- and
hindlimbs, such as Scutellosaurus .
Within a single limb, the bones of the stylopodium (humerus,
femur) have higher relative growth rates than the bones of the
zeugo- and autopodium, because the latter are often smaller in
overall size and share functions, such as weight bearing or muscle
activity, among each other. The absolute forces acting on each of
them are therefore smaller than in the stylopodium. This is the
case for the less vascularized radii and ulnae compared to the
humeri and femora in Thrinaxodon  and to the femora in
Scylacops , and for the ulnae of Allosaurus and Tenontosaurus
compared to the other sampled bones of the respective studies
[34,68]. Nevertheless, whenever bones of the zeugo- and
autopodium are fused (e.g. to the tibiotarsus and tarsometatarsus
in birds), are much more prominent than their neighbors (e.g. the
tibiae in many dinosaurs), or are exclusively used for powerful
movements (e.g. the wing phalanges of pterosaurs), their relative
growth rates should be more comparable to the bones of the
stylopodium (see , Dysalotosaurus , respectively). In all these
cases, the fused bones are also larger than usual.
In the end, the relative size of a bone in a skeleton reveals its
importance in weight bearing and/or movement and its relative
growth rate compared to other elements is therefore predictable to
a certain degree.
Variation between Different Cross Sectional
Units. Cross sections with very consistent outlines (especially
distal and mid diaphyseal humeri; Fig. 1) reveal much less
variation of bone tissues than cross sections with irregular outlines
and acute corners, such as femoral sections (Figs. 16), distal tibial
sections (Figs. 1GH; 9; 12), and prepubic sections (Fig. 15). Some
of the intrasectional variation is caused by differences in bone wall
thickness. The thicker posteromedial and posterolateral corners in
femora and the anteromedial corner and medial bend in tibiae
have a high density of weakly-organized primary osteons (e.g.
Figs. 3AB; 9C) and osteocyte lacunae. The collagen fibrils in
these areas are also hardly organized so that there is only a weak
birefringence under polarized light. Finally, the slow growing
zones are weaker and the distances between them are larger than
in the thinner bone wall units (see below; Fig. 1H, L). The opposite
trend of the noted features takes place in the latter (in the anterior
corner of femora and in the anterolateral corner of tibiae)
(Figs. 2EF; 3EF; 8C; 9AB, E). A similar pattern can be seen in
the largest sampled femur of Dryosaurus altus .
The variation in relative growth rates due to variable bone wall
thickness is superimposed by another source of variation in
femora, distal tibiae, and proximal humeri. The anterior corner in
distal femora, the anterolateral corner in distal tibiae, and
sometimes the anterolateral corner in proximal humeri, consist
of an internal wedge of CCCB (femora, tibiae) or of endosteal
lamellae (mainly humeri). The external periosteal regions possess
here well organized primary osteons in a low density, osteocyte
lacunae are also rarer than in other units, and the collagen fibrils
are mainly transversely organized (Fig. 2EF; 3EF; 13C). All
growth cycles (including annuli/LAGs) are closer together (Fig. 1H,
L). The bone wall of the opposite side of the cross sections
(posterior bend in femora, medial sides in distal tibiae and
proximal humeri) is distinctly resorbed internally by the marrow
cavity (Figs. 1B, D, H, L; 2AB; 5E) and is more similar to thick
bone wall units (Figs. 1B, D, H, L, 2AD; 3AB; 5CD, FG; 9C;
10CF; 13G). Thus, the latter units were deposited by much
higher relative growth rates than the former units.
These differences in growth rate of opposing cross sectional
units are explained by the drift of the marrow cavity towards the
side with the suggested higher relative growth rate. The
combination with the bending orientation of the respective long
axes of the bone shafts indicates that the marrow cavity always
drifts from the convex side of the long axis to the concave side to
maintain the overall bone wall thickness during growth. The
convex side of the long axis is located anteriorly in femora and
laterally in distal tibiae and proximal humeri, respectively. This
also explains why there is still unresorbed CCCB left in the
mentioned units of relative slow growth, because this metaphyseal
tissue is necessary for a consistent bone wall thickness during
ontogeny . For the same reason, juvenile bone tissue (small
longitudinal primary osteons, knitting pattern of the matrix) is still
preserved in the internal areas even in respective units of large
cross sections (Figs. 3GH; 6EH; 8C; 9AB, EG). The typical
intrasectional variation caused by osseous drift is well described in
Enlow  for rats and monkeys and is also shown for Varanus (see
figure 2E in ) and for the small lizard Gallotia (see figure 13 in
). In contrast, this typical variation is rarely described in fossil
tetrapods, although it is documented in the multituberculate
mammal Nemegtbataar (see figures 6 and 7 in  and indicated in
the dinosaurs Scutellosaurus (see figure 2 in  and Psittacosaurus
). As a result, cortical drift is supposed to be the normal case in
long bones with a bent long axis [8,71] and should be considered
before histological sampling, due to its strong influence on the
microstructure and on estimating growth rates.
The described special bone tissue of the Posterolateral Plug in
femora (Figs. 4AD; 5CE), of the anterolateral corner in tibiae
and humeri (Figs. 9E, H; 13C), of the medial wall in fibulae
(Fig. 14EF), and of the lateroventral corner in prepubic processes
(Fig. 15AB), are suggested to be the result of muscle and/or
tendon forces acting on these cross sectional units. This is
indicated by the relationship of these special structures with
external processes or attachment sites for muscles. The tissue
structures also display the potential orientations of the acting
muscle forces, because Sharpeys fibers are most abundant in these
units and the vascular canals are often oriented in a dominant
direction. These Plugs are also very restricted with sharp borders
(Figs. 4AB; 5E; 9E; 15B) and show more secondary remodeling.
Scattered secondary osteons are sometimes even developed close
to the external surface, which is very unusual for the normal bone
tissue in Dysalotosaurus independently of ontogenetic stage.
Such unusual restricted areas in cross sections are already
mentioned for the femur in Hypsilophodon and described for the
femur in Iguanodon . There were also sharply delimited and
more strongly remodeled areas (also visible in Hypacrosaurus ) in
possible connection with muscle attachment sites. As in
Dysalotosaurus, these special areas can also be sharply restricted to a certain
level in the shaft and vanish over a short distance within the shafts
long axis. Possible Plug-like structures are mostly known in the
literature as local areas with unusually intensive secondary
remodeling almost reaching the external surface (e.g.
[30,34,36,38,52]). Horner et al.  already noted the possibility
of muscle strain as a reason for these above-average remodeled
areas, which was also pointed out by Currey .
Variation of Growth Cycles. The number, relative
distances, and developmental degree of growth cycles are highly
variable in Dysalotosaurus. Their number is naturally strongly
influenced by ontogeny (the larger/older the more) and by the
primary bone wall thickness. This can be seen between different
elements of the skeleton. The thickest primary bone walls are
developed in femora and tibiae with 12.5 and 11 mm, respectively.
These elements preserve the highest number of growth cycles,
which counts up to nine in the largest sections alone and up to 12
after ontogenetic correlations in all sections. Humeri, which have a
maximum primary bone wall thickness of 5.3 mm in the samples,
have only up to five cycles in a single section and up to ten after
the correlation. The much thinner primary bone wall in the
prepubic process can preserve only two cycles at maximum. The
relative distances between growth cycles are also dependent on the
cutting level within the shaft, because the average thickness of the
periosteal bone wall is increasing towards the mid diaphysis and
the portion of CCCB at the total bone wall thickness is here
insignificant [8,71]. The resulting differences in the course of
calculated growth curves derived from these distances are even
stronger between cutting levels than between methods for
calculating body mass (Fig. 20).
In contrast to the results of Chinsamy , there are indeed
annuli/LAGs preserved in Dysalotosaurus, but they are rather rare,
especially in femora (Tab. 1). They are slightly more abundant in
tibiae and prepubic processes and most abundant in humeri
(Tabs. 2; 3). There is also no distinct pattern predicting the
occurrence of annuli/LAGs, because a medium-sized femur can
possess a LAG and a large femur none at all (Fig. 1B, D). In tibiae
and humeri, the number of LAGs increases with increasing bone
wall thickness, but this is the same pattern as for all growth cycles,
and LAGs are only part of them (see e.g. Fig. 13EF).
Interestingly, some of the prepubic processes, with their extremely
thin primary bone wall, possess more annuli/LAGs than the
Thus, the development of annuli/LAGs in the sampled skeletal
elements of Dysalotosaurus seems to be dependent on several factors,
where relative growth rate (the lower the more) might be
dominating over bone wall thickness. In elements with relatively
high growth rates (femora, tibiae), only unfavorable environmental
conditions (e.g. long draughts) or dramatic events in the
individuals life history (e.g. injury, disease) may have resulted in
the rare development of annuli/LAGs.
The relatively random and rare formation of clearly defined
annuli/LAGs is in striking contrast to the pattern seen in many
other dinosaurs. In theropods (e.g. [25,28,75]), mainly primitive
and/or smaller sauropodomorphs (e.g. [29,32]), and some
ornithischians studied (e.g. [27,34,36,38]), annuli/LAGs occur
much more regularly and not as an exception, as in Dysalotosaurus.
Especially large and derived sauropods have much weaker cycles,
such as polished lines  or zonal differences in vascularization
[37,56,76], which are assumed to be annual markers as well.
