Paleobiology, 45(2), 2019, pp.

317–330
DOI: 10.1017/pab.2019.13

Bone histology of Azendohsaurus laaroussii: Implications for the

evolution of thermometabolism in Archosauromorpha

Jorge Cubo and Nour-Eddine Jalil

Abstract.—This paper is aimed at constraining the phylogenetic frame of the acquisition of endothermy by
Archosauromorpha. We analyzed the bone histology of Azendohsaurus laaroussii. Stylopodial and zeugo-
podial bones show three tissue types: (1) avascular lamellar zonal bone formed at low growth rates; (2) a
scaffold of parallel-fibered bone containing either small primary osteons or simple vascular canals; and (3)
fibrolamellar bone formed at high growth rates. We used quantitative histology to infer the thermometa-
bolic regime of this taxon. We define endothermy as the presence of any mechanism of nonshivering
thermogenesis that increases both body temperature and resting metabolic rate. Thus, estimating the rest-
ing metabolic rate of an extinct organism may be a good proxy to infer its thermometabolic regime (endo-
thermy vs. ectothermy). High resting metabolic rates have been shown to be primitive for the clade
Prolacerta–Archosauriformes. Therefore, we inferred the resting metabolic rates of A. laaroussii, a sister
group of this clade, and of 14 extinct related taxa, using phylogenetic eigenvector maps. All the inferences
obtained are included in the range of variation of resting metabolic rates measured in mammals and birds,
so we can reasonably assume that all these taxa (including Azendohsaurus) were endotherms. A parsimony
optimization of the presence of endothermy on a phylogenetic tree of tetrapods shows that this derived
character state was acquired by the last common ancestor of the clade Azendohsaurus–Archosauriformes
and that there is a reversion in Crocodylia.

Jorge Cubo. Sorbonne Université, Muséum national d’Histoire naturelle, CNRS, Centre de Recherche en
Paléontologie–Paris (CR2P), 4 place Jussieu, BC 104, 75005 Paris, France. E-mail: jorge.cubo_garcia@sorbonne-
universite.fr
Nour-Eddine Jalil. Muséum national d’Histoire naturelle, Sorbonne Université, CNRS, Centre de Recherche en
Paléontologie–Paris (CR2P), 75005 Paris, France; and Laboratory of Biodiversity and Dynamic of Ecosystems,
Department of Geology, Faculty of Sciences Semlalia, University Cadi Ayyad, Marrakesh, Morocco. E-mail:
[email protected]

Accepted: 27 March 2019

Data available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.qc86g11

Introduction
The bone histology of non-archosaurian of basal metabolic rates, because they directly
archosauromorphs has received great attention reflect growth rates […]. The sustained depos-
in the last decade (Ricqlès et al. 2008; Nesbitt ition of fast-growing bone tissues, as displayed
et al. 2009; Botha-Brink and Smith 2011; Wern- by mammals, birds and other dinosaurs, must
ing and Irmis 2011; Legendre et al. 2013, 2016; reflect sustained high basal metabolic rates”
Ezcurra et al. 2014; Mukherjee 2015; Veiga (p. 123). Montes et al. (2007) provided evidence
et al. 2015; Werning and Nesbitt 2016; Jaquier for this hypothesis by showing that the vari-
and Scheyer 2017). Ricqlès et al. (2008) sug- ation of bone growth rates significantly
gested that the capacity of reaching and main- explains the variation of resting metabolic
taining very high bone growth rates is an rates in a sample of extant amniotes. They
apomorphic feature of archosauriforms. argued that this is due to the fact that high
Botha-Brink and Smith (2011) showed that rates of protein synthesis and degradation
this capacity may have been acquired in a involved in the periosteal osteogenesis and endo-
more inclusive node, by the last common ances- steal osteolysis are energy consuming (Montes
tor of the clade Prolacerta–archosaurifoms, with et al. 2007). Consistently, Legendre et al. (2016)
a reversion in Vancleavea (see also Nesbitt et al. inferred resting metabolic rates of a sample of
2009). Padian and Horner (2002) hypothesized Archosauromorpha using quantitative bone hist-
that “the types of tissue deposited in the bones ology and phylogenetic eigenvector maps. Inter-
of extinct animals are the most direct evidence estingly, they inferred a resting metabolic rate of
© 2019 The Paleontological Society. All rights reserved. 0094-8373/19

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 318 JORGE CUBO AND NOUR-EDDINE JALIL