None of these studies have mentioned the kind of growth cycles
found here. Their identity as possible annual markers is now,
however, unambiguously demonstrated. Despite the often
relatively weak appearance (Figs. 2AD; 5CH; 10EF; 13F), the
cyclic occurrence of fast and slow growing zones is striking. As for
annuli/LAGs: (1) their preserved number increases with related
body size and is quite constant (with a maximum deviation of 2)
throughout a single ontogenetic stage of a certain element (see also
Tabs. 13); (2) the thickness of the slow growing zones is relatively
constant, whereas the fast growing zones become thicker in the
thick bone wall units and thinner in the thin bone wall units; (3) the
zonation becomes weaker in thicker bone wall units and more
distinct in thinner units of a cross section; and (4) the plot of the
maximum growth rate with age, which is derived from the
correlated growth cycles under the assumption of their annual
signal, fits almost perfectly into the linear regression line of
maximum growth rates developed for dinosaurs (Fig. 22, see also
[7,77,78]). Cyclical fluctuations found in juvenile Maiasaura ,
in Hypacrosaurus , and in Coelophysis  are probably another
kind of growth cycles, but their significance as annual markers is
questioned by these authors and has still to be proved.
It is important to note that the type of growth cycles described
for Dysalotosaurus probably exists in a wider range of taxa, because
the cyclicity between zones of oriented collagen fibrils is also
mentioned in Alligator ( see figures 2JL; 3IK; 4 therein), and
is probably present in an extinct crurotarsian (pers. comm.
Bronowicz, 2009) and in Tenontosaurus (pers. comm. Werning,
2010). Thus, this kind of growth cycles will probably be found in
more tetrapods in the future and should provide age estimations
especially in taxa with an otherwise poor record of annuli/LAGs.
Correlation and Comparison of Ontogenetic Growth
Since all the sampled elements are isolated and microstructural
details vary between different elements of a skeleton, the
correlation of ontogenetic stages in femora, tibiae, and humeri
of Dysalotosaurus is only preliminary.
The second ontogenetic stage of all three elements (early
juvenile or juvenile stage; Figs. 8AD; 12) compares favorably,
because each section belongs to the smallest available specimens
and is located close to, or at, the left margin within the respective
size-frequency distribution (Fig. 23). Furthermore, primary osteons
are often incompletely developed (in humeri more advanced but
very small; Fig. 13B, D), there is not more than one completed
growth cycle, secondary osteons are extremely rare, and
histological differences between sectional units are weak (Fig. 8A
D; 12). This correlated juvenile stage is similar to large nestlings in
Maiasaura , to small juveniles in Orodromeus , and is located
in between the perinate and juvenile stages of Dryosaurus .
The correlation of the older stages is more difficult, because
there are different numbers of distinguishable stages in femora,
tibiae, and humeri. Femora in the third and fourth stage of
development (late juvenile (Figs. 3EF; 6F; 8EH) and sexually
immature stages (Figs. 1CD; 2EF; 5AB; 6GH) are correlated
with the third stage of tibiae (late juvenile to sexually immature
stage (Figs. 1EF; 10CF), and the post-juvenile stage of humeri
(Figs. 1IL; 13AB, D, G). Individual cross sections in humeri are
only assignable to either sexually immature or sexually mature
stages by their absolute size within the two-peaked size-frequency
distribution of the Dysalotosaurus herd (Tab. 3). The respective cross
sections of femora, tibiae, and humeri possess more than one
growth cycle (up to five in the fourth femoral stage), the vascular
pattern of vascular canals is mainly laminar to plexiform, primary
osteons are abundant and well developed, secondary osteons, Plug
structures, and osseous drift are present, and the cross sectional
units are well diversified (less prominent in humeri). The closest
similarities to described growth stages of other ornithopods were
found in the large juvenile and subadult stages of Orodromeus ,
the juvenile and smallest subadult stages of Dryosaurus , and the
juvenile stage in Maiasaura . Both the late juvenile stage and
sexually immature stage of Dysalotosaurus femora are also similar to
the subadult stage in Orodromeus and to the small subadult stage in
The last represented ontogenetic stage is considered here as the
sexually mature stage. This is clearly different to somatic maturity,
because none of the sampled specimens show an External
Fundamental System (EFS) as a sign for ceasing growth [6,7]. In
that sense, all sampled large specimens would represent
somatically subadult individuals. The differentiation to younger stages is
unambiguous in femora and tibiae, but only the absolute size and
the position within the two-peaked size-frequency distribution of
the Dysalotosaurus herd are helpful in humeri (Fig. 23; Tab. 3).
Shared features of the sexually mature stage are well diversified
cross sectional units with strong differences in bone wall thickness
(less distinct in humeri), numerous growth cycles (up to nine in
femora, seven in tibiae, five in humeri), often interrupted by
strongly developed Plug structures (Figs. 1B, H; 4AB; 5CE; 9E),
numerous and dense primary osteons, more abundant secondary
osteons (Figs. 2GH; 4CD; 9H), and highly advanced osseous
drift (Figs. 1B, H, L; 2AB; 5E; depends especially in humeri on
cutting level). This ontogenetic stage is comparable to the subadult
stage in Orodromeus and the medium-sized subadult femur of
Dryosaurus . It does not match the subadult stage in Maiasaura
due to the lack of extensive remodeling in the deep cortex and the
lack of a starting EFS .
The ontogeny of the bone histology in Dysalotosaurus is most
similar to Dryosaurus  regarding the overall size of skeletal
elements as well as the respective cross sectional dimensions,
vascularization pattern, and degree of secondary remodeling.
Orodromeus, on the other hand, reveals a vascularization pattern,
which is usually found in skeletal elements of Dysalotosaurus with
relatively lower growth rates, such as humeri or prepubic processes
(Figs. 13; 15EF). There, mainly longitudinal and smaller primary
osteons are common, which are well described for Orodromeus
[43,52]. LAGs are also more common as in Dysalotosaurus and a
possible EFS is known, which indicates nearly cessation of growth
in the somatically mature adults. It confirms that this ornithopod,
which has reached a smaller maximum body size than
Dysalotosaurus, grew with a lower overall growth rate than the
latter genus (other examples are e.g. [41,43,77,80]).
The opposite case is the much larger hadrosaur Maiasaura. The
vascularization pattern is not very different, but the thicker
primary bone walls experienced more intensive secondary
remodeling. Large and widespread resorption cavities or dense
Haversian bone, which can obscure the primary bone in the
deeper cortex, is completely unknown in the sampled elements of
Dysalotosaurus. The intensity of secondary remodeling is therefore
probably not only an indicator of individual age and longevity (e.g.
[57,81]), but also an indicator of maximum body size . This is
probably the case in primates (compare e.g. Castanet et al. 
and Burr  with Mulhern & Ubelaker , see also Singh et al.
), ornithopods (see above), and sauropodomorphs (compare
e.g. Klein  with Klein & Sander ). The comparison of the
largest sampled femur of Dysalotosaurus (33cm calculated length)
with the largest femur of Dryosaurus (49cm length; see ), which
shows much more extensive secondary remodeling, either
confirms this assumption, or the latter was indeed individually
older than the former . This femur is even larger than the
largest preserved, Dysalotosaurus femur, which has a calculated
length of 38 cm. Together with the observations of increasing
secondary remodeling within the ontogenetic stages of
Dysalotosaurus, the influence of individual age on remodeling intensity is
probably most important, but maximum body size might be an
Finally, Horner et al.  noted that the largest Dryosaurus
femur was still actively growing, because it lacks an EFS and
therefore belonged to a somatically subadult individual. If this is
true, then even the largest known individuals of Dysalotosaurus were
still somatically subadults.
The Life History of Dysalotosaurus
The embryonic or perinatal ontogenetic stage is not preserved
in Dysalotosaurus, but the longitudinal section of the smallest known
femur (Fig. 7DF) belonging to the early juvenile stage is very
distinctive regarding possible behavior of hatchlings. This stage is
very similar to the structures described for younger stages of
Orodromeus and Troodon , although the pads of calcified cartilage
reach naturally much deeper at this early ontogenetic stage than in
the sample of Dysalotosaurus. It is also in strong contrast to the
situation seen in some hadrosaurs  where pads of calcified
cartilage are not constricted to the preserved epiphysis, but reach
through the whole metaphysis into the diaphysis. Endochondral
bone is here much rarer and apparently lacks transverse struts
crossing the long tubular structures, which consist of connected
cartilage canals and marrow processes. In the large nestling of
Maiasaura , thin coatings of endochondral bone are developed
along the wall of the marrow processes, but noticeable transverse
struts were only observed deeper within the metaphysis. Since
large nestlings of Maiasaura are here tentatively correlated with the
early juvenile stage of femora in Dysalotosaurus, the degree of
epiphyseal ossification in Dysalotosaurus at this stage was strongly
different from Maiasaura and other hadrosaurs, but similar to
Orodromeus and Troodon, which would implicate precociality in
Dysalotosaurus hatchlings [51,87]. Thus, they could follow their
parents short after hatching, but experienced rather moderate
growth rates compared to the probably semi- to fully altricial
hadrosaurs . By the way, the precocial behavior is also
assumed for the closest relative of Dysalotosaurus, Dryosaurus altus,
whereas an embryo of the larger taxon Camptosaurus was probably
altricial similar to Maiasaura .
Moderate growth rates are visible in the four growth curves of
Dysalotosaurus (Fig. 20). The early and late juvenile stages of the
femur cover the moderately sloping part of the growth curves up
to approximately six years of age. The sexually immature stage
correlates with the age of six up to ten years. The latter date
most likely marks the initiation of sexual maturity and therefore
separates the sexually immature members of the Dysalotosaurus
herd from the sexually mature individuals. This hypothesis was
derived from five out of the six sampled large femora belonging
to the most mature histological ontogenetic stage observed (see
above). A mark or transition (MISM) is visible in these cross
sections (Figs. 1B; 4A; 5EF; 6AE). This demarcation shows an
overall slow-down of bone apposition rates (the usual fast
growing zones are weak or absent), which interestingly starts in
each of the five concerning femora at almost the same relative
position within the cross sections (Figs. 6E; 16; 19; 20). Thus, this
mark represents not an individual event, but a real physiological
signal indicating an important change in the life history of
The achievement of sexual maturity is the most likely
explanation supported by several reasons: (1) This event is
commonly combined by a slow-down of growth rate in many
other tetrapods (e.g. [7,57,62,89,90]); (2) The timing of sexual
maturity occurs well before somatic maturity as in other dinosaurs
(e.g. [32,57,62,75,91]); (3) This event plots in diagrams with body
size versus age almost exactly at the curves point of inflection (
but see below); (4) The preservation of medullary bone tissue in a
large fibula and a large tibia (Figs. 11; 14AD), which belong to
the group of large individuals in the size-frequency distribution
(Fig. 23; Tab. 2); and (5) By correlating the respective value of this
mark with femoral size, the mark plots well within the gap between
the dominating groups of small and large individuals of the
Dysalotosaurus herd (Fig. 23).
This gap shows the underrepresentation of individuals and is
probably the result of banishment and/or increased mortality of
this size class. In recent and at least temporarily gregarious
ungulate mammals, young males predominantly suffer increased
mortality around the time of sexual maturity, because they are
driven out of the herd very early by prime-aged males (e.g. Impala
) or they leave on their own (e.g. Kudu ). They are then
vulnerable to predators and have higher stress levels due to their
low rank within bachelor herds. In other species, young males
suffer high mortality during their first rut (e.g. bighorn sheep 
and rhinos ). Young females also suffer increased mortality due
to inexperience in reproduction, high reproduction costs, and
competition with prime-aged females (e.g. red deer ). Higher
mortality rates resulting from early sexual maturity were also
suggested for the tyrannosaur Albertosaurus . Thus, the position
of the mark right within the gap of the size-frequency distribution
(Fig. 23) supports the assumption that it is indeed the Mark of
Initial Sexual Maturity (MISM).
The decrease in bone apposition rate observed in the cross
sections at the MISM apparently conflicts with its relative position
within the growth curves (body mass versus age; Fig. 20), because
it is located here within the lower third of the exponential growth
phase and growth rate is still accelerating. This is similar to other
dinosaur taxa, where the time of sexual maturity is strongly
indicated by the occurrence of medullary bone  and/or
increased midlife mortality . The time of sexual maturity for
Tenontosaurus (8 years) and Allosaurus (10 years) is located, as in
Dysalotosaurus, within the lower third of the exponential growth
phase and not at the curves point of inflection, where growth rate
reaches its maximum . In the case of Tyrannosaurus, the
estimate of 18 years is close to the inflection point, which is similar
to Albertosaurus (compare  with ), although the exact time of
sexual maturity is probably an upper bound for Tyrannosaurus .
It is suggested that the phenomenon of contradicting features in
Dysalotosaurus is an effect of allometric scaling between increasing
body mass and increasing body size (including bone apposition),
where the ratio would be 8:1 (compare also Box 3a with 3b in ).
Furthermore, the scaling effect of body mass is neutralized by
plotting a variable representing body size versus age (Fig. 21),
where the time of sexual maturity in Dysalotosaurus is indeed located
almost exactly at the curves point of inflection.
It should also be noted that the MISM is completely absent in
all large tibiae and humeri of respective position within the
sizefrequency distributions. This indicates an only moderate slow
down of bone apposition rate, which is probably not visible in
elements of slightly lower relative growth rates compared to the
rates in femora. Finally, the relative body size at time of sexual
maturity compared to maximum known body size in Dysalotosaurus
is approximately 62 to 64%, which is strikingly similar to the
remarked 60% to the recorded maximum size known in
Albertosaurus  and close to the estimated value of 70% in
Barosaurus . Thus, the apparent contradiction between
decelerating bone apposition and accelerating body mass in
Dysalotosaurus in young sexually mature individuals is treated here
as rather insignificant.
The location of the largest sampled femur (SMNS F2 group
two) within the growth curves is well below the estimated
asymptote at approximately 16.4 years of age (Fig. 20). Additional
features of still active growth are the open vascular canals at the
periphery, well vascularized tissue in the external bone wall areas,
and the complete absence of an EFS. The third largest known
femur (R12277) is also located below the asymptotic level of the
growth curves indicating that this individual has also not reached
somatic maturity. The subsequent sampling of the largest known
femur (MB.R.Ig374; similar to the specimen used to calculate
maximum body size and mass (MB.R.2144)) also revealed still
active growth. The absence of EFS in a much larger femur of the
closely related taxon Dryosaurus altus  suggests that this species
obviously grew to larger body sizes than Dysalotosaurus and that
both taxa most likely experienced indeterminate growth as
Chinsamy  already suggested.
Many of the Dysalotosaurus individuals could be reproductively
active for more than five years, but none of them obviously
reached somatic maturity. Dysalotosaurus was highly vulnerable to
most of the contemporaneous predators due to its relative small
body size and the lack of any defensive structures (as in
Kentrosaurus). This may be a reason, why sexual maturity was
delayed until the ninth year of life. The cost of reproduction was
too high for small individuals due to high vulnerability to
Another factor for the high mortality rate around time of sexual
maturity and, especially, the prolonged exponential growth phase
in sexually mature individuals might be intraspecific competition
within a herd. Larger/stronger individuals surely had a more
dominant role within the herd and a better chance for
reproduction than smaller/weaker individuals. Fast and extended
indeterminate growth could therefore be regarded as a survival
advantage for Dysalotosaurus.
Implications for the Growth Pattern in Other
Like in some other small ornithopods and many sauropods
[18,37,48,52,55,56,57,76,80,97] Dysalotosaurus exhibits a growth
pattern, where annuli/LAGs as representatives of a zonal bone
tissue are rather scarce, completely absent, or are replaced by less
obvious growth cycles. On the other hand, large ornithopods,
other ornithischians, prosauropods, and all theropods more
derived than Herrerasaurus (see [6,18]) show a relative consistent
growth pattern with annuli/LAGs representing the usual kind of
growth cycles (e.g. [25,27,28,32,35,36,38,68,75]).
Klevezal  has found a relationship between the abundance
and uniformity of annuli/LAGs and environmental conditions in
recent mammalian populations, which could partially explain the
sorting of dinosaurs into such multiform groups. Populations
inhabiting regions with strong seasonality consist mainly of
individuals with distinct and weakly variable annuli/LAGs in
their recording structures (e.g. bone microstructure), which is
mostly a two-phase annual rhythm. In contrast, populations of the
same species, inhabiting regions with moderate conditions, exhibit
mostly weakly-developed annuli/LAGs and a higher variability in
number (poly-phase annual rhythm). However, exceptions always
occur. So, although it is likely that a single fossil specimen
represents the usual growth pattern of its population, it is also
possible that it represents the anomalous minority. An unusual
growth pattern found in a single specimen should therefore be
treated with caution (see [48,52,98]).
The regular development of annuli/LAGs in highly seasonal
regions is advantageous compared to irregular cyclicity, because
the former is synchronized to the seasonal changes of
environmental conditions. Irregular or asynchronous growth is
disadvantageous in strongly seasonal regions, because growth phases
reaching into harsh times cost naturally more energy than arrested
growth. Poly-phase growing individuals have therefore to fit their
growth regime to the seasonal conditions or die. In less seasonal
regions, it does not matter, which growth regime an individual
possess, because the effects on its energy balance is not so
disadvantageous and the variability of growth patterns in the
population is therefore much higher .
The results for Dysalotosaurus have shown that the abundance
and development of annuli/LAGs depends either on relative
growth rate (annuli/LAGs in faster growing femora are less
abundant) and environmental conditions (by far not all growth
cycles are completed by an annulus/LAG). For the Tendaguru
region with its reconstructed seasonal change of humidity ,
long droughts would be such harsh times accompanied by a
shortage of food and water. This is also indicated by the
depositional area of the Tendaguru Beds, which are very unlikely
to be the usual habitat for the preserved dinosaurs .
LAGs are obviously more common in ornithopods than
previously thought  and completely azonal bone is rather
unlikely (in contrast to e.g. [6,18,48]). LAGs occur in Orodromeus,
Dysalotosaurus, Tenontosaurus, and Maiasaura at first in the late
juvenile stage (this study and [34,36,52]). In Dysalotosaurus, LAGs
are very rare and close to the periphery at this stage (except in
humeri). The first LAG in Tenontosaurus is also not consistently
developed in all specimens and is sometimes substituted by a
band of differing oriented collagen fibrils . In Dryosaurus altus,
LAGs were found in all three subadult femora, but at
nonoverlapping relative positions indicating at least three different
growth cycles for the two smaller specimens and up to 15, if one
includes the largest femur and calculates the number of LAGs by
back counting . If Dryosaurus is indeed similar to Dysalotosaurus
in its growth pattern, which is implicated by a similar
vascularization pattern and the absence of an EFS, then the
number of developed LAGs would be still rare in the large
femora of Dryosaurus. In Dysalotosaurus, ten out of 14 femora
(excluding the juvenile stages) bear one (in one case two) LAG or
annulus (Fig. 5AB), respectively (in Tab. 1 six out of nine,
excluding the femora not usable for the age calculations), but
these annuli/LAGs represent at least three to four
nonoverlapping positions, which confirms a very inconsistent and
highly variable growth pattern. It is therefore possible that
Chinsamy  sampled specimens, where LAGs are not
developed among the other growth cycles.
Orodromeus differs from both Dysalotosaurus and Dryosaurus by its
lower overall growth rate (see above) and the presence of an EFS
in the largest individuals . Another difference is the quiet
consistent development of LAGs in the tibiae and femora of
subadult and adult individuals. This could be the consequence of
overall lower growth rates in Orodromeus . The development of
LAGs is more likely, because the seasonal slow-down in growth
starts from an already lower level than in Dysalotosaurus and
Dryosaurus. However, Orodromeus seems to be rather an exception
among small to medium-sized ornithopods regarding its growth
pattern, although LAGs and annuli were recently also found in
small ornithopods from high latitudes .