Prolacerta included within the range of variation laaroussii (Dutuit 1976). The analysis of these
of extant mammals (Legendre et al. 2016). Rest- specimens argues strongly against a phylogen-
ing metabolic rate depends on bone growth rate etic position inside Dinosauria (Jalil and Knoll
(as noted earlier), but it depends also on the pres- 2002). Flynn et al. (2010) named the new species
ence of thermogenetic mechanisms. We define Azendohsaurus madagaskarensis from the Trias-
endothermy as the presence of any mechanism sic of Madagascar and provided for the first
of nonshivering thermogenesis (e.g., Lowell time a comprehensive description of the cranial
and Spiegelman 2000; Rowland et al. 2015; Now- anatomy of the genus. More recently, Nesbitt
ack et al. 2017) that increases both body tempera- et al. (2015) provided an extensive description
ture and resting metabolic rate. Thus, the resting the postcranial anatomy of A. madagaskarensis,
metabolic rate of an extinct organism may be a which is now one of the best-known early
good proxy to infer its thermometabolic regime archosauromorphs.
(endothermy vs. ectothermy). Using this proxy, All the available postcranial material in the
the results obtained by Legendre et al. (2016) sug- MNHN collection attributed to A. laaroussii is
gest that endothermy may have been acquired by monotypic (Khaldoun 2014). It was collected
the last common ancestor of the clade Prolacerta– from the locus typicus of A. laaroussii during the
archosauriforms. The next step to better constrain same initial field season (Dutuit 1976) and
the phylogenetic frame of the acquisition of endo- strongly resembles the postcranial skeleton of
thermy by Archosauromorpha involves inferring A. madagaskarensis (Fig. 1). This material can be
the resting metabolic rate of Allokotosauria, the confidently attributed to A. laaroussii. Among
sister-group of the clade Prolacerta–archosauri- the diagnostic postcranial characters of the Mala-
forms (Nesbitt et al. 2015; see also Ezcurra gasy form, only two can be checked on the avail-
2016). Therefore, we analyzed the bone histology able material of A. laaroussii, and both are
and inferred the resting metabolic rates of Azen- present: a posteriorly expanded T-shaped inter-
dohsaurus laaroussii (Archosauromorpha: Alloko- clavicle and hyposphene-hypantra interverte-
tosauria; Nesbitt et al. 2015; Ezcurra 2016) using bral articulations in anterior trunk vertebrae.
phylogenetic eigenvector maps. This later character state was considered autapo-
morphic for A. madagaskarensis by Nesbitt et al.
(2015), but it may be a synapomorphy for the
Materials and Methods
genus Azendohsaurus. The proportions of the
Materials studied skeletal elements MNHN.F.ALM 435
We histologically analyzed skeletal remains (right humerus), MNHN.F.ALM 497 (left
from the non-archosauriform archosauro- femur), and MNHN.F.ALM 369 (right tibia)
morph A. laaroussii from the Upper Triassic show that they are from different specimens.
(Carnian) of the Argana Basin (Morocco)
deposited at the Muséum National d’Histoire Histological Methods
Naturelle (MNHN). Azendohsaurus laaroussii Skeletal elements were molded and cast
Dutuit, 1972 was originally described on the before sectioning to preserve morphological
basis of a dental fragment and two isolated information. Casts and bone remains after
teeth specialized for plant feeding (Dutuit histological sampling were reposited at the
1972). On the basis on comparison with orni- paleontology collection of the MNHN, where
thischian dinosaurs such as Fabrosaurus and they are available upon a request to the curator.
Scelidosaurus, it was attributed to a new orni- Mid-diaphyses were embedded in epoxy resin
thischian dinosaur, making it one of the earliest and processed histologically following stand-
dinosaurs to be identified (Dutuit 1972; Thul- ard procedures (Lamm 2013). Transverse sec-
born 1973, 1974; Bonaparte 1976; Galton 1985, tions as well as longitudinal sections (at the
1990; Weishampel 1990; Gauffre 1993; Hunt anterior, posterior, dorsal, and ventral sides
and Lucas 1994; Flynn et al. 1999). In addition for humerus and femur and at the anterior, pos-
to the craniodental remains, hundreds of disar- terior, medial, and lateral sides for the tibia)
ticulated and monospecific postcranial remains were obtained and mounted on glass slides.
have been recovered from the locus typicus of A. These were analyzed and photographed in a

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 ORIGIN OF ENDOTHERMY IN ARCHOSAUROMORPHA 319

FIGURE 1. Azendohsaurus laaroussii Dutuit, 1972 (Argana Basin, Irohalène Member, Carnian): A, B, MNHN.F.ALM 586 left
humerus (A) and mold of MNHN.F.ALM 435 right humerus (B) in dorsal (dor), ventral (ven), anterior (ant), proximal (pro),
and distal (dis) views; C, D, MNHN.F. ALM 498 and 502, two ends of the same left femur glued (C) and mold of
MNHN.F.ALM 497 left femur (D) in ventral, dorsal, and posterior (pos) views; E, F, MNHN.F.ALM 398 right tibia (E)
and MNHN.F.ALM 369 mold of right tibia (F) in medial (med) and lateral (lat) views. The arrowheads show the level
of the thin sections. Other abbreviations: a.fi, articulation for fibula; cap, capittellum; dp, deltopectoral crest; ect, ectepicon-
dyle; gr, groove; int, internal trochanter; tro, trochlea.

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 320 JORGE CUBO AND NOUR-EDDINE JALIL

Nikon Eclipse E600POL microscope using nor- Quantitative Histology

mal light and cross-polarized light with a We quantified the vascular density (number
lambda compensator. Thin sections were of vascular canals by square millimeter) in
reposited at the vertebrate hard tissues histo- four orthogonal transects of the transverse sec-
logical collection of the MNHN, where they tions of humerus, femur, and tibia of A. laarous-
are available upon request to the curator (thin sii. We used the data set published by Legendre
sections MNHN-F-HISTOS 2348 to 2362). et al. (2016) for comparative purposes. It con-
tains the resting metabolic rates and the vascu-
Qualitative Histology lar densities of humerus, femur, and tibia of a
The nomenclature and classification of bone sample of 13 extant species and 14 extinct spe-
tissues is based on bone organization at differ- cies of tetrapods. Data for Mus musculus were
ent integration levels (Francillon-Vieillot et al. removed, because this species was an outlier.
1990). At the level of bone matrix organization,
three character states have been defined based Phylogenetic Comparative Methods
on the organization of collagen fibers: the pres- We tested whether vascular density explains
ence (or absence) of woven-fibered bone matrix a significant fraction of the variation of resting
(WB), parallel-fibered bone matrix (PFB), and metabolic rate using phylogenetic generalized
lamellar bone matrix (LB). At the level of organ- least-squares regressions (PGLS; Grafen 1989),
ization of bone as a tissue, two character states with the ‘caper’ package (Orme et al. 2013) in
have been defined based on the organization of R (R Core Team 2016). For this, we used the
the collagen fibers, the vascular density, and data set and the phylogeny published by
the cyclicity of bone apposition: the presence Legendre et al. (2016).
of lamellar-zonal bone (LZB) and the presence Moreover, we inferred the resting metabolic
of the fibrolamellar bone (FLB). See Francillon- rates of A. laaroussii and those for 14 species of
Vieillot et al. (1990) for a detailed description of extinct tetrapods using phylogenetic eigen-
these character states. This pattern-oriented vector maps (Guenard et al. 2013) with the
approach has been recently expanded toward ‘MPSEM’ package (Guenard et al. 2013) in R
a process-oriented view based on the develop- (R Core Team 2016). For this, we used our
mental origin of the components of bone tissue data from A. laaroussii and the data set and the
(Prondvai et al. 2014). In process-oriented phylogeny published by Legendre et al. (2016).
paleohistology, the developmental mechan- Finally, we performed an optimization of the
isms involved in the formation of bone tissue presence of fibrolamellar bone (using our data
are inferred from the fine architecture of the for A. laaroussii and published data for other
bone matrix. PFB and LB are formed by the pro- Archosauromorpha) and an optimization of
cess of dynamic osteogenesis, whereas WB is the presence of endothermy (using observed
formed by the process of static osteogenesis resting metabolic rates for extant species and
(Ferretti et al. 2002; Palumbo et al. 2004; Marotti inferred resting metabolic rates for extinct
2010; Prondvai et al. 2014; Stein and Prondvai ones as a proxy for the thermometabolic
2014; Cubo et al. 2017). Cortices can also have regime) onto phylogenies of Archosauromor-
a composite nature, including (1) a scaffold of pha and Tetrapoda, respectively, using parsi-
WB formed by the process of static osteogenesis mony in Mesquite (Maddison and Maddison
and (2) centripetal layers infilling primary 2015).
osteons made of PFB or LB formed by the pro-
cess of dynamic osteogenesis. This last tissue
Results
type has been named “woven-parallel com-
plex” by Prondvai et al. (2014). In the context Bone Histology of Azendohsaurus laaroussii
of process-oriented paleohistology, the widely Supplementary Fig. 1 shows composite
used term “FLB” is a special case of a woven- images of the entire diaphyseal cross sections
parallel complex in which there is abundant of humerus, femur, and tibia. The medial
woven bone and dense vascularity in the region of the transverse section of the tibia is
form of primary osteons (Prondvai et al. 2014). avascular (Fig. 2A). The bone matrix appears