The age of Orodromeus at the beginning of somatic maturity is
estimated by Horner et al.  at five to six years. This is relatively
short for a dinosaur of this size, because other small dinosaur taxa
reached ages of at least nine and eight to 18 years, respectively
[27,75]. Scheetz  described four additional bands of highly
birefringent bone tissue alternating with weakly birefringent
darker bands in a juvenile femur of Orodromeus (see also figure
2C in ). At a first glance, it has some similarities to the
alternation of fast and slow growing zones in Dysalotosaurus,
although such a suggestion should be treated with caution. If these
bands are indeed annual cycles, than the age of Orodromeus would
be about ten years at time of reaching somatic maturity. This
would fit much better to the estimated ages of other small
The three larger ornithopods Tenontosaurus, Maiasaura, and
Hypacrosaurus developed much higher numbers of LAGs in the
subadult and adult stages than Dysalotosaurus and Dryosaurus before
reaching somatic maturity [34,36,38]. They experienced very high
growth rates during the juvenile stages (e.g. ), as the growth
curve of Tenontosaurus also shows in comparison to the averaged
growth curve of Dysalotosaurus (Fig. 24). Thus, all three large
ornithopods had higher initial and juvenile growth rates and
reached their asymptotic growth plateau relatively earlier than
Dysalotosaurus and most of the other small ornithopods.
By using the mentioned relationship between strength of
seasonality of environmental factors and occurrence and
uniformity of annuli/LAGs , the abundance of numerous annuli/
LAGs in subadults and adults of larger ornithopod taxa would
indicate higher seasonal stress than in the smaller Dysalotosaurus
and Dryosaurus. Another example is the absence of annuli/LAGs in
the small Proctor Lake ornithopod compared to their occurrence
in a large hadrosaur of the same locality . The zonation in just
a single femur of Gasparinisaura (assuming that the others lack it
) probably represents similar intraspecific variation of cyclical
growth patterns than in Dysalotosaurus, although LAGs are even
Thus, many small ornithopods had probably less seasonal
environmental stress than large ornithopods and different growth
patterns had existed in large and small taxa, respectively. Two
reasons are proposed for these differences:
(1) Food demands and migration: Small ornithopods were
predominantly selective low-browsers  and probably not able for
supra-regional migration . They needed less absolute
amounts of food than large ornithopods, which would also have
a weaker effect on their growth rates during dry (or cold) seasons
than in large taxa. The ability to alternative nutrition, such as
insectivory [20,53], might also have played a role. Large
ornithopods cleared their local habitat of food much faster than
small ornithopods, not only, because of their higher absolute food
demands, but also due to their much more effective chewing
ability (e.g. [101,102]) and their assumed gregarious behavior (e.g.
[101,103,104,105]). For many of them, migration was therefore
essential to survive and this meant additional seasonal stress.
Furthermore, some small ornithopods were probably able to
endure bad times by specialized adaptations, such as the fossorial
Oryctodromeus (, see also ), to which larger ornithopods
were unable to do so . However, the recent discovery of
annuli/LAGs in the small polar ornithopods from southern
Australia ( in contrast to ) demonstrates that even low
seasonal liability did not prevent them from the severe polar
winters of their habitat so that they had to stop growth for saving
vital energy during the dark season.
In conclusion, higher food demands and seasonal migration of
large ornithopods could be one reason for the much more
consistent development of annuli/LAGs in their long bones
compared to small ornithopods. Exceptions may be the
ornithopods Telmatosaurus and Zalmoxes, which are treated as secondarily
downsized taxa due to their restricted island habitat .
(2) Breeding strategy and courtship/rut: Dysalotosaurus, Orodromeus, and
other smaller ornithopods were probably precocial as hatchlings
(see above; [51,55]), whereas hadrosaurs were mainly altricial
[36,51,104]. Parents of precocial offspring only have to care for
the eggs and have to protect and lead the young within the herd.
The latter task could also be managed by other members of the
herd, so that the individual stress of single parents was even lower.
Altricial behavior, in contrast, means the possibility of
extraordinary high juvenile growth rates on the one hand, but also more
stress for the caring parents on the other hand. Parents of an
altricial offspring have to feed their young and have to protect
them against other members of the colony as well as against
carnivores of all sizes. Colonial nesting is also a stress factor in
itself, because many individuals are concentrated in a
comparatively small area . In addition, at least the sexually dimorphic
lambeosaurine hadrosaurs could have had a seasonal rut or
courtship , which also would mean higher seasonal stress for
sexually mature individuals. Thus, the large hadrosaurs likely
suffered much more stress as sexually mature individuals, but their
altricial behavior equalized this disadvantage due to the ability to
outgrow other dinosaurs as juveniles, especially all
contemporaneous theropods . The growth pattern of Tenontosaurus
([34,62]; Fig. 24) is similar to hadrosaurs, so that altricial behavior
can be assumed as well. Thus, altricial behavior was probably one
of the key strategies within Ornithopoda to become large in a short
time and the resulting growth pattern (higher juvenile growth
rates, early sexual and somatic maturity compared to small
ornithopods, consistent development of annuli/LAGs) reflects this
seasonally much more stressful strategy.
It is important to note that the remarks on the reasons for
different growth patterns in ornithopods are tentative hypotheses.
The variability of growth patterns, especially in smaller
ornithopods, is striking and ontogenetic histological studies of more taxa
are urgently needed to strengthen or disprove them. Nevertheless,
the occurrence and/or consistency of annuli/LAGs in ornithopods
is dependent on a mixture of absolute growth rates (which depends
on maximum body size), relative growth rates (depends on the
sampled skeletal element and its ontogenetic stage), the degree of
seasonality of the respective habitat, and the liability of the taxon
to seasonal effects including temperature, humidity, food supply,
migration, and behavior (e.g. precocial or altricial breeding
strategy). Phylogeny plays a rather unimportant role, as already
indicated by Werning .
The large amount of specimens, representing a wide range of
ontogenetic stages, offered the unique opportunity to learn more
about the modes and reasons of variation in bone tissues and
allowed insight into the growth pattern and life history of the
ornithopod dinosaur Dysalotosaurus. For this purpose, up to 70
individual bones were sampled, comprising femora, tibiae, humeri,
fibulae, and prepubic processes.
Variation within the bone tissue was mainly found between
different skeletal elements and between different units of single
cross sections. The former is the result of different relative growth
rates, which are dependent on the individual size of a certain
element and its degree of utilization within the skeleton. Skeletal
elements with a large absolute size, with main weight bearing
functions, and elements intensively used for movements (e.g. for
locomotion) experience higher relative growth rates than other
elements. Some elements have of course combined these
characters, which explain the highest growth rates in the femur
for instance. Accordingly, the only predictable model on the
occurrence of annuli/LAGs in Dysalotosaurus is their increasing
abundance in skeletal elements with lower relative growth rate.
The number of growth cycles naturally increases during ontogeny,
but this definitely is not the case for annuli/LAGs. The
extraordinary variation in the development of annuli/LAGs in
Dysalotosaurus eliminates prediction of their existence and relative
number in skeletal elements of different ontogenetic stages.
Intra-cortical variation in bone tissue is mainly the result of
osseous drift and variation in bone wall thickness during growth.
The relationship between osseous drift, bone wall thickness, bone
tissue variation, and resulting relative growth rates, can now be
N A long bone with a bended long axis experiences osseous drift
from the convex to the concave side of this long axis.
N Relative growth rates, derived from the organizational degree
and the density of vascular canals, are lower on the convex side
of the bended long axis and higher on its concave side.
N Growth rates are also relatively higher in thicker cross sectional
units than in thinner units.
N Variation in bone tissue within a cross section decreases the
more consistent and round the transverse shape of a bone is. A
shaft with a triangular transverse outline contains more
variation than a shaft with a circular transverse outline.
N In the case of partial sampling of a bended long bone, the part
with the best potential record of ordinary bone tissue and
growth cycles is the flat wall on its concave side.
The bone histology of Dysalotosaurus is most similar to Dryosaurus
altus in respect of ontogenetic stages, rarity of annuli/LAGs,
variation of bone tissues, low degree of secondary remodeling, and
the absence of an External Fundamental System. This confirms
the close relationship and a similar growth pattern and general life
style of these taxa.
A new type of growth cycles was used to reconstruct the life
history of Dysalotosaurus, despite the scarcity and variability of
annuli/LAGs. Growth curves of femora (derived from this
alternation of fast and slow growing zones) revealed that
Dysalotosaurus grew with a moderate rate in its juvenile stage until
approximately six years of age, experienced accelerated growth
during its sexually immature stage until reaching sexual maturity
at approximately ten years of age, and had its exponential growth
phase as sexually mature individual until the 14th year of life,
where the maximum growth rate was reached. Afterwards, the
growth rate decelerated and might have reached asymptotic
growth well after 20 years. However, most likely none of the
members of the Dysalotosaurus herd reached the growth plateau of
The group of large individuals within the size-frequency
distribution obviously consists of sexually mature individuals,
because medullary bone was found in a tibia and a fibula of this
size range. The time of initial sexual maturity was discovered as a
transitional mark (MISM) in five large femora representing a slight
slow-down of bone apposition rates.
Indeterminate growth, combined with delayed sexual maturity,
is assumed to represent the optimal growth strategy of
Dysalotosaurus to withstand intra-specific competition and its high liability
The results of the bone histological study of Dysalotosaurus were
finally combined with a relationship between abundance and
consistency of annuli/LAGs in recent mammals and their
respective seasonal environment. Smaller species of ornithopods
are less exposed to seasonal effects than the large species mainly
based on differences in food demands, growth rates, and breeding
strategy. In fact, the achievement of large size within Ornithopoda
was probably linked to a change in breeding strategy from
precocial to altricial behavior.
Materials and Methods
The key literature for an introduction into bone histology,
where also the here used terms are explained, comprises Castanet
et al. , Chinsamy-Turan , Erickson , Francillon-Vieillot
et al. , Klevezal , and Ricqles et al. .