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 ORIGIN OF ENDOTHERMY IN ARCHOSAUROMORPHA 321

FIGURE 2. Transverse (A, B, E, F) and longitudinal (C, D, G, H) sections of the medial (A–D) and posterior (E–H) regions of
the right tibia MNHN.F.ALM 369 of Azendohsaurus laaroussii under normal light (A, C, E, G) and cross-polarized light with
lambda compensator (B, D, F, H). Periosteum is on the left. Arrows, primary osteons; arrowheads, erosion cavities lined
along its periphery by a coating of endosteal lamellar tissue. Scale bars, 300 µm.

isotropic (red in online figure) under cross- avascular (Fig. 2C). In the longitudinal section,
polarized light with a lambda compensator the bone matrix appears anisotropic (blue in
(Fig. 2B) and black in cross-polarized light online figure, but it can be yellow depending
without a lambda compensator (not shown). on the orientation to the polarizers) under
This can be interpreted either as woven bone cross-polarized light with a lambda compensa-
or as parallel-fibered bone with fibers perpen- tor (Fig. 2D). These complementary patterns of
dicular to the transverse plane of section interference suggest that the cortex is com-
(Stein and Prondvai 2014). The longitudinal posed of parallel-fibered bone, with collagen
section of the same (medial) region is also fibers running parallel to the longitudinal axis

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 322 JORGE CUBO AND NOUR-EDDINE JALIL

FIGURE 3. Transverse sections of the left femur MNHN.F.ALM 497 (A, anterior region; B, posterior region; E, F, postero-
dorsal region) and the right humerus MNHN.F.ALM 435 (C, anterior region; D, posterior region; G, posterior region with
higher magnification) and a longitudinal section of the same humerus (H, posterior region) of Azendohsaurus laaroussii
under normal light (A–E) and cross-polarized light with lambda compensator (F–H). Periosteum is on the left in A and
C, on the right in B and D, and at the bottom in G. Endosteum is at the bottom in E, F, H. Arrow, primary osteon; arrowhead,
woven bone. Scale bars, 1 mm (A–D); 300 µm (E–H).

of the tibia. Osteocyte lacunae have a rounded lamellar tissue around the medullary cavity
aspect in the transverse section (Fig. 2A,B) (Fig. 2A,B).
and an elongate spindle shape in the longitu- The posterior region of the transverse section
dinal section (Fig. 2C,D). Their main axis is of the tibia is well vascularized (Fig. 2E). Vascu-
thus aligned with the collagen fibers of the sur- lar density is higher in the deep cortex than in
rounding matrix, whereas the canaliculi are the outer cortex. Vascular canals are longitu-
perpendicular (Fig. 2C,D). We can observe ero- dinal: they are rounded in transverse section
sion cavities lined by a coating of endosteal (Fig. 2E,F) and elongated (with the main axis

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 ORIGIN OF ENDOTHERMY IN ARCHOSAUROMORPHA 323

FIGURE 4. Resting metabolic rates quantified in a sample of extant tetrapods and inferred in a sample of extinct Archosaur-
omopha using the phylogeny and the vascular density of the humerus. Data for Azendohsaurus laaroussii were obtained in
this study and data for other taxa as well as the phylogeny were taken from Legendre et al. (2016). Resting metabolic rates
inferred for Azendohsaurus, Prolacerta, Garjainia, and Euparkeria are in the range of variation of extant mammals, whereas the
value inferred for Postosuchus is in the range of variation of extant birds.