The sections used by Chinsamy  could not be re-examined,
so that this study is completely based upon newly produced thin
30 femora, 12 tibiae, 13 humeri, seven fibulae, and eight
prepubic processes were sampled, but not all of the obtained thin
sections were well preserved. Thus, 11 femora, two tibiae, and four
humeri were inappropriate to be considered for measurements and
correlations and are therefore also not included in the Tables 1, 2,
Location and Production of Thin Sections
The bones used for thin sectioning were loaned from the
collections of the Geowissenschaftliches Zentrum, University of
Go ttingen (GZG), the Staatliches Museum f ur Naturkunde,
Stuttgart (SMNS), and the Institut und Museum fu r Geologie
und Palaontologie, University of Tu bingen (GPIT). Measurements
of additional specimens were also made in the Museum f ur
Naturkunde, Berlin (MB) and the Natural History Museum,
London (R/NHMUK). All sampled bones (femora, tibiae, humeri,
fibulae, and pubii) where already broken, lacking either the distal
or proximal ends. In case of the femora, it was also possible to use
isolated shafts, because the distal beginning of the fourth
trochanter or the medial depression helped to clarify its orientation
and the best position for the thin section. The prepubic process of
the pubis represents the only non-long bone element and was
chosen to highlight further variability within the skeleton of
Dysalotosaurus. It is important to note that it was impossible to take
thin sections from a standard level, because only incomplete
specimens were used. Furthermore, it was aimed to cause as less
damage as possible, so that most of the cuts were carried out close
to broken surfaces. Thus, the sections are standardized to a single
interval along the bone shaft and not to a single level (Fig. 25).
Distinct processes or expansions helped to verify the relative
position of the section.
The bones were cross cut with a diamond powder disk on a
precision saw. Due to the brittle nature of many bones, they were
temporally embedded in acetone dissolvable two-component
epoxy-resin (Technovit 5071) during the sawing process. By the
following well known process of grinding and cutting (see e.g.
[9,22,109]), the produced thin section got a final thickness of
approximately 100 mm.
Sorting of Thin Sections
The description of bone tissue types and structures found in the
sampled bones of Dysalotosaurus generalizes the observations for
each of the elements. Some cross sectional units have very special
and recognizable features, which helped to orientate them even
without the actual bone. Distances are only measured along the
anteroposterior axis or the mediolateral axis (see Tables 1, 2, and
3). All steps beyond the description, which incorporates the count
and correlation of growth cycles, were only done with femora,
tibiae, and humeri. The other sampled elements were too close to
the metaphysis (fibulae) or they had a too thin periosteal bone wall
(pubii) to gain enough quantitative information.
All thin sections with growth cycles were sketched using Adobe
Photoshop 7.0 software. Due to the large error in taking
standardized thin sections, it was impossible to simply superimpose
the sketches of different ontogenetic stages of a sampled element to
get a complete record of all growth cycles from the smallest to the
largest specimen. Thus, thin sections of femora were sorted into
four groups and the sections of the tibiae into two groups
depending on cutting level and cross sectional shape. Humeri were
not sorted due to the relative constancy of the outer cross sectional
shape (Tables 1, 2, and 3).
Conversion of Growth Cycles into Absolute Age
The basic assumption is the annual character of the present
growth cycles (see above). It was the goal to correlate the cycles of
all cross sections of one group of a single skeletal element, to count
the final number of cycles, and to equalize them into years.
Superimposition of sketches did not lead to a good correlation of
growth cycles due to variation in cross sectional shape and the
course and distances of growth cycles to each other. Hence,
another way was chosen to get a correlation, which was also
carried out by using Adobe Photoshop 7.0 software.
The end of each growth cycle was marked in the sketches by a
permanent line. A standard location within the cross sections,
which usually revealed the best record of growth cycles, was
determined for femora, tibiae, and humeri respectively. In tibiae,
two fitting locations were found and the final growth cycle values
were then averaged.
The first step towards the correlation of cycles was the definition
of an unambiguous and repeatable midpoint for every used cross
section (Fig. 26). Femoral cross sections mostly have a triangular
shape, so that two types of geometric triangles were generated.
The vertices of the first triangle were set on the utmost extremity of
each of the three corners of the cross section (Fig. 26A). The
vertices of the second triangle were generated by three straight
lines, which were placed on the external edge of the three straight
walls. Each line was then graphically shifted onto the utmost
extremity of the opposing corner in the cross section and the
respective vertex was set. The midpoint of both triangles was
generated, but the midpoints of both triangles did not coincide in
most cases. The midpoint of a straight line, drawn between both
triangle midpoints, was therefore defined (Fig. 26B). To minimize
possible error, a circle was additionally drawn as large as possible
to fit right on the outer contour of the femoral cross section.
Another straight line was created between the midpoint of this
circle and the combined midpoint of both triangles, so that the
actual midpoint of the whole femoral cross section was the
midpoint of this line (Fig. 26C).
The cross sectional shape of tibiae and humeri were much more
oval in shape. Here, the midpoints of two circles were used to
determine the midpoint of the cross section. One circle was
graphically scaled down as small as possible to enclose the cross
section externally and just tangent the outer edge. The second
circle was scaled up as large as possible to tangent the outer cross
sectional edge internally. The midpoint of a straight line, which
was drawn between the two obtained circle midpoints, was then
determined as the midpoint of the cross section (Fig. 26D).
During the next step, the distance between the cross sectional
midpoint and each of the recorded growth cycles was measured
and transformed into partial percentages of the distance between
the midpoint and the external cross sectional edge. The reference
measurement for each of the cross sections, representing 100%
from midpoint to external edge, was already measured before at
the respective sampled specimen. Since not the same reference
measurement could be taken from each of the femora, tibiae, and
humeri, regression equations were calculated with Microsoft
Office Excel 2007 software to get the allometric relationships for
these distances. In the end, all reference measurements were
transformed into diaphyseal circumference and distal mediolateral
width in femora, distal mediolateral width for tibiae, and
mediolateral width at the deltopectoral crest for humeri
(Tables 1, 2, and 3). The data for the allometric calculation was
taken from the measurement dataset of complete specimens of
these long bones (Table S1).
It is important to note that each measurement from the cross
sectional midpoint to a growth cycle was taken perpendicular to
the course of the latter. In all femora, the best growth record was
preserved in the lateral part of the posterior wall, close to the
posterolateral corner. A straight line was drawn from the midpoint
parallel to the course of the growth cycles and the measurement
was then taken laterally from the midpoint and perpendicular to
the course of the growth cycles (Fig. 26C). In all tibiae and humeri,
such an additional line was not necessary and all measurements
were directly taken from the midpoint. The best growth record in
tibiae was preserved in anterior and medial direction and in the
humeri in anterior direction only.
A special cycle, observed in five large femora and marking initial
sexual maturity (MISM), was measured in the same way as the
All measured percentages of growth cycles were then
transformed, in a third step, into partial values of the reference distance
of the respective cross section (representing 100%) and recorded in
an Excel file. The values of each cross section were sorted in their
respective group, one in humeri, two in tibiae, and four in femora.
The following correlation of growth cycles was therefore done only
within a single group. The still uncorrelated growth cycles of each
group were related to age in years. A diagram was then created,
were the x-axis represents age and the y-axis partial reference
values of the growth cycles of each cross section of this group. The
correlation of growth cycles to age in years started by fitting the
lowest known value to an age of one year. This could be done,
because the respective value was derived from the smallest
sampled specimen (Tab. 1), where the corresponding growth
cycle was found at the outer edge of the completely unremodelled
and unresorbed bone wall (Fig. 8AB). The distance of successive
growth cycles of the other cross sections in the dataset as well as
the diagram revealed the general distance of values between two
successive years. First, all values of a single cross section were
shifted, so that the smallest value of a cross section fit onto a value
of another one. In this way, the values of every single section of this
group were fitted to get a single curve in the diagram, where
possible outliers are minimized. It occurred especially in large or
strongly obscured cross sections that the successive growth cycle
values could be separated, because the large distance between
them could be filled by successive values of other sections. The
MISM was separately signed into the diagrams of two groups of
Calculation of Body Mass
Two out of four groups of sampled femora were chosen to
convert their age related growth cycles into body mass estimates.
The samples of the other two groups are not appropriate, because
their location within the shaft is either too proximal or too distal,
and their small number of recorded growth cycles only covers
three to four years. In contrast, growth cycles of several samples in
femoral group one and two were often placed within the same year
of age during correlation. In this case, the average of all values of
this year was used as the basis for the body mass calculation.
Two methods of calculating body mass by skeletal elements
were considered. The first method was derived by Anderson et al.
 by using the combined humeral and femoral shaft
circumference to calculate body mass in quadruped animals. For
biped animals, only the femoral shaft circumference was necessary.
The following equation was therefore used for Dysalotosaurus
femora, W = 0.16 CF2.73, where W is the weight and CF is the
circumference of the femur.
The accuracy of this method was recently doubted .
However, the conventional model predicts the body mass of small
to medium-sized animals much better than the proposed
alternative . It is also more reliable to the natural variability
of body mass in different size categories than the proposed
nonlinear alternative . Thus, it is assumed that the method of
Anderson et al.  used here is still the best model to predict the
body mass in the rather small-bodied dinosaur Dysalotosaurus.
The second method was derived by Erickson & Tumanova 
known as Developmental Mass Extrapolation (DME). The basis
for this body mass calculation, which emphasizes the effect of
ontogeny on mass increase, is the assumption that the
approximately third power of femoral length corresponds to body mass in
Alligator (data in ) and the California Gull (data in ).
Both species represent members of outgroups of non-avian
dinosaurs (Extant Phylogenetic Bracket ), so that the ratio
of femoral length to body mass could also be used for
Dysalotosaurus. This was also done for the respective values of the
Establishing the Growth Curve
To compare the life history of Dysalotosaurus to other dinosaurs
and recent animals, a type of growth curve had been chosen,
which was already used by Erickson et al. .
The calculated body mass of the averaged growth cycles was
therefore plotted against their respective age in years. The
equation y = a/(1+exp (b * (x+c)))+d describes the sigmoidal course
of this type of growth curve (y = body mass; x = age in years;
a = largest known body mass; b, c, d = parameters to fit). The
variable a was derived from the largest known femur with a
calculated body mass of 115.3 kilograms. Only the secured growth
cycle values were integrated and all unsecured values, including
the values externally to the MISM, were excluded. The latter
values were entered afterwards into the curves to evaluate their
significance and possible age correlation. The MISM itself was
included with the corresponding age of 9.5 years in femur group
one and 10.5 years in femur group two. A total of four curves were
created, including the calculated body masses by the methods of
Anderson et al.  and Erickson & Tumanova  for femoral
group one and two, respectively. The dataset was entered into the
software Microcal Origin and the non-linear curve fit function
(basing on least-square regression analysis) was performed using
the equation mentioned above.
Growth Rates and Age/Size Frequency Distribution
To get yearly and daily growth rates, the calculated yearly body
masses were derived by using the sigmoidal equations and the four
parameters of each of the four growth curves. One version
corresponds to the growth rate in a recent year (365 days) and the
second version corresponds to a year in the Late Jurassic
(Kimmeridge, 150 million years ago), which contained
approximately 377.76 days [27,115]. The maximum growth rate per day,
calculated in gram, was then plotted into the diagram of Erickson
et al. .
The final step was the combination of the absolute age estimates
with the size frequency distribution of all femora (Fig. 23), so that
one can assign a certain position within this distribution to a
certain age. First, the allometric relationships for the femoral distal
mediolateral width and femoral circumference were determined
by combining the values of all measured specimens with the
sectioned samples. The allometric relationship for the femoral
circumference and length was obtained from the measured
specimens with both distances preserved (Table S1). Second, the
calculation of age for all femora was carried out by conversion of
the sigmoidal equation to x (age in years), which resulted in the
following equation: x = ln ((a/(y2d)21)/b+c (y = body mass
calculated by either the method of  or ; the parameters
a, b, c, d were derived from each of the four growth curves). The
obtained ages of the separately calculated versions for both
femoral groups were averaged for the dataset derived from the
Anderson et al.  body mass calculation and for the dataset
derived from the Erickson & Tumanova  body mass
calculation. These average estimates were then correlated with
the circumferences and distal mediolateral widths of the femora.
Thus, every single value of both measured distances can now be
assigned to a specific age (see Tab. 1).
Figure S1 Detail of cross section of tibia SMNS T 13,
under polarized light; Anterolateral unit internally;
Marrow cavity at top left. The original vascularization is
obviously altered by postmortem dissolution of bone tissue.
Former primary osteons are lost during this process and the
vascular canals are widened. Scale bar = 500 mm.
Table S1 List of all specimens and measured data of
humeri, tibiae, and femora, which were used for the
allometric calculation of the reference values necessary
Text S1 This text comprises a more comprehensive
description of the thin sections of all five skeletal
elements of Dysalotosaurus and additionally includes
notes on the modes of preservation of the bone
microstructure as well as on the occurrence and
distribution of osteocyte lacunae and Sharpeys fibers.
I want to thank Martin Sander (Bonn) for introducing me into the depths of
bone histology. Without him, I would never have properly understood this
tricky topic with all its sometimes unmanageable variation. He also was the
first who recognized the medullary bone in a sample, which led to many
important results. I am very thankful for his help and time. I also thank
Nicole Klein and Koen Stein, both from Martin Sanders team in Bonn, for
helpful discussions and hints as well as Olaf D ulfer (also Bonn) for his
comments on the production of thin sections. I further want to thank
Oliver Rauhut (Munich) for his support and advice, Cathleen Helbig
(Munich) for her skilled preparation of the thin sections, Frank Melcher
(Hannover) for giving me access to microscopy equipment to make
additional photos of the thin sections, and Wiete Hu bner for her help with
the Excel and Microcal Origin software. I especially want to thank the
persons who were responsible for the collections providing the sampled and
measured material. These are Rainer Schoch (SMNS), Philipe Havlik and
Alexander Hohloch (GPIT), Daniela-Schwarz-Wings (MB), Mike Reich
and Tanja Stegemann (GZG), and Sandra Chapman (NHMUK). Richard
Butler and Rod Scheetz kindly helped improve the English of the text. I
finally want to thank all reviewers who helped me greatly to improve this
Analyzed the data: TRH. Contributed reagents/materials/analysis tools:
TRH. Wrote the paper: TRH.
1. Lehman TM ( 2007 ) Growth and population age structure in the horned dinosaur Chasmosaurus . In: CarpenterK, ed. Horns and beaks: ceratopsian and ornithopod dinosaurs . Bloomington: Indiana University Press. pp 259 - 317 .
2. Sampson SD , Ryan MJ , Tanke DH ( 1997 ) Craniofacial ontogeny in centrosaurine dinosaurs (Ornithischia: Ceratopsidae): taxonomic and behavioural implications . Zool J Linn Soc 121 : 293 - 337 .
3. Tumarkin-Deratzian AR ( 2003 ) Bone surface textures as ontogenetic indicators in extant and fossil archosaurs: Macroscopic and histological evaluations . PhD Thesis . Philadelphia: University of Pennsylvania. 333 p.
4. Tumarkin-Deratzian AR ( 2009 ) Evaluation of long bone surface textures as ontogenetic indicators in centrosaurine ceratopsids . Anat Rec 292 : 1485 - 1500 .
5. Castanet J , Francillon-Vieillot H , Meunier FJ , de Ricqle`s A ( 1993 ) Bone and individual aging . In: Hall BK, ed. Bone. Volume 7: Bone growth - B. Boca Raton: CRC Press . pp 245 - 283 .
6. Chinsamy-Turan A ( 2005 ) The microstructure of dinosaur bone - Deciphering biology with fine-scale techniques . Baltimore/London: Johns Hopkins University Press. 195 p.
7. Erickson GM ( 2005 ) Assessing dinosaur growth patterns: a microscopic revolution . Trends Ecol Evol 20 ( 12 ): 677 - 684 .
8. Francillon-Vieillot H , de Buffrenil V , Castanet J , Geraudie J, Meunier FJ , et al. ( 1990 ) Microstructure and mineralization of vertebrate skeletal tissues . In: Carter JE, ed. Skeletal biomineralization: Patterns, processes and evolutionary trends . New York : Van Nostrand Reinhold . pp 471 - 530 .
9. Hubner TR ( 2011 ) Ontogeny in Dysalotosaurus lettowvorbecki . PhD Thesis . Munich: Ludwig Maximilians Universitat. 318 p.
10. de Ricqle`s A, Meunier FJ , Castanet J , Francillon-Vieillot H ( 1991 ) Comparative microstructure of bone . In: Hall BK, ed. Bone. Volume 3 : Bone matrix and bone specific products . Boca Raton: CRC Press . pp 1 - 78 .
11. Janensch W ( 1914 ) Die Gliederung der Tendaguru-Schichten im TendaguruGebiet und die Entstehung der Saurier-Lagerstatten . Arch Biont 3 ( 3 ): 227 - 261 .
12. Maier G ( 2003 ) African Dinosaurs Unearthed . Bloomington: Indiana University Press. 512 p.
13. Bussert R , Heinrich W-D , Aberhan M ( 2009 ) The Tendaguru Formation (Late Jurassic to Early Cretaceous, southern Tanzania): definition, palaeoenvironments, and sequence stratigraphy . Fossil Record 12 ( 2 ): 141 - 174 .
14. Heinrich W-D ( 1999 ) The taphonomy of dinosaurs from the Upper Jurassic of Tendaguru (Tanzania) based on field sketches of the German Tendaguru Expedition (1909-1913). Mitt Mus Natkd Berl , Geowiss Reihe 2 : 25 - 61 .
15. Reck H ( 1925 ) Grabungen auf fossile Wirbeltiere in Deutsch-Ostafrika . Geol Char 31 : 1 - 36 .
16. Russell D , Beland F , McIntosh JS ( 1980 ) Paleoecology of the dinosaurs of Tendaguru (Tanzania) . Mem Soc Geol France 59 ( 139 ): 169 - 175 .
17. Aberhan M , Bussert R , Heinrich W-D , Schrank E , Schultka S , et al. ( 2002 ) Paleoecology and depositional environments of the Tendaguru Beds (Late Jurassic to Early Cretaceous , Tanzania). Mitt Mus Natkd Berl, Geowiss Reihe 2 : 201 - 205 .
18. Chinsamy A ( 1995 ) Ontogenetic changes in the bone histology of the Late Jurassic ornithopod Dryosaurus lettowvorbecki . J Vert Pal 15 ( 1 ): 96 - 104 .
19. Galton P ( 1977 ) The ornithopod dinosaur Dryosaurus and a LaurasiaGondwanaland connection in the Upper Jurassic . Nature 268 : 230 - 232 .
20. Hubner TR , Rauhut OWM ( 2010 ) A juvenile skull of Dysalotosaurus lettowvorbecki (Ornithischia: Iguanodontia), and implications for cranial ontogeny, phylogeny, and taxonomy in ornithopod dinosaurs . Zool J Linn Soc 160 : 366 - 396 .
21. Hutton JM ( 1986 ) Age determination of living Nile crocodiles from the cortical stratification of bone . Copeia 1986 ( 2 ): 332 - 341 .
22. Klevezal GA ( 1996 ) Recording structures of mammals: Determination of age and reconstruction of life history . Rotterdam: Balkema . 274 p.