parallel to the main axis of the tibia) in the lon- the lateral region contains a line of arrested
gitudinal section (Fig. 2G,H). The scarce vascu- growth split into two lines.
lar canals of the outer cortex are simple primary The stylopodial bones (femur and humerus)
canals, whereas those of the deep cortex form show a well-developed vascularization in the
small primary osteons (Fig. 2F). An avascular deep cortex, mainly formed by vascular canals
outer circumferential layer is absent. The bone parallel to the bone periphery (circular canals)
matrix is isotropic (it appears red in online fig- connected through abundant radial anasto-
ure) under cross-polarized light with a lambda moses (Fig. 3A–D). This pattern of vascular
compensator in the transverse section (Fig. 2F) orientation is more regular in the femur
and anisotropic (blue in online figure) in the (Fig. 3A,B) than in the humerus (Fig. 3C,D),
longitudinal section (Fig. 2H). This pattern sug- but vascular density decreases from the inner
gests that bone matrix in the posterior region of to the outer cortex in both cases. The bone
the tibia is composed of parallel-fibered bone, matrix is made of either a scaffold of woven
with collagen fibers running parallel to the lon- bone containing primary osteons (fibrolamellar
gitudinal axis of the tibia. Osteocyte lacunae bone; e.g., Fig. 3E,F) or a scaffold of parallel-
have a rounded aspect in the transverse section fibered bone containing both simple vascular
(Fig. 2E,F) and an elongate spindle shape in the canals and primary osteons (e.g., Fig. 3G,H).
longitudinal section (Fig. 2G,H). The anterior We can observe the anisotropic (blue in online
and the lateral regions of the tibia are similar figure) aspect of the scaffold of bone matrix in
to the posterior region in all respects but one: a longitudinal section of the humerus under

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 324 JORGE CUBO AND NOUR-EDDINE JALIL

FIGURE 5. Resting metabolic rates quantified in a sample of extant tetrapods and inferred in a sample of extinct Archosaur-
omopha using the phylogeny only. Data for Azendohsaurus laaroussii were obtained in this study and data for other taxa as
well as the phylogeny were taken from Legendre et al. (2016). Resting metabolic rates inferred for Azendohsaurus, Proter-
osuchus, and Maiasaura are in the range of variation of extant mammals; values inferred for Calyptosuchus, Lesothosaurus,
Allosaurus, and Troodon are in the range of variation of extant birds, whereas the value inferred for Rutiodon is intermediate.

polarized light, suggesting that it is made (we observed vascular canals near the bone
of parallel-fibered bone with collagen fibers periphery in many regions; e.g., Fig. 2E) and
oriented parallel to the main axis of the bone that lines of arrested growth are absent in all
(Fig. 3H). Please note that the matrix around analyzed regions but one (the lateral region of
vascular canals appears isotropic (red in online the tibia), we conclude that the analyzed speci-
figure) in this longitudinal section (Fig. 3H), mens were subadults. The two lines of arrested
whereas it appears anisotropoic (blue in online growth observed in the lateral region of the
figure) in the transverse section (Fig. 3G), sug- tibia probably correspond to the split of a single
gesting that it is made of parallel-fibered bone line formed during bone cortical drift, because
forming primary osteons with collagen fibers they are absent in the anterior, posterior, and
perpendicular to the main axis of the humerus. medial sides of the same section.
Avascular outer circumferential layers or lines
of arrested growth are absent in both femur Inferring the Resting Metabolic Rate of
and humerus (Fig. 3A–D). Azendohsaurus laaroussii
In summary, vascular canals are more abun- First, we tested whether vascular density
dant in the deep cortex than in the outer cortex explains a significant fraction of the variation
in all three bones, suggesting an ontogenetic of resting metabolic rate using the sample of
decrease of bone growth rates (e.g., Fig. 2E). extant tetrapods published by Legendre et al.
Considering that the avascular outer circumfer- (2016). For this, we used phylogenetic general-
ential layer of parallel-fibered bone is absent ized least-squares regressions (Grafen 1989). In

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 ORIGIN OF ENDOTHERMY IN ARCHOSAUROMORPHA 325

FIGURE 6. Resting metabolic rates quantified in a sample of extant tetrapods and inferred in a sample of extinct Archosaur-
omopha using the phylogeny and the vascular density of the tibia. Data for Azendohsaurus laaroussii were obtained in this
study and data for other taxa as well as the phylogeny were taken from Legendre et al. (2016). Resting metabolic rate
inferred for Azendohsaurus, Prolacerta, Proterosuchus, Garjainia, Lesothosaurus, Maiasaura, and Thecodontosaurus are in the
range of variation of extant mammals; the value inferred for Allosaurus is in the range of variation of extant birds, whereas
the value inferred for Lourinhanosaurus is intermediate.

all three cases (humerus, femur, and tibia), vas- Legendre et al. [2016]). According to the AIC
cular density explained highly significant por- procedure, the variable that maximizes the R 2
tions of the variation of resting metabolic rate and minimizes the AIC value should be
(85%, 73.3%, and 78.9%, respectively). Thus, a selected. In the case of the humerus, we
priori we can confidently infer the resting meta- obtained the same R 2 when using vascular
bolic rate of A. laaroussii using the vascular density and the phylogeny and when using
density values measured in the humerus, the only the phylogeny, but the AIC value was
femur, and the tibia (65.56, 34.83, and 47.19 lower when using both vascular density and
vascular canals by square millimeter, respect- the phylogeny (Fig. 4). We inferred a resting
ively) and the comparative data published by metabolic rate of 3.11 ml O2 h−1 g−0.67 and a
Legendre et al. (2016). confidence interval of 2.87–3.36 ml O2 h−1
Phylogenetic eigenvector maps (Guenard g−0.67 for A. laaroussii using the phylogeny
et al. 2013) includes an Akaike information cri- and vascular density of the humerus (Fig. 4).
terion (AIC) procedure to select the explanatory In the case of the femur, we found a slightly
variables (here, vascular density and/or the higher R 2 and a lower AIC value when using
phylogeny) to be used to infer the response the phylogeny only (Fig. 5). We inferred a rest-
variable (here, the resting metabolic rate) of ing metabolic rate of 2.20 ml O2 h−1 g−0.67 and a
extinct taxa (here, A. laaroussii and the sample confidence interval of 1.55–2.85 ml O2 h−1
of extinct Archosauromorpha published by g−0.67 for A. laaroussii using only the phylogeny

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 326 JORGE CUBO AND NOUR-EDDINE JALIL

FIGURE 7. Optimization using parsimony of the presence of fibrolamellar bone in the stylopodial and zeugopodial bones of
the Archosauromorpha analyzed to date (data summarized in Table 1). The topology is based on Nesbitt et al. (2009) for
Archosauriformes and on Nesbitt et al. (2015) for non-archosauriform Archosauromorpha. The minimum ages of taxa are
from the Paleobiology Database. Black branches indicate presence of fibrolamellar bone, white branches indicate absence,
and black and white branches indicate uncertain character state.