23. Peabody FE ( 1961 ) Annual growth zones in living and fossil vertebrates . J Morph 108 : 11 - 62 .
24. Botha J , Chinsamy A ( 2000 ) Growth patterns from the bone histology of the cynodonts Diademodon and Cynognathus . J Vert Pal 20 : 705 - 711 .
25. Chinsamy A ( 1990 ) Physiological implications of the bone histology of Syntarsus rhodesiensis (Saurischia: Theropoda) . Palaeont Afr 27 : 77 - 82 .
26. Erickson GM , Brochu CA ( 1999 ) How the ''terror crocodile'' grew so big . Nature 398 : 205 - 206 .
27. Erickson GM , Tumanova TA ( 2000 ) Growth curve of Psittacosaurus mongoliensis Osborn (Ceratopsia: Psittacosauridae) inferred from long bone histology . Zool J Linn Soc 130 : 551 - 566 .
28. Horner JR , Padian K ( 2004 ) Age and growth dynamics of Tyrannosaurus rex . Proc R Soc Lond B 271 : 1875 - 1880 .
29. Sander PM , Mateus O , Laven T , Knotschke N ( 2006 ) Bone histology indicates insular dwarfism in a new Late Jurassic sauropod dinosaur . Nature 441 : 739 - 741 .
30. Varricchio DJ ( 1993 ) Bone microstructures of the Upper Cretaceous theropod dinosaur Troodon formosus . J Vert Pal 13 : 99 - 104 .
31. Castanet J , Meunier FJ , de Ricqle`s A ( 1977 ) L'enregistrement de la croissance cyclique par le tissu osseux chez les vertebres poikilothermes: donnees comparatives et essai de synthese . Bull Biol Fr Belg 111 : 183 - 202 .
32. Klein N ( 2004 ) Bone histology and growth of the prosauropod dinosaur Plateosaurus engelhardti Meyer 1837 from the Norian bonebeds of Trossingen (Germany) and Frick (Switzerland) . PhD Thesis . Bonn: Rheinische FriedrichWilhelms-Universitat . 128 p.
33. Klein N , Sander PM ( 2007 ) Bone histology and growth of the prosauropod dinosaur Plateosaurus engelhardti (von Meyer 1837) from the Norian bonebeds of Trossingen (Germany) and Frick (Switzerland) . Spec Pap Pal 77 : 169 - 206 .
34. Werning S ( 2005 ) Long bone histology of Tenontosaurus tilletti Ostrom 1970 (Early Cretaceous, North America), with comments on ontogeny . Master's Thesis . Norman: University of Oklahoma. 150 p.
35. Chinsamy A ( 1993 ) Bone histology and growth trajectory of the prosauropod dinosaur Massospondylus carinatus Owen . Mod Geol 18 : 319 - 329 .
36. Horner JR , de Ricqle`s A , Padian K ( 2000 ) Long bone histology of the hadrosaurid dinosaur Maiasaura peeblesorum: growth dynamics and physiology based on an ontogenetic series of skeletal elements . J Vert Pal 20 ( 1 ): 115 - 129 .
37. de Ricqle`s A ( 1983 ) Cyclical growth in the long limb bones of a sauropod dinosaur . Acta Pal Pol 28 : 225 - 232 .
38. Horner JR , de Ricqle`s A , Padian K ( 1999 ) Variation in dinosaur skeletochronology indicators: implications for age assessment and physiology . Paleobiology 25 ( 3 ): 295 - 304 .
39. Amprino R ( 1947 ) La structure du tissu osseux envisage`e comme expression de differences dans la vitesse de l' accroisement . Arch Biol 58 : 315 - 330 .
40. de Buffrenil V , Houssaye A , Bohme W ( 2008 ) Bone vascular supply in monitor lizards (Squamata: Varanidae): Influence of size, growth, and phylogeny . J Morph 269 : 533 - 543 .
41. Castanet J , Curry Rogers K , Cubo J , Boisard J-J ( 2000 ) Periosteal bone growth rates in extant ratites (ostriches and emu). Implications for assessing growth in dinosaurs . C R Acad Sci Paris, Sci Vie 323 : 543 - 550 .
42. Erickson GM , Makovicky PJ , Currie PJ , Norell MA , Yerby SA , et al. ( 2004 ) Gigantism and comparative life-history parameters of tyrannosaurid dinosaurs . Nature 430 : 772 - 775 .
43. Padian K , Horner JR , de Ricqle`s A ( 2004 ) Growth in small dinosaurs and pterosaurs: the evolution of archosaurian growth strategies . J Vert Pal 24 ( 4 ): 555 - 571 .
44. Turvey ST , Green OR , Holdaway RN ( 2005 ) Cortical growth marks reveal extended juvenile development in New Zealand moa . Nature 435 : 940 - 943 .
45. Starck JM , Chinsamy A ( 2002 ) Bone microstructure and developmental plasticity in birds and other dinosaurs . J Morph 254 : 232 - 246 .
46. de Margerie E , Cubo J , Castanet J ( 2002 ) Bone typology and growth rate: Testing and quantifying ''Amprino's rule'' in the mallard (Anas platyrhynchos) . C R Acad Sci Paris, Biol 325: 221 - 230 .
47. de Margerie E , Robin J-P , Verrier D , Cubo J , Groscolas R , et al. ( 2004 ) Assessing the relationship between bone microstructure and growth rate: A fluorescent labelling study in the king penguin chick (Aptenodytes patagonicus) . J Exp Biol 207 : 869 - 879 .
48. Chinsamy A , Rich T , Vickers-Rich P ( 1998 ) Polar dinosaur bone histology . J Vert Pal 18 : 385 - 390 .
49. Cooper LN , Lee AH , Taper ML , Horner JR ( 2008 ) Relative growth rates of predator and prey dinosaurs reflect effects of predation . Proc R Soc Lond B: doi:10.1098/rspb.2008 .0912.
50. Horner JR , Currie PJ ( 1994 ) Embryonic and neonatal morphology and ontogeny of a new species of Hypacrosaurus (Ornithischia, Lambeosaurinae) from Montana and Alberta . In: Carpenter K, Hirsch KF , Horner JR, eds. Dinosaur eggs and babies . Cambridge : Cambridge University Press . pp 312 - 336 .
51. Horner JR , Padian K , de Ricqle`s A ( 2001 ) Comparative osteohistology of some embryonic and perinatal archosaurs: developmental and behavioral implications for dinosaurs . Paleobiology 27 : 39 - 58 .
52. Horner JR , de Ricqle`s A , Padian K , Scheetz RD ( 2009 ) Comparative long bone histology and growth of the 'Hypsilophodontid'' dinosaurs Orodromeus makelai, Dryosaurus altus, and Tenontosaurus tilletti (Ornithischia: Euornithopoda) . J Vert Pal 29 ( 3 ): 734 - 747 .
53. Scheetz RD ( 1999 ) Osteology of Orodromeus makelai and the phylogeny of basal ornithopod dinosaurs . PhD Thesis . Bozeman: Montana State University. 186 p.
54. Cerda IA , Chinsamy-Turan A ( 2008 ) Bone histology of Gasparinisaura cincosaltensis, a basal ornithopod dinosaur from the Upper Cretaceous of Patagonia . Ameghiniana 45 ( 4 ): 24R .
55. Winkler DA ( 1994 ) Aspects of growth in the Early Cretaceous Proctor Lake ornithopod . J Vert Pal 14 ( 3 ): 53A .
56. Curry KA ( 1999 ) Ontogenetic histology of Apatosaurus (Dinosauria: Sauropoda): New insights on growth rates and longevity . J Vert Pal 19 ( 4 ): 654 - 665 .
57. Sander PM ( 2000 ) Long bone histology of the Tendaguru sauropods: implications for growth and biology . Paleobiology 26 ( 3 ): 466 - 488 .
58. Woodward HN , Lehman TM ( 2009 ) Bone histology and microanatomy of Alamosaurus sanjuanensis (Sauropoda: Titanosauria) from the Maastrichtian of Big Bend National Park, Texas . J Vert Pal 29 ( 3 ): 807 - 821 .
59. Lee AH ( 2004 ) Histological organization and its relationship to function in the femur of Alligator mississippiensis . J Anat 204 : 197 - 207 .
60. Chinsamy A , Tumarkin-Deratzian A ( 2009 ) Pathologic bone tissues in a turkey vulture and a nonavian dinosaur: Implications for interpreting endosteal bone and radial fibrolamellar bone in fossil dinosaurs . Anat Rec 292 : 1478 - 1484 .
61. Holmes JR ( 1958 ) Experimental transmission of avian osteopetrosis . J Comp Pathol 68 : 439 - 451 .
62. Lee AH , Werning S ( 2008 ) Sexual maturity in growing dinosaurs does not fit reptilian growth models . Proc Nat Acad Sci 105 : 582 - 587 .
63. Schweitzer MH , Wittmeyer JL , Horner JR ( 2005 ) Gender-specific reproductive tissue in ratites and Tyrannosaurus rex . Science 308 : 1456 - 1460 .
64. Anderson JF , Hall-Martin A , Russell DA ( 1985 ) Long-bone circumference and weight in mammals, birds and dinosaurs . J Zool Lond 207 : 53 - 61 .
65. Ray S , Chinsamy A ( 2004 ) Diictodon feliceps (Therapsida, Dicynodontia): Bone histology, growth, and biomechanics . J Vert Pal 24 ( 1 ): 180 - 194 .
66. Enlow DH , Brown SO ( 1958 ) A comparative histological study of fossil and recent bone tissues . Part III. Tex J Sci 10 : 187 - 230 .
67. de Ricqle`s AJ , Padian K , Horner JR , Francillon-Vieillot H ( 2000 ) Palaeohistology of the bones of pterosaurs (Reptilia: Archosauria): anatomy, ontogeny, and biomechanical implications . Zool J Linn Soc 129 ( 3 ): 349 - 385 .
68. Bybee PJ , Lee AH , Lamm E-T ( 2006 ) Sizing the Jurassic theropod dinosaur Allosaurus: assessing growth strategy and evolution of ontogenetic scaling of limbs . J Morph 267 ( 3 ): 347 - 359 .
69. Botha J , Chinsamy A ( 2005 ) Growth patterns of Thrinaxodon liorhinus, a nonmammalian cynodont from the Lower Triassic of South Africa . Palaeontology 48 ( 2 ): 385 - 394 .
70. Ray R , Botha J , Chinsamy A ( 2004 ) Bone histology and growth patterns of some nonmammalian therapsids . J Vert Pal 24 : 634 - 648 .
71. Enlow DH ( 1962 ) A Study of the Post-Natal Growth and Remodeling of Bone . Am J Anat 110 : 79 - 101 .
72. Chinsamy A , Hurum JH ( 2006 ) Bone microstructure and growth patterns of early mammals . Acta Pal Pol 51 ( 2 ): 325 - 338 .
73. Reid REH ( 1984 ) The histology of dinosaurian bone, and its possible bearing on dinosaur physiology . Symp Zool Soc Lond 52 : 629 - 662 .
74. Currey JD ( 1984 ) Comparative Mechanical Properties and Histology of Bone . Am Zool 24 : 5 - 12 .
75. Erickson GM , Rogers KC , Varricchio DJ , Norell MA , Xu X ( 2007 ) Growth patterns in brooding dinosaurs reveals the timing of sexual maturity in nonavian dinosaurs and genesis of the avian condition . Biol Lett: doi:10.1098/ rsbl.2007 .0254.
76. Sander PM , Klein N , Buffetaut E , Cuny G , Suteethorn V , et al. ( 2004 ) Adaptive radiation in sauropod dinosaurs: bone histology indicates rapid evolution of giant body size through acceleration . Org Div Evol 4 : 165 - 173 .
77. Erickson GM , Curry Rogers K , Yerby SA ( 2001 ) Dinosaurian growth patterns and rapid avian growth rates . Nature 412 : 429 - 433 .
78. Lehman TM , Woodward HN ( 2008 ) Modeling growth rates for sauropod dinosaurs . Paleobiology 34 ( 2 ): 264 - 281 .
79. de Ricqle`s AJ , Padian K , Horner JR ( 2003 ) On the bone histology of some Triassic pseudosuchian archosaurs and related taxa . Ann Paleontol 89 : 67 - 101 .
80. Case TJ ( 1978 ) Speculations on the growth rate and reproduction of some dinosaurs . Paleobiology 4 : 320 - 328 .
81. Klein N , Sander PM ( 2008 ) Ontogenetic stages in the long bone histology of sauropod dinosaurs . Paleobiology 34 ( 2 ): 247 - 263 .
82. de Ricqle`s AJ ( 1976 ) On bone histology of fossil and living reptiles, with comments on its functional and evolutionary significance . In: Bellairs A d'A, Cox CB, eds. Morphology and Biology of Reptiles. Linn Soc Symp Ser 3 : 123 - 150 .
83. Castanet J , Croci S , Aujard F , Perret M , Cubo J , et al. ( 2004 ) Lines of arrested growth in bone and age estimation in a small primate: Microcebus murinus . J Zool Lond 263 : 31 - 39 .
84. Burr DB ( 1992 ) Estimated intracortical bone turnover in the femur of growing macaques: Implications for their use as models in skeletal pathology . Anat Rec 232 : 180 - 189 .
85. Mulhern DM , Ubelaker DH ( 2003 ) Histologic examination of bone development in juvenile chimpanzees . Am J Phys Anth 122 : 127 - 133 .
86. Singh IJ , Tonna EA , Gandel CP ( 1974 ) A comparative histological study of mammalian bone . J Morph 144 : 421 - 431 .
87. Horner JR , Weishampel DB ( 1988 ) A comparative embryological study of two ornithischian dinosaurs . Nature 332 : 256 - 257 .
88. Chure DJ , Turner C , Peterson F ( 1994 ) An embryo of Camptosaurus from the Morrison Formation (Jurassic, Middle Tithonian) in Dinosaur National Monument, Utah . In: Carpenter K, Hirsch KF , Horner JR, eds. Dinosaur eggs and babies . Cambridge : Cambridge University Press . pp 298 - 310 .
89. Andrews RM ( 1982 ) Patterns of growth in reptiles . In: Gans C, Pough FH, eds. Biology of the Reptilia . London: Academic Press. pp 273 - 320 .
90. Chinsamy A , Codorniu L , Chiappe L ( 2008 ) Developmental growth patterns of the filter-feeder pterosaur, Pterodaustro guinnazui . Biol Let 4 : 282 - 285 .
91. Erickson GM , Currie PJ , Inouye BD , Winn AA ( 2006 ) Tyrannosaur life tables: An example of nonavian dinosaur population biology . Science 313 : 213 - 217 .
92. Jarman PJ , Jarman MV ( 1973 ) Social behavior, population structure, and reproduction potential in impala . E Afr Wild J 11 : 329 - 338 .
93. Owen-Smith N ( 1993 ) Comparative mortality rates of male and female kudus: the costs of sexual size dimorphism . J Anim Ecol 62 : 428 - 440 .
94. Jorgenson JT , Festa-Bianchet M , Gaillard J-M , Wishart WD ( 1997 ) Effects of age, sex, disease, and density on survival of bighorn sheep . Ecology 78 ( 4 ): 1019 - 1032 .
95. Mihlbachler MC ( 2003 ) Demography of late Miocene rhinoceroses (Teleoceras proterum and Aphelops malacorhinus) from Florida: linking mortality and sociality in fossil assemblages . Paleobiology 29 : 412 - 428 .
96. Proaktor G , Coulson T , Milner-Gulland EJ ( 2008 ) The demographic consequences of the cost of reproduction in ungulates . Ecology 89 ( 9 ): 2604 - 2611 .
97. Rimblot-Baly F , de Ricqle`s A , Zylberberg L ( 1995 ) Analyse paleohistologique d'une serie de croissance partielle chez Lapparentosaurus madagascariensis (Jurassique Moyen): Essai sur la dynamique de croissance d'un dinosaure sauropode . Ann Pal 81 : 49 - 86 .
98. Woodward HN , Rich TH , Chinsamy A , Vickers-Rich P ( 2011 ) Growth dynamics of Australia's polar dinosaurs . PLoS ONE 6 ( 8 ) : e23339 . doi:10.1371/ journal.pone.0023339.
99. Norman DB , Sues H-D , Witmer LM , Coria RA ( 2004 ) Basal Ornithopoda . In: Weishampel DB, Dodson P , Osmolska H, eds. The Dinosauria. Second edition . Berkeley: University of California Press. pp 393 - 412 .
100. Bell PR , Snively E ( 2008 ) Polar dinosaurs on parade: a review of dinosaur migration . Alcheringa 32 : 271 - 284 .
101. Carrano MT , Janis CM , Sepkoski JJ , Jr. ( 1999 ) Hadrosaurs as ungulate parallels: Lost lifestyles and deficient data . Acta Pal Pol 44 ( 3 ): 237 - 261 .
102. Norman DB , Weishampel DB ( 1985 ) Ornithopod feeding mechanisms: Their bearing on the evolution of herbivory . Am Nat 126 : 151 - 164 .
103. Forster CA ( 1990 ) Evidence for juvenile groups in the ornithopod dinosaur Tenontosaurus tilletti Ostrom . J Pal 64 ( 1 ): 164 - 165 .
104. Horner JR , Makela R ( 1979 ) Nest of juveniles provides evidence of family structure among dinosaurs . Nature 282 ( 5736 ): 296 - 298 .
105. Horner JR , Weishampel DB , Forster CA ( 2004 ) Hadrosauridae . In: Weishampel DB, Dodson P , Osmolska H, eds. The Dinosauria. Second edition . Berkeley: University of California Press. pp 438 - 468 .
106. Varricchio DJ , Martin AJ , Katsura Y ( 2007 ) First trace and body fossil evidence of a burrowing, denning dinosaur . Proc R Soc Lond B 274 ( 1616 ): 1361 - 1368 .
107. Huh M , Lee D-G , Kim J-K , Lim J-D , Godefroit P ( 2011 ) A new basal ornithopod dinosaur from the Upper Cretaceous of South Korea . N Jb Geol Palaont Abh 259 ( 1 ): 1 - 24 .
108. Redelstorff R , Csiki Z , Grigorescu D ( 2009 ) The heritage of Nopcsa: Dwarf status of Hateg ornithopods supported by the histology of long bones . J Vert Pal 29 ( 3 ): 170A .
109. Klein N ( 2010 ) Long Bone Histology of Sauropterygia from the Lower Muschelkalk of the Germanic Basin Provides Unexpected Implications for Phylogeny . PLoS ONE 5 ( 7 ) : e11613 . doi:10.1371/journal.pone.0011613.
110. Packard GC , Boardman TJ , Birchard GF ( 2009 ) Allometric equations for predicting body mass of dinosaurs . J Zool Lond 279 : 102 - 110 .
111. Cawley GC , Janacek GJ ( 2010 ) On allometric equations for predicting body mass of dinosaurs . J Zool Lond 280 : 355 - 361 .
112. Dodson P ( 1975 ) Functional and ecological significance of relative growth in Alligator . J Zool Lond 175 : 315 - 355 .
113. Carrier D , Leon LR ( 1990 ) Skeletal growth in the California gull (Larus californicus) . J Zool Lond 222 : 375 - 389 .
114. Witmer LM ( 1995 ) The extant phylogenetic bracket and the importance of reconstructing soft tissues in fossils . In: Thomason JJ, ed. Functional morphology in vertebrate paleontology. Cambridge : Cambridge University Press . pp 19 - 33 .
115. Wells JW ( 1963 ) Coral growth and geochronometry . Nature 197 : 948 - 950 .