(this is equivalent to an optimization; Fig. 5). lamellar or nonlamellar parallel-fibered bone)

Finally, in the case of the tibia, we obtained (Prondvai et al. 2014). Therefore the capacity for
again the same R squared when using vascular reaching and maintaining very high bone growth
density and the phylogeny and when using rates during ontogeny, recorded by the presence
only the phylogeny, but the AIC value was of fibrolamellar bone in the bone cortex, may be a
lower when using both vascular density and reliable proxy to infer the metabolic rate of extinct
the phylogeny (Fig. 6). We inferred a resting animals (see “Introduction”). Four previous stud-
metabolic rate of 2.53 ml O2 h−1 g−0.67 and a ies have analyzed the evolution of the presence of
confidence interval of 2.25–2.82 ml O2 h−1 fibrolamellar bone in Archosauromorpha in a
g−0.67 for A. laaroussii using the phylogeny phylogenetic context (Botha-Brink and Smith
and the vascular density of the tibia (Fig. 6). 2011; Mukherjee 2015; Veiga et al. 2015; Jaquier
and Scheyer 2017). Here, we performed parsi-
mony optimizations of this character state
Discussion
(Fig. 7, Supplementary Figs. 2, 3) using our obser-
Evolution of Fibrolamellar Bone in vations on the bone histology of A. laaroussii and
Archosauromorpha previously published observations on other
The formation of fibrolamellar bone is very Archosauromorpha (summarized in Table 1).
energy consuming, because it involves the com- We observed three bone tissue types in the
bined action of static osteogenesis (to form a scaf- humerus, femur, and tibia of A. laaroussii:
fold of woven bone at high growth rates (1) avascular lamellar zonal bone (e.g.,
involving high rates of protein synthesis) and Fig. 2A–D); (2) a cortex formed by a scaffold of
dynamic osteogenesis (osteonal infilling with parallel-fibered bone (with collagen fibers

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 ORIGIN OF ENDOTHERMY IN ARCHOSAUROMORPHA 327

TABLE 1. Bibliographic compilation of the presence/absence of fibrolamellar complex in the long bones of the
non-archosaurian Archosauromorpha analyzed to date.

Presence of
fibrolamellar Figures within
Taxa complex Skeletal elements References references
Aenigmastropheus 1 Humerus (?) Ezcurra et al. (2014) Fig. 15A–D
parringtoni
Azendohsaurus laaroussii 1 Humerus, femur Cubo and Jalil, this study Fig. 2E,F
Chanaresuchus 1 “Long bone” Ricqlès et al. (2008) Plate 2, Fig. 3
Erythrosuchus africanus 1 Fibula Ricqlès et al. (2008) Plate 1, Fig. 4
Erythrosuchus africanus 1 Radius, tibia Botha-Brink and Smith (2011) Fig. 4A,B
Euparkeria capensis 0 Humerus or femur Ricqlès et al. (2008) Plate 2, Fig. 4
Euparkeria capensis 0 Humerus, femur, Botha-Brink and Smith (2011) Figs. 5A–D and 6A, B
tibia, fibula
Euparkeria capensis 1 Humerus Legendre et al. (2013) Fig. 1A,B
Hyperodapedon huxleyi 1 Humerus, radius, Mukherjee (2015) Figs. 2–7
femur, tibia
Hyperodapedon tikiensis 1 Humerus, radius, Mukherjee (2015) Figs. 2–7
femur, tibia
Macrocnemus bassanii 0 Jaquier and Scheyer (2017)
Prolacerta broomi 1 Tibia Botha-Brink and Smith (2011) Fig. 2A–C
Proterosuchus fergusi 1 Femur, tibia, fibula Botha-Brink and Smith (2011) Fig. 3A–D
Stenaulorhynchus stockleyi 0 Femur, tibia Werning and Nesbitt (2016) Figs. 3, 4
Tanystropheus 0 Humerus, femur Jaquier and Scheyer (2017) Fig. 2A,B
Teyumbaita sulcognathus 1 Tibia Veiga et al. (2015) Fig. 1B
Trilophosaurus buettneri 0 Humerus, ulna Werning and Irmis (2011)
Vancleavea campi 0 Humerus Nesbitt et al. (2009) Fig. 21B

parallel to the longitudinal axis of the bone) con- (Supplementary Fig. 3). All three analyses
taining either small primary osteons infilled by showed a flickering on-and-off pattern that pre-
lamellar parallel-fibered bone (inner part of the vents inferring the primitive condition for this
cortex; Fig. 2F) or simple vascular canals (outer clade. Moreover, the parsimony method fails
part of the cortex; Fig. 2F), and (3) fibrolamellar to find the condition for the four more inclusive
bone (e.g., Fig. 3E,F). Therefore, we coded the (basal) nodes of the clade. The only robust con-
presence of FLB for Azendohsaurus (Fig. 7, Sup- clusion obtained in the three analyses is con-
plementary Figs. 2, 3). gruent with that published by Botha-Brink
We coded the presence of FLB for Archo- and Smith (2011): the fibrolamellar bone was
sauria, because this character state has been acquired by the last common ancestor of Prola-
reported both in Pseudosuchia (e.g., Ricqlès certa–archosauriforms, with a reversion in Van-
et al. 2003; Padian et al. 2004; Nesbitt 2007; cleavea (Fig. 7, Supplementary Figs. 2, 3).
Tumarkin-Deratzian 2007) and in Ornithodira
(Ricqlès et al. 2000; Padian et al. 2004; Cubo Constraining the Phylogenetic Frame of the
et al. 2015). Consistently, Cubo et al. (2012) Acquisition of Endothermy by
inferred a high bone growth rate for the last Archosauromorpha
common ancestor of archosaurs. Lepidosauro- The resting metabolic rate may be a good
morpha has been chosen as an outgroup and proxy to infer the thermometabolic regime
is characterized by the absence of this character (endothermy vs. ectothermy) of extinct organ-
state (Fig. 7, Supplementary Figs. 2, 3). isms, because thermogenesis is very energy
We performed an optimization of the pres- consuming (see “Introduction”). Legendre
ence of FLB onto the phylogeny of Archosauro- et al. (2016) used this proxy in Archosauromor-
morpha published by Nesbitt et al. (2015) pha and concluded that endothermy may have
(Fig. 7). Moreover, we carried out two supple- been acquired by the last common ancestor of
mentary optimizations of the presence of the clade Prolacerta–archosauriforms. We
FLB using the phylogenies of Archosauro- inferred the resting metabolic rate of a member
morpha published by Pritchard et al. (2015) of the sister group of this clade (Azendohsaurus
(Supplementary Fig. 2) and Ezcurra (2016) laaroussii: Allokotosauria) as the next step to

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 328 JORGE CUBO AND NOUR-EDDINE JALIL

FIGURE 8. Optimization using parsimony of the presence of endothermy in a sample of tetrapods. The presence of endo-
thermy was inferred in extinct tetrapods using resting metabolic rates estimated through phylogenetic eigenvector maps.
Dark gray (red online) corresponds to the presence of endothermy and light gray (blue online) to the presence of ecto-
thermy. The phylogeny has been taken from Legendre et al. (2016).

better constrain the phylogenetic frame of the thermogenesis is very energy consuming, we
acquisition of endothermy by Archosauromor- can reasonably assume that all these taxa
pha. Moreover, we inferred resting metabolic were endotherms. A parsimony optimization
rate values for the sample of 14 extinct Archo- of the presence of endothermy in the whole
sauromorpha analyzed by Legendre et al. sample (including extant taxa of known ther-
(2016). The values inferred for all these taxa mometabolism and extinct taxa of inferred
are included in the range of variation measured thermometabolism) shows that endothermy
in the sample of extant endotherms (mammals was acquired twice, by mammals and by
and birds) (Figs. 4–6). Considering that the last common ancestor of the clade

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 ORIGIN OF ENDOTHERMY IN ARCHOSAUROMORPHA 329

Azendohsaurus–archosauriforms, and that Bonaparte, J. F. 1976. Pisanosaurus mertii Casamiquela and the ori-
gin of the Ornithischia. Journal of Paleontology 50:808–820.
within this last clade there is a reversion in Cro- Botha-Brink, J., and R. M. H. Smith. 2011. Osteohistology of the Tri-
codylia (Fig. 8). Considering that many archo- assic Archosauromorphs Prolacerta, Proterosuchus, Euparkeria, and
sauriforms have been found in the Permian Erythrosuchus from the Karoo Basin of South Africa. Journal of
Vertebrate Paleontology 31:1238–1254.
and that endothermy may have been acquired Buffrenil, V. de, and J .M. Mazin 1990. Bone histology of the
in a more inclusive clade (Azendohsaurus–archo- ichthyosaurs: comparative data and functional interpretation.
sauriforms), this key evolutionary event prob- Paleobiology 16:435–447.
Cubo, J., N. Le Roy, C. Martinez-Maza, and L. Montes. 2012. Paleo-
ably took place in the Permian. Bernard et al. histological estimation of bone growth rate in extinct archosaurs.
(2010) inferred fully developed endothermy Paleobiology 38:335–349.
for ichthyosaurs and plesiosaurs and incipient Cubo, J., H. Woodward, E. Wolff, and J. R. Horner. 2015. First
reported cases of biomechanically adaptive bone modeling in
endothermy for mosasaurs using a geochemical non-avian dinosaurs. PLoS ONE 10:e0131131.
approach. These conclusions are congruent Cubo, J., M. Hui, F. Clarac, and A. Quilhac. 2017. Static osteogen-
with those obtained using a paleohistological esis does not precede dynamic osteogenesis in periosteal ossifica-
tion of Pleurodeles (Caudata, Amphibia) and Pogona (Squamata,
approach (Buffrénil and Mazin [1990] for Lepidosauria). Journal of Morphology 278:621–628.
ichthyosaurs; Fleischle et al. [2018] for plesio- Davesne, D., F. J. Meunier, M. Friedman, R. B. J. Benson, and
saurs; and Houssaye et al. [2013] for mosa- O. Otero 2018. Histology of the endothermic opah (Lampris sp.)
suggests a new structure–function relationship in teleost fish
saurs). Thermogenetic mechanisms have also bone. Biology Letters 14:20180270. doi: 10.1098/rsbl.2018.0270.
been described in Actinopterygii: at the extrao- Dutuit, J.-M. 1972. Découverte d’un Dinosaure ornithischien dans
cular muscles that warm the eyes and brain in le Trias supérieur de l’Atlas occidental marocain. Comptes
Rendus de l’Académie des Sciences, Paris 275:2841–2844.
billfishes (Xiphioidei) and in the butterfly king- ——. 1976. Introduction à l’étude paléontologique du Trias contin-
fish Gasterochisma melampus (Scombridae: Gas- ental marocain. Description des premiers Stégocéphales recueillis
terochismatinae) (Davesne et al. 2018). Future dans le Couloir d’Argana (Atlas occidental). Mémoires du
Muséum National d’Histoire Naturelle, nouvelle série C 36:1–253.
research on the biochemical basis of thermogen- Ezcurra, M. D. 2016. The phylogenetic relationships of basal arch-
esis in extant taxa and on paleobiological infer- osauromorphs, with an emphasis on the systematics of proterosu-
ence of resting metabolic rates in extinct taxa are chian archosauriforms. PeerJ 4:e1778.
Ezcurra, M. D., T. M. Scheyer, and R. J. Butler. 2014. The origin and
needed to elucidate the origin and the evolu- early evolution of Sauria: Reassessing the Permian saurian fossil
tionary patterns of endothermy in Osteichthyes. record and the timing of the crocodile-lizard divergence. PLoS
ONE 9:e89165.
Ferretti, A., C. Palumbo, M. Contri, and G. Marotti. 2002. Static and
Acknowledgments dynamic osteogenesis: two different types of bone formation.
Anatomy and Embryology 206:21–29.
We thank R. Allain, curator of paleontology at Fleischle, C. V., T. Wintrich, and P. M. Sander 2018. Quantitative
histological models suggest endothermy in plesiosaurs. PeerJ
the MNHN, for giving us permission to perform 6:25. doi: 10.7717/peerj.4955.
histological analyses in the right humerus Flynn, J. J., J. M. Parrish, B. Rakotosamimanana, W. F. Simpson, R.
MNHN.F ALM 435, the left femur MNHN.F B. Whatley, and A. R. Wyss. 1999. A Triassic fauna from Mada-
gascar, including early dinosaurs. Science 286:763–765.
ALM 497, and the right tibia MNHN.F ALM Flynn, J. J., S. J. Nesbitt, J. M. Parrish, R. Ranivoharimana, and A.
369 of A. laaroussii. We also thank D. Germain, R. Wyss. 2010. A new species of Azendohsaurus (Diapsida: Arch-
curator of the vertebrate hard tissues histological osauromorpha) from the Triassic Isalo Group of southestern
Madagascar: cranium and mandible. Palaeontology 53:669–688.
collection of the MNHN, for including the thin Francillon-Vieillot, H., V. de Buffrénil, J. Castanet, J. Geraudie,
sections produced in this study in that collection. F. Meunier, J. Y. Sire, L. Zylberberg, and A. de Ricqlès. 1990. Micro-
Thanks also to H. Lamrous for the preparation of structure and mineralization of vertebrate skeletal tissues. Pp. 471–
548 in J. G. Carter, ed. Skeletal biomineralization: patterns, processes
the thin sections, to P. Loubry for the photos, and and evolutionary trends. Van Nostrand Reinhold, New York.
to S. Fernandez for the drawings. Finally, we Galton, P. M. 1985. Diet of Prosauropod dinosaurs from the Late
thank H. Woodward and an anonymous Triassic and Early Jurassic. Lethaia 18:105–123.
——. 1990. Basal Sauropodomorpha–Prosauropoda. Pp. 320–344
reviewer for their constructive comments. in D. B. Weishampel, P. Dodson, and H. Osmólska, eds. The Dino-
sauria. University of California Press, Berkeley.
Gauffre, F.-X. 1993. The prosauropod dinosaur Azendohsaurus laarous-
Literature Cited sii from the Upper Triassic of Morocco. Palaeontology 36:897–908.
Grafen, A. 1989. The phylogenetic regression. Philosophical Trans-
Bernard, A., C. Lécuyer, P. Vincent, R. Amiot, N. Bardet, N., actions of the Royal Society of London B 326:119–157.
E. Buffetaut, G. Cuny, F. Fourel, F. Martineau, J. M. Mazin, and Guenard, G., P. Legendre, and P. Peres-Neto. 2013. Phylogenetic
A. Prieur. 2010. Regulation of body temperature by some Meso- eigenvector maps: a framework to model and predict species
zoic marine reptiles. Science 328:1379–1382. traits. Methods in Ecology and Evolution 4:1120–1131.

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user
 330 JORGE CUBO AND NOUR-EDDINE JALIL

Houssaye, A., J. Lindgren, R. Pellegrini, A. H. Lee, D. Germain, and Orme, D., R. Freckleton, G. Thomas, T. Petzoldt, S. Fritz, N. Isaac,
M. J. Polcyn. 2013. Microanatomical and histological features in and W. Pearse. 2013. Caper: comparative analyses of phyloge-
the long bones of mosasaurine mosasaurs (Reptilia, Squamata) netics and evolution in R. https://CRAN.R-project.org/
—implications for aquatic adaptation and growth rates. PLoS package=caper.
ONE 8(10):1–12. doi: 10.1371/journal.pone.0076741. Padian, K., and J. R. Horner. 2002. Typology versus
Hunt, A. P., and S. P. Lucas. 1994. Ornithischian dinosaurs from the transformation in the origin of birds. Trends in Ecology and Evo-
Upper Triassic of the United States. Pp. 227–241 in N. C. Fraser lution 17:120–124.
and H.-D. Sues, eds. In the shadow of the dinosaurs. Early Meso- Padian, K., J. R. Horner, and A. De Ricqles. 2004. Growth in small
zoic tetrapods. Cambridge University Press, Cambridge. dinosaurs and pterosaurs: the evolution of archosaurian growth
Jalil, N. E., and F. Knoll. 2002. Is Azendohsaurus laaroussii (Carnian strategies. Journal of Vertebrate Paleontology 24:555–571.
Morocco) a dinosaur? Journal of Vertebrate Paleontology 22 Palumbo, C., M. Ferretti, and G. Marotti. 2004. Osteocyte dendro-
(Suppl. to No. 3):70A. genesis in static and dynamic bone formation: an ultrastructural
Jaquier, V. P., and T. M. Scheyer. 2017. Bone histology of the Middle study. Anatomical Record A 278A:474–480.
Triassic long-necked reptiles Tanystropheus and Macrocnemus Pritchard, A. C., A. H. Turner, S. J. Nesbitt, R. B. Irmis, and N.
(archosauromorpha, Protorosauria). Journal of Vertebrate Pale- D. Smith. 2015. Late Triassic tanystropheids (Reptilia, Archo-
ontology 37:e1296456. sauromorpha) from northern New Mexico (Petrified Forest Mem-
Khaldoun, F. 2014. Les vertébrés du Permien et du Trias du Maroc ber, Chinle Formation) and the biogeography, functional
(Bassin d’Argana, Haut Atlas Occidental) avec la réévaluation morphology, and evolution of Tanystropheidae. Journal of Verte-
d’Azendohsaurus laaroussii (Reptilia, Archosauromorpha) et la brate Paleontology 35:e911186.
description de Reptilia Moradisaurinae et Rhynchosauria nou- Prondvai, E., K. H. W. Stein, A. de Ricqles, and J. Cubo. 2014.
veaux: anatomie, relations phylogénétiques et implications bios- Development-based revision of bone tissue classification: the
tratigraphiques. Unpublished Ph.D. thesis. University Cadi importance of semantics for science. Biological Journal of the Lin-
Ayyad, Marrakesh, Morocco. 342 pp. nean Society 112:799–816.
Lamm, E. T. 2013. Preparation and sectioning of specimens. Pp. 55– R Core Team. 2016. R: a language and environment for statistical
160 in Bone histology of fossil tetrapods: advancing methods, computing. R Foundation for Statistical Computing, Vienna,
analysis, and interpretation. University of California Press, Austria. http://www.R-project.org.
Berkeley. Ricqlès, A. de, K. Padian, J. R. Horner, and H. Francillon-Vieillot.
Legendre, L. J., L. Segalen, and J. Cubo. 2013. Evidence for high 2000. Palaeohistology of the bones of pterosaurs (Reptilia : Arch-
bone growth rate in Euparkeria obtained using a new paleohisto- osauria): anatomy, ontogeny, and biomechanical implications.
logical inference model for the humerus. Journal of Vertebrate Zoological Journal of the Linnean Society 129:349–385.
Paleontology 33:1343–1350. Ricqlès, A. de, K. Padian, and J. R. Horner. 2003. On the bone hist-
Legendre, L. J., G. Guenard, J. Botha-Brink, and J. Cubo. 2016. ology of some Triassic pseudosuchian archosaurs and related
Palaeohistological evidence for ancestral high metabolic rate in taxa. Annales de Paléontologie 89:67–101.
Archosaurs. Systematic Biology 65:989–996. Ricqlès, A. de, K. Padian, F. Knoll, and J. R. Horner. 2008. On the
Lowell, B. B., and B. M. Spiegelman. 2000. Towards a molecular origin of high growth rates in archosaurs and their ancient rela-
understanding of adaptive thermogenesis. Nature 404:652–660. tives: complementary histological studies on Triassic archosauri-
Maddison, W. P., and D. R. Maddison. 2015. Mesquite: a modular forms and the problem of a “phylogenetic signal” in bone
system for evolutionary analysis. http://www.mesquiteproject.org. histology. Annales de Paléontologie 94:57–76.
Marotti, G. 2010. Static and dynamic osteogenesis. Italian Journal Rowland, L. A., N. C. Bal, and M. Periasamy. 2015. The role of
of Anatomy and Embryology/Archivio italiano di anatomia ed skeletal-muscle-based thermogenic mechanisms in vertebrate
embriologia 115:123–126. endothermy. Biological Reviews 90:1279–1297.
Montes, L., N. Le Roy, M. Perret, V. De Buffrenil, J. Castanet, and J. Stein, K., and E. Prondvai. 2014. Rethinking the nature of fibrola-
Cubo. 2007. Relationships between bone growth rate, body mass mellar bone: an integrative biological revision of sauropod plexi-
and resting metabolic rate in growing amniotes: a phylogenetic form bone formation. Biological Reviews 89:24–47.
approach. Biological Journal of the Linnean Society 92:63–76. Thulborn, R. A. 1973. Teeth of ornithischian dinosaurs from the
Mukherjee, D. 2015. New insights from bone microanatomy of the Upper Jurassic of Portugal. Serviços Geológicos de Portugal,
Late Triassic Hyperodapedon (Archosauromorpha, Rhyncho- Memoria 22: 89–134.
sauria): implications for archosauromorph growth strategy. ——. 1974. A new heterodontosaurid dinosaur (Reptilia:
Palaeontology 58:313–339. Ornithischia) from the Upper Triassic Red Beds of Lesotho.
Nesbitt, S. 2007. The anatomy of Effigia okeeffeae (Archosauria, Zoological Journal of the Linnean Society of London 55:151–175.
Suchia), theropod-like convergence, and the distribution of Tumarkin-Deratzian, A. R. 2007. Fibrolamellar bone in wild
related taxa. Bulletin of the American Museum of Natural History adult Alligator mississippiensis. Journal of Herpetology
302:1–84. 41:341–345.
Nesbitt, S. J., M. R. Stocker, B. J. Small, and A. Downs. 2009. The Veiga, F. H., M. B. Soares, and J. M. Sayao. 2015. Osteohistology of
osteology and relationships of Vancleavea campi (Reptilia: Archo- hyperodapedontine rhynchosaurs from the Upper Triassic of
sauriformes). Zoological Journal of the Linnean Society Southern Brazil. Acta Palaeontologica Polonica 60:829–836.
157:814–864. Weishampel, D. B. 1990. Dinosaurian distribution. Pp. 63–139 in D.
Nesbitt, S. J., J. J. Flynn, A. C. Pritchard, J. M. Parrish, B. Weishampel, P. Dodson, and H. Osmólska, eds. The Dino-
L. Ranivoharimanana, and A. R. Wyss. 2015. Postcranial oste- sauria. University of California Press, Berkeley.
ology of Azendohsaurus madagaskarensis (?middle to Upper Trias- Werning, S., and R. B. Irmis. 2011. Reconstructing growth of the
sic, Isalo Group, Madagascar) and its systematic position among basal archosauromorph Trilophosaurus. Integrative and Compara-
stem archosaur reptiles. Bulletin of the American Museum of Nat- tive Biology 51:E147.
ural History 398:1–126. Werning, S., and S. J. Nesbitt. 2016. Bone histology and growth in
Nowack, J., S. Giroud, W. Arnold, and T. Ruf. 2017. Muscle non- Stenaulorhynchus stockleyi (Archosauromorpha: Rhynchosauria)
shivering thermogenesis and its role in the evolution of endo- from the Middle Triassic of the Ruhuhu Basin of Tanzania. Comp-
thermy. Frontiers in Physiology 8:ar889. tes Rendus Palevol 15:163–175.

Downloaded from https://pubs.geoscienceworld.org/paleobiol/article-pdf/45/2/317/4700607/s0094837319000137a.pdf

by Yonsei University